Lend Academy Network Forum

Lending Club Discussion => Investors - LC => Topic started by: brycemason on June 14, 2014, 11:52:15 AM

Title: Loan Information Asymmetrical Timing
Post by: brycemason on June 14, 2014, 11:52:15 AM
Debated posting this, but so many ramifications I felt it best to do so. Original post in my forum. Reported here for broad applicability. The question has been why my systems (and I imagine those of many) have been less capable of doing business since the latest LC change. I detail the change and consequences below.

The issue is that P2P-Picks relies on LC providing its loan application data via the CSV or API, and that these are no longer the most timely way to get loan application data. It is affecting me personally and all of my clients, too, and there isn't much I can do about it except complain to LC, which I have done.

It used to be that when you clicked the link to get the CSV, LC would instruct its systems to generate a fresh CSV based on the loan applications that were available at that very moment. This, I imagine, was expensive for them when thousands of people were clicking at the magic hours. Hence why we saw rate limiters in place last November. In the last few days, they moved to a model of linking to a static CSV available for download that is updated for freshness just once every minute. While this relieves the stress off their servers, it has had some unintended (or perhaps intended) consequences.

Here is the rub. LC releases the loans onto their platform around 30-40 second after the hour. The loans are available to website users a full 20-30 seconds before they ever even appear in the CSV. The API has always been out of date a number of seconds, just like the CSV is now. Now, the absolute fastest way to get loan application data is to use the website interactively via their filter processes (or to screen scrape for automation). This puts all of us who want to do active loan selection in a position of either sitting at our terminals every release twitching like some kid playing a video game or breaking the rules (even more than we already have to do) and resorting to screen scraping loan data off the website in order to automate things. Of course, the web pages don't even have a majority of the credit variables anyway, so this may be of limited value.

I have witnessed this occur on both the fractional and whole loan side. Reliably, at 30 after the hour, new loans will appear. Within a few seconds, the "good stuff" (to the extent you wish to believe in such a concept) is either purchased outright (in the whole loan market) or is mostly already full (on the fractional side) by the time the other two loan data retrieval methods are updated. Here is an example from 6pm a couple nights ago to illustrate:


5:55:00 - Initial count: 31 whole loans on platform
6:00:05 - Unused whole loans expire (12 hour rule): 19 whole loans on the platform
6:00:30 - 170 new whole loans added to platform (as viewed from the webpage) - total inventory = 189
6:00:45 - 24 whole loans disappear off the platform - total inventory = 165
6:01:00 - browseNotes.csv and LoansBrowseLoans update with just 165 whole loans in them

Why would LC do this? I don't want to go into pet theories, even though I have some, before allowing LC a chance to respond. And to be clear LC is absolutely still providing an even playing field. Same rules for everyone. It's just that this change has benefitted a certain type of user (the twitching kid or the scraper).

My recommendation is to synch the timing of all forms of loan application data, website, CSV, and API. If the Royal Bank of Canada wrote Thor to synchronize orders across multiple stock exchanges, I'm pretty sure this is a feasible request.


Title: Re: Loan Information Asymmetrical Timing
Post by: Ran on June 14, 2014, 01:17:06 PM
What happened to LC's own autoinvest? Does it get loans as fast as screen scrapping?
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on June 14, 2014, 01:46:45 PM
That is a very important question. All third party automated investment systems designed for retail investors that I know of rely on the now-outdated CSV and API for loan input data. Either there are a lot of twitch buyers sitting on the site, or LC's own Prime service is getting first crack at automated investing. Those are the only two options for explaining why so many fractional loans are largely filled by the time of the CSV/API update. If it's the latter, it could be argued that their internal systems have an unfair advantage, as they wouldn't have to resort to screen scraping to organize the data (which takes time to poll individual web pages).
Title: Re: Loan Information Asymmetrical Timing
Post by: sociallender on June 14, 2014, 02:38:55 PM
This is troublesome. I don't understand the need for an API if the information is stale.  I can understand limiting the CSV file and metering to deter screen scraping.  However, an API should provide the same level of service as their front end.  This may have the opposite affect of now requiring auto invest users to use their front end (screen scrape) to place orders.

PLS does have the feature of using the web site front end to place orders based on previously saved LC filters.  However, there is no way to apply sophisticated logic such as AI, ratios,etc..  In other words, we can still place orders faster than those who are doing it the manual way.  However, we won't have the ability to analyze the loan data until the API or CSV is updated.  IMHO, the reason for the API is the ability to automate the loan selection process.  I don't understand why they would give the web site an unfair advantage?  I see this as a migration from the API to the web site which may be an unanticipated consequence. 

Bryce, in your analysis, is the API updated at the same time as the CSV?  Are there any advantages with one over the other? 

 
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on June 14, 2014, 02:59:21 PM
The API appears to update simultaneously with the CSV, as least when it matters at 1 minute after the hour. Historically the API was a little stale in my experience. This is all just my limited experience. I'm not going about this in a super scientific way, but have observed a number of times.
Title: Re: Loan Information Asymmetrical Timing
Post by: lascott on June 14, 2014, 04:38:27 PM
Timing seems to matter!  I was quite happy with P2P-Picks and BV the way it was working!  I suspect LendingRobot is affected as well.  We have another account using it but it is pretty selective so not sure if the note investing have slowed down yet. They seem to be.

Title: Mad Rush at Lending Club Loan Release Time: Part I
Posted by Anil Gupta | 08 Jun 2014 22:11:48 | Category: Lending Club
https://www.peercube.com/blog/post/21
Title: Re: Loan Information Asymmetrical Timing
Post by: sociallender on June 14, 2014, 05:09:28 PM
Quote
I suspect LendingRobot is affected as well.

Indeed, it affects all auto invest 3rd party software.  LC is basically preventing the download of loan data until it decides to update the CSV or API.  I do not see a way around it even with screen scrape.  It basically gives an advantage to those using the shopping cart hole or those using their filtering technology provided by the web site. 
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on June 14, 2014, 05:14:48 PM
I encourage others to verify my own observations.
Title: Re: Loan Information Asymmetrical Timing
Post by: sociallender on June 14, 2014, 05:19:28 PM
Another issue is when to download the CSV or API data?  Prior, the note count could be used to determine when to download the CSV/API loan data.  As soon as the note count increased by a significant amount, you could download the loan data in one shot.  Now, I can't think of a way to determine when to download the API/CSV.  There is no note count in the API (which is absolutely ridiculous).  The only thing I can think of is to continue to monitor the note count then continuously poll the CSV/API for an update?  If that is the case, then we are back to the main problem of the expense/load to handle the poll traffic.  That is, we have to resort to a continuous download of the CSV/API until it catches up with the note count.  Any ideas on how to make this a smoother process? 
Title: Re: Loan Information Asymmetrical Timing
Post by: GS on June 14, 2014, 10:00:00 PM
I can't even begin to imagine why the website, API, and CSV can't be updated simultaneously ...

LC has set up system that encourages screen scrapping and exploiting the shopping cart lock out while the crafty users wait for the CSV to update.  I mean, if I wanted the best loans, and was willing to break LC's rules and possibly the law (is willfully delaying an open market order, and preventing others from placing an order, while one waits for updated information to be posted, illegal market manipulation? -- well, we know people are doing it already), I would design my automation to screen scrape my prospects from the website based on the limited info posted there, put them in the shopping cart, then wait for the expanded CSV to be posted before I select which ones to keep....

This can't be the vision that LC has in mind for the future of Automation ....

Title: Re: Loan Information Asymmetrical Timing
Post by: Lovinglifestyle on June 14, 2014, 10:38:06 PM
I use the shopping cart for recent listings of filters-applied EFG loans because I have more uninvested cash than the new listings will consume.  If LC had a way to sort KEEP from DISCARD into two piles I could get through the list faster.  I don't waste time "waiting" for a .csv to update.  I  go straight to work on what's there and check the 3rd sites when I think they've had time to publish.  Waiting instead of loading the cart would mean losing all the lower amount, 36 month loans without even being able to see them in order to make a decision! HOW ON EARTH IS THIS CHEATING??  It seems like common sense to me.

If LC's filters had the same filter options the 3rd party sites do there would be a chance I could use their automation.

A point of interest to me in all this recent discussion about the timing of available loans is that Interest Radar is very, very fast to have them all.  Rev must have a secret sauce.  Unfortunately for me, I'm addicted to an additional data point only available on a different service, but it too is ready while I'm still paring down my list.

Title: Re: Loan Information Asymmetrical Timing
Post by: GS on June 14, 2014, 11:04:28 PM
I don't know if that last post was directed at me, or not (I never called anything "hoarding" or "cheating", nor did anyone else), but I completely agree that as long as LC is going to tilt the field to encourage screen scrapping and use of the shopping cart to retain notes while you think about whether or not you really want to buy it, well, the crafty investors will do just that.  In fact, I posted on the PLS section (http://www.lendacademy.com/forum/index.php?topic=1907.0) several months ago that I'd love to see PLS modified to do a quick LC filter grab of notes, then exploit the shopping cart lock out while it runs it's models.

As you pointed out, it works best if you have lots of cash to invest, and grab all or nearly all the new notes, which is why I think this has probably become the choice tactic of the institutional investors. 

SO, I'm not being judgmental, I'm right there with you, that it is the best way to get the best notes.  When I used to "hand invest", I'd do the same thing.  Refresh like a mo-fo until the new loans appeared, quick filter, sort, click-click-click the top notes with the highest yields, add to cart .... then start looking at the quality of what I selected.  However, that strategy stopped working about 6-7 months ago when I found I could not do it by hand fast enough to get the top notes in my filter, which led me to believe the game had changed and the computers were now doing what I was doing, but doing it much faster. 


But, while I do admit to doing this myself, I've always wondered if the tactic was legal, in that I was preventing others from participating in a fast paced market while I took my sweet time to decide what I wanted to keep, and what I wanted to throw back.  Can you imagine if the SEC discovered that some hedgefund had discovered a way to prevent other's from trading a stock while they ran models to see if they wanted to buy it or not .... I'd imagine they would not be amused.  That's a much closer analogy to consider than taking two pairs of pants into the dressing room. 
Title: Re: Loan Information Asymmetrical Timing
Post by: mchu168 on June 14, 2014, 11:45:12 PM
I see people locking up loans from high income borrowers right at feeding time, but then release them after further review.  There's definitely a lot of gaming the system going on to get some perceived "edge."  Someone needs to write a version of "Flash Boys" as it pertains to the Lending Club world.
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on June 15, 2014, 12:01:24 AM
It's not cheating to use the website to buy a bunch of loans right at release. Nobody said it was. To the contrary my original post said the playing field is indeed even. Everyone gets the same info at the same time. We are just complaining that the supported official ways to get loan applications are unequal. LendingClub is my main hobby and I have spent hundreds of hours building models, culturing relationships, and writing code to automate loan selection that is based on the premise that use of the API or CSV will allow me to be competitive in the marketplace. For two years that was true. Last week it got turned on its head. We simply can't do the type of analysis we want to do now.
Title: Re: Loan Information Asymmetrical Timing
Post by: Fred on June 15, 2014, 02:03:10 AM
It is not too difficult for LC to implement a "shopping cart" like that of Amazon:

http://www.amazon.com/gp/help/customer/display.html?nodeId=468468
Quote
Placing an available item in your shopping cart does not reserve that item.

Quote
Available inventory is only assigned to your order after you click Place your order and receive an e-mail confirmation that we've received your order.
Title: Re: Loan Information Asymmetrical Timing
Post by: rawraw on June 15, 2014, 07:45:03 AM
Curious to hear LC's reply.  You sent this to them Bryce?
Title: Re: Loan Information Asymmetrical Timing
Post by: Lovinglifestyle on June 15, 2014, 09:53:59 AM
I'm also curious to hear LC's reply to Bryce.

Fairness is quite the sticky wicket.  API, CSV, and the site all feeding at the same time seems fairest to me.

GS, sorry about the pants analogy.  That was bad.  I like your stock analogy much better.  I'll edit mine out.

Fred, the Amazon system would create a lot of one-note orders from individuals.  It is fair, but the order proliferation could be a problem if enough manual selectors hang on.

Bryce and GS, I don't remember who used the words hoarding and cheating in a different thread yesterday, but it wasn't you and that viewpoint is not new.  It has long been a criticism of the system and has sometimes been directed at individuals (not at me, but feel free to do so now!).

btw, IR was slow this morning and too late for some loans.  I'm grateful that the 3rd sites are still operating at all and for all the work their developers have done.  Don't know why LC ignores the main reasons people are using them while trying to be so "improved".


 

Title: Re: Loan Information Asymmetrical Timing
Post by: GS on June 15, 2014, 11:35:03 AM
Don't know why LC ignores the main reasons people are using them while trying to be so "improved".

I couldn't agree more.  If LC offered a filtering service in which one could filter all the available data, and write SQL type statements like (Mo. Payment/Mo. Income < 0.10), etc, as part of their filtering, with an instant auto purchase option, priority given to buyers who have gone the longest without having an order placed .... then there would be hardly any need for a 3rd party site.   Fact is, the LC website doesn't even display all the criteria available in the CSV, much less let you utilize it.   I didn't even know more criteria was available until I found this site and saw Bryce and others talking about it.  ("Say there, how did you know this guy has 5 mortgages?") 

 I don't think there is any question now that these changes are a way to make it very difficult, or impossible, for the 3rd party sites to operate.  LC needs to focus on their own shortcomings rather than trying to shut down the sites that are getting it right!
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on June 15, 2014, 01:05:12 PM
I'm sure the third party sites appreciate the acknowledgement of value added, but let's keep it on topic. It is more likely that this was a change that wasn't so well thought out and had unintended consequences.

I sent the observations to LC (one of the gentlemen in the investor services dept.) and haven't yet heard a reply beyond acknowledgement of receipt and forwarding. But, it is the weekend. Hopefully we can see equal treatment across all info feeds in the near future.
Title: Re: Loan Information Asymmetrical Timing
Post by: BruiserB on June 15, 2014, 07:58:50 PM
Bryce,

Do you have any feel for how many of the loans that seem to be getting scooped up before you can get the data to run your analysis would have been top P2P-Picks?  Is it possible for you to analyze the loans that are selling quickly to determine any pattern of what might be making them desirable?
Title: Re: Loan Information Asymmetrical Timing
Post by: Fred on June 15, 2014, 08:43:36 PM
I think the concept of "feeding times" (currently, 4 daily) should now be abandoned.

None of this asymmetric timing issues would matter.  It should also solve a lot of bursty demand of CPU, memory, network, etc.
Title: Loan Information Asymmetrical Timing
Post by: brycemason on June 15, 2014, 09:21:24 PM
On the whole loan side, my picks went from many per release to zero or one, so it was a great impact. On my personal account and the picks list, the volume declined precipitously. I have over $1100 in my personal account which has not happened in years.
Title: Re: Loan Information Asymmetrical Timing
Post by: BruiserB on June 16, 2014, 11:39:51 AM
On the whole loan side, my picks went from many per release to zero or one, so it was a great impact. On my personal account and the picks list, the volume declined precipitously. I have over $1100 in my personal account which has not happened in years.

Yea, that's significant.  Hmm.  Seems odd to me, because the loans P2P-Picks chooses seem quite different than I would get using any kind of conventional filter.  I can't imagine people using Lending Club's Automated Investing are clamoring for the same loans your service chooses.  Also does Lending Club even offer a version of Automated Investing on the whole loan side? 
Title: Re: Loan Information Asymmetrical Timing
Post by: rawraw on June 16, 2014, 02:32:22 PM
I'd be willing to bet most people are just trying to gobble up high interest rate notes indiscriminately. 
Title: Re: Loan Information Asymmetrical Timing
Post by: sociallender on June 16, 2014, 03:52:37 PM
Quote
I'd be willing to bet most people are just trying to gobble up high interest rate notes indiscriminately.

My bet is most are using the LC cart or browse notes filtering (either manually or scripted).  The former requires cash drag (not good with my account), the latter uses a limited set of information and a lot of time (if manual).  In either case, I don't see any way to analyze loan information at the same time (or before) this happens.  Looks like I will be using scripted browse notes filtering (very tight saved filter) then fall back to loan analysis once the CSV/API is updated.

I sent a note to LC for explanation.  No response yet...  Hopefully this was not intended, and resolved soon.
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on June 16, 2014, 05:14:55 PM
A fundamental question is whether a note locked up in a shopping cart (which is fully funded) will even appear in a CSV/API call. You need that to be true in order to get the full credit variables for a future analysis.
Title: Re: Loan Information Asymmetrical Timing
Post by: sociallender on June 16, 2014, 06:47:31 PM
Looks like the CSV delay is permanent (from Lending Club support):

Thank you for your email and interest in Lending Club. If you are using an API that gets loan information from the "Download All" in our Browse Notes tab, there is a small delay. The "Download All" information is updated once per minute rather than continuously. This is a permanent feature to allow equal access to the platform for all investors.

My question also asked about the API.  However, the response appears to be referencing the web site as the API (which doesn't make sense).  I replied with clarification regarding the API and will repost.  However, my own analysis confirms that the API is indeed delayed (I believe in sync with the CSV). 

I am trying to stay open minded about this whole thing.  However, I just don't see this as a viable solution.  Although the rules are the same for everyone, this is definitely not "equal" access to all investors.  If you are one type of investor, you get the front of the bus.  If you are another type, you get the back of the bus.  Just change your investor type and you get to the front of the bus.  Therefore, I believe most savvy investors are going to have to  change their investor type, and now we are all competing in much the same way we did before things changed.  Thus cutting out the so called "equal" access.  In addition, you now have to use the limited filtering LC provides or use the cart hole. 

Ugghhh.  I wonder if it's time to make the switch to Prosper?


 
Title: Re: Loan Information Asymmetrical Timing
Post by: sociallender on June 16, 2014, 06:59:07 PM
Maybe off tangent but applicable:

Even if the API isn't delayed by content(which I believe it is), it is appears to be delayed in execution.  On average, the download time of the CSV file takes me about 4-6 seconds without throttle on repeat downloads.  Downloading the notes via API, appears to be about 5-10 seconds on first attempt.  Then up to a minute on repeat attempts.  Since there is no indicator on when the API is updated, you are forced to repeatedly download the CSV file.  Hence, migration from the API. 
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on June 16, 2014, 08:03:20 PM
I declare the API/CSV dead. Well, the API was always dead. But now it's like a zombie, infecting the CSV with its malaise. Dead and greyed, patchy flesh hands grasping at the last wisps of higher-interest borrowers as they run away from the pack, just out of reach.

Time to adjust.
Title: Loan Information Asymmetrical Timing
Post by: BruiserB on June 16, 2014, 11:06:43 PM

Looks like the CSV delay is permanent (from Lending Club support):

Thank you for your email and interest in Lending Club. If you are using an API that gets loan information from the "Download All" in our Browse Notes tab, there is a small delay. The "Download All" information is updated once per minute rather than continuously. This is a permanent feature to allow equal access to the platform for all investors.

My question also asked about the API.  However, the response appears to be referencing the web site as the API (which doesn't make sense).  I replied with clarification regarding the API and will repost.  However, my own analysis confirms that the API is indeed delayed (I believe in sync with the CSV). 

I am trying to stay open minded about this whole thing.  However, I just don't see this as a viable solution.  Although the rules are the same for everyone, this is definitely not "equal" access to all investors.  If you are one type of investor, you get the front of the bus.  If you are another type, you get the back of the bus.  Just change your investor type and you get to the front of the bus.  Therefore, I believe most savvy investors are going to have to  change their investor type, and now we are all competing in much the same way we did before things changed.  Thus cutting out the so called "equal" access.  In addition, you now have to use the limited filtering LC provides or use the cart hole. 

Ugghhh.  I wonder if it's time to make the switch to Prosper?

But what "type" of investor is getting front of bus access?  Feels like we're all at the back now.....clearly someone is up front, but if the only way to get there is to screen-scrape or throw everything in your cart and sort it out later, then this clearly can't be the "equal" access Lending Club says they're trying to provide.  I hope Lending Club knows we're all watching carefully and this doesn't seem like a good time to p!$$ off those of us who have gotten them to this stage of their development.

I was getting frustrated with trying to keep my funds invested near the end of last year and had about given up.  Early this year I discovered third party automation and am thankful for the efforts of Bryce, Nathan, Emmanuel, and others.  I actually more than doubled my investment in Lending Club since discovering these tools!  If Lending Club makes it impossible for these services to function, then I'll be right back where I was late last year.....considering throwing in the towel.   I hope someone from Lending Club is paying attention!


Sent from my iPad using Tapatalk
Title: Re: Loan Information Asymmetrical Timing
Post by: Fred on June 16, 2014, 11:16:13 PM
Thank you for your email and interest in Lending Club. If you are using an API that gets loan information from the "Download All" in our Browse Notes tab, there is a small delay. The "Download All" information is updated once per minute rather than continuously. This is a permanent feature to allow equal access to the platform for all investors.

If PRIME (Automated Investing) has access to all data prior to CSV publication, then the claim "to allow equal access to the platform for all investors" is definitively not true. 
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on June 16, 2014, 11:53:48 PM
Fred, that's where I said I wasn't going to go in my original post. But that's my thought, too. Only one way to tell...find a few people with PRIME, identify the loans bought into, and compare with lists of loans that never made it to my archived CSVs. I really hope that isn't happening, because it would be an implicit acknowledgement that there's a "best" set of loans, something they've vociferously argued against the entire time.
Title: Re: Loan Information Asymmetrical Timing
Post by: rockinray on June 17, 2014, 07:28:44 AM
Bryce,

I am testing the "Automated Investing" service with them right now.

How can I help with this?

Ray
Title: Re: Loan Information Asymmetrical Timing
Post by: Fred93 on June 17, 2014, 02:29:00 PM
Looks like the CSV delay is permanent (from Lending Club support):

Thank you for your email and interest in Lending Club. If you are using an API that gets loan information from the "Download All" in our Browse Notes tab, there is a small delay. The "Download All" information is updated once per minute rather than continuously. This is a permanent feature to allow equal access to the platform for all investors.

Keep in mind you just got a form leter response from a low level employee.  This has to be escalated.  There are obviously some people in the middle who don't understand, and need to be educated.
Title: Re: Loan Information Asymmetrical Timing
Post by: Fred on June 17, 2014, 09:51:53 PM
I declare the API/CSV dead...

IMO, CSV can still be useful. 

With CSV polling and addToPortfolio.action re-tries (to catch loans that have been unlocked from shopping cart), there should be enough 'good' loans to invest in.

Not ideal, but not dead either.
Title: Re: Loan Information Asymmetrical Timing
Post by: Fred93 on June 18, 2014, 04:28:32 AM
I sent LC a note tonite telling them what I think about the recent API/CSV/Web changes.

Titled "Lending Club ecosystem, battles with your customers, and the API"

I've attached a copy which is identical, except that I redacted my name.

Hope you guys back me up on this.  If you agree, send them a note telling them so.  If you think I'm full of crap, send them a note telling them so.

I sent it to my normal contact, and also Renaud Lapanche (CEO).
Title: Re: Loan Information Asymmetrical Timing
Post by: hippo387 on June 18, 2014, 08:04:20 AM
I'm a retail investor who has never used API or a 3rd party service, and admittedly I know nothing about the IT infrastructure of Lending Club or Prosper. But my dollar is just as valuable as anyone else's, so here's my opinion. From a big picture perspective, any move that puts individual investors (not concerned about institutions for this point) on a common playing field is a good move in my opinion. The API and 3rd party services, by using advanced speeds and data, have frequently put the common investor like me at a disadvantage. I shouldn't need to know how to code to invest. I shouldn't need to sign up for a 3rd party service to gain greater speeds to invest. I should be able to log onto LC or Prosper directly, or use their automated investing tools, and put money to work myself. And more and more, that's what I can do. Personally I prefer Prosper and I've been finding many loans that meet my criteria each day. I don't expect them to cater to me, and I'm not sure why anyone else expects to be catered to. No 3rd party service is a part of Lending Club, so I wouldn't think they have any obligation here. 
Title: Re: Loan Information Asymmetrical Timing
Post by: Half Right on June 18, 2014, 10:22:31 AM
Just speculating, but I believe the point LC is attempting to make is that they want you to invest through their Automated Prime service or don't bother. In this way they enable you to pick the note allocation, or rate of return you are looking to achieve, while simultaneously killing off the entire "High Frequency Front-running" (so to speak) created by all the programs using the API.

 With the overwhelming majority of investors, or more importantly the minority of investors with the majority of the money, using the Automated Investing Service they can foresee the upcoming Demand and try to match it up with the Supply of new notes in an orderly manner. Although this may crimp the style of the few who have been able to cherry pick notes to-date, in the long run the new approach seems "fairer" to the "Independent Observer" and is beneficial to lending Club from an operational standpoint.

Although this may upset the forum readers , it is exactly what I would if I was running LC.
Title: Re: Loan Information Asymmetrical Timing
Post by: BruiserB on June 18, 2014, 10:35:50 AM
Just speculating, but I believe the point LC is attempting to make is that they want you to invest through their Automated Prime service or don't bother. In this way they enable you to pick the note allocation, or rate of return you are looking to achieve, while simultaneously killing off the entire "High Frequency Front-running" (so to speak) created by all the programs using the API.

 With the overwhelming majority of investors, or more importantly the minority of investors with the majority of the money, using the Automated Investing Service they can foresee the upcoming Demand and try to match it up with the Supply of new notes in an orderly manner. Although this may crimp the style of the few who have been able to cherry pick notes to-date, in the long run the new approach seems "fairer" to the "Independent Observer" and is beneficial to lending Club from an operational standpoint.

Although this may upset the forum readers , it is exactly what I would if I was running LC.

If that's the case, then they should just shut off the API and force everyone to Automated Investing.  And if they do that then I will likely look elsewhere to invest my money.

Since they haven't done that, I would like to give them the benefit of the doubt and hope that maybe they are bumbling this and their changes have had unintended consequences that they will quickly resolve.  I know nothing about the programming side of what the third parties do, but Fred93's letter makes it clear in plain English that things could be done better.

I have no problem if Lending Club wants to offer Automated Investing that is on equal footing with the access given to third party clients.  They say they are trying to promote fairness, but changes seem to be having the opposite effect.  As Fred93 says, they should really promote this ecosystem and encourage the third party clients as these bring more customers to Lending Club.  They are not "front-running" the system if they are given access equal to what Automated Investing users are getting.  It's much like when Apple first introduced the iPhone with just the factory apps.  Then people had to jailbreak and create apps.  Finally Apple introduced the AppStore and it created the ecosystem that makes iPhone hard to leave.  Lending Club should be open to the same.
Title: Re: Loan Information Asymmetrical Timing
Post by: SBryantMS on June 18, 2014, 12:44:44 PM
If you read LC Automated Investing information, they are ranking investors internally.  The more cash that you have the better your access to loans:

"The frequency of orders is based on the cash balance of your account"  -- sounds like to me they are going to rank the accounts from highest cash to lowest.   IMO, someone that is a retail investor like me will always be at the bottom of the list as I try and keep my cash invested.

As written in an earlier post, cutoff the third party tools and I will be leaving LC too.
Title: Re: Loan Information Asymmetrical Timing
Post by: Half Right on June 18, 2014, 04:21:57 PM
"Your Automated Investing account will be reviewed automatically up to four times per day to determine the level of available cash in your account, overall demand on the platform, and availability of Notes matching your investment criteria.  Automated Investing prioritizes accounts with more available cash as a percent of the overall account size and places orders for those accounts first. Automated Investing does NOT prioritize accounts based on overall account size or investment criteria."   (   http://kb.lendingclub.com/investor/articles/Investor/How-does-PRIME-determine-when-to-make-an-order-for-my-account  )

I love when you spin a document anyway you want, but if you bothered to read the entire paragraph you will see you are totally wrong.
Title: Re: Loan Information Asymmetrical Timing
Post by: rawraw on June 18, 2014, 05:50:00 PM
"Your Automated Investing account will be reviewed automatically up to four times per day to determine the level of available cash in your account, overall demand on the platform, and availability of Notes matching your investment criteria.  Automated Investing prioritizes accounts with more available cash as a percent of the overall account size and places orders for those accounts first. Automated Investing does NOT prioritize accounts based on overall account size or investment criteria."   (   http://kb.lendingclub.com/investor/articles/Investor/How-does-PRIME-determine-when-to-make-an-order-for-my-account  )

I love when you spin a document anyway you want, but if you bothered to read the entire paragraph you will see you are totally wrong.
+2

One point for the reply.  One point for the user name that compliments the reply so well.
Title: Re: Loan Information Asymmetrical Timing
Post by: lascott on June 18, 2014, 06:26:24 PM
"Your Automated Investing account will be reviewed automatically up to four times per day to determine the level of available cash in your account, overall demand on the platform, and availability of Notes matching your investment criteria.  Automated Investing prioritizes accounts with more available cash as a percent of the overall account size and places orders for those accounts first. Automated Investing does NOT prioritize accounts based on overall account size or investment criteria."   (   http://kb.lendingclub.com/investor/articles/Investor/How-does-PRIME-determine-when-to-make-an-order-for-my-account  )  I love when you spin a document anyway you want, but if you bothered to read the entire paragraph you will see you are totally wrong.

Seems like the sentences contradict each other as overall account size does matter.  A spreadsheet with some varying 'available cash' and 'account size' sorted by that percentage would be interesting to see how one would be prioritized.
Title: Loan Information Asymmetrical Timing
Post by: BruiserB on June 18, 2014, 07:28:32 PM


Seems like the sentences contradict each other as overall account size does matter.  A spreadsheet with some varying 'available cash' and 'account size' sorted by that percentage would be interesting to see how one would be prioritized.

Not really a contradiction. They are saying that someone with a $2,000 overall balance with $200 cash available (10% of account value available) will have priority over someone with a $100,000 overall balance with $5,000 cash available (5% of account value available).  If anything this seems more favorable to the smaller investor as they will have less cash waiting for investment.  It also might slightly favor people who have rather restrictive criteria.....they will buy fewer loans, but as their cash balance increases they will move to higher priority making them more likely to find loans that meet their criteria.....but any increase in returns because of this may be dragged down by having more "dead money" tied up!



Sent from my iPad using Tapatalk
Title: Re: Loan Information Asymmetrical Timing
Post by: Fred on June 18, 2014, 09:24:04 PM
With the overwhelming majority of investors, or more importantly the minority of investors with the majority of the money, using the Automated Investing Service they can foresee the upcoming Demand and try to match it up with the Supply of new notes in an orderly manner....

I don't see why LC needs to entice investors to use their Automated Investing in order to foresee upcoming demand.

LC loan demand is driven by both Automated Investing and non-Automated Investing investors.  LC has data on all loans, all notes, all investors, all borrowers.  LC has enough data to run all kinds of predictive analytics.  Having more investors with Automated Investing does not improve the predictions.
Title: Re: Loan Information Asymmetrical Timing
Post by: Fred on June 19, 2014, 02:40:05 AM
From your PDF file:

So how do I look to see how much cash is in my account? Amazingly, there is no API function to do this.

You can run the following URL:

https://www.lendingclub.com/browse/cashBalanceAj.action?rnd=<some_random_number>

to get a JSON response containing your cash balance:

Quote
{"result": "success","cashBalance": "$xxx.xx"}
Title: Re: Loan Information Asymmetrical Timing
Post by: Fred93 on June 19, 2014, 12:09:49 PM
From your PDF file:

So how do I look to see how much cash is in my account? Amazingly, there is no API function to do this.

You can run the following URL:

https://www.lendingclub.com/browse/cashBalanceAj.action?rnd=<some_random_number>

to get a JSON response containing your cash balance:

Quote
{"result": "success","cashBalance": "$xxx.xx"}

Let me know when they put that in the API document.
Title: Re: Loan Information Asymmetrical Timing
Post by: sociallender on June 19, 2014, 01:15:16 PM
Quote
You can run the following URL:

https://www.lendingclub.com/browse/cashBalanceAj.action?rnd=<some_random_number>

to get a JSON response containing your cash balance:

My understanding is you need to be logged in to perform that action which requires screen scrape.

@Fred93: Well written note to LC.  Hopefully, it will make a difference.
Title: Re: Loan Information Asymmetrical Timing
Post by: TonySaunders on June 19, 2014, 01:18:10 PM
There is some merit to attempting to keep LC attractive to new investors (giving them a chance to buy loans manually before they are gone). But in reality this just increases pressure to break the rules with automated screen-scraping. And the most dedicated will succeed.

I've already been in trouble with my screen-scraping automation and had to cut it back and use the API exclusively for fear of having my LC accounts limited. I'm not keen on taking that risk in order to be more competitive. It's really frustrating that they put us into the position of choosing between these things:

1) Be at my computer 4 times a day and try to click faster than screen-scraping automation that I'm not allowed to compete with.
2) Break the rules.
3) Settle for notes I value less.

Title: Re: Loan Information Asymmetrical Timing
Post by: TonySaunders on June 19, 2014, 01:19:30 PM
Y'know... none of us know the distribution of notes to lenders. This is something LC can examine. Maybe this policy DOES result in a more even distribution, and the frustration we feel is because we are losing an unfair advantage we had become accustomed to.
Title: Re: Loan Information Asymmetrical Timing
Post by: TonySaunders on June 19, 2014, 01:21:58 PM
The only way to really correct this problem is to remove the race condition. Everyone should be able to select notes to 'buy' with successful purchases assigned at random after like... an hour.

That's probably not as good a deal for automation people like us... since we'd have to compete with point-and-clickers, but this debate would end.
Title: Re: Loan Information Asymmetrical Timing
Post by: sisraelwi on June 19, 2014, 01:24:58 PM
From your PDF file:

So how do I look to see how much cash is in my account? Amazingly, there is no API function to do this.

You can run the following URL:

https://www.lendingclub.com/browse/cashBalanceAj.action?rnd=<some_random_number>

to get a JSON response containing your cash balance:

Quote
{"result": "success","cashBalance": "$xxx.xx"}

Is there a similar URL to run the number of notes on the platform?
Title: Re: Loan Information Asymmetrical Timing
Post by: GS on June 19, 2014, 03:13:48 PM
The only way to really correct this problem is to remove the race condition. Everyone should be able to select notes to 'buy' with successful purchases assigned at random after like... an hour.

That's probably not as good a deal for automation people like us... since we'd have to compete with point-and-clickers, but this debate would end.

I completely agree about removing the race condition.  Several months ago their was a long thread about ideas to "fix" LC.  Someone (maybe it was you) suggested that LC have a loan "pre-order" period (perhaps the 4 hour gap between "drops"), where lenders could line up pre-orders to be executed, with priority, at drop times .... if orders outweighed the availability, then orders would be filled based some logical priority, such as users who have gone the longest without having an order filled... I thought that was the most brilliant idea on the thread, but I doubt LC would implement anything that actually worked to solve the problem.
Title: Re: Loan Information Asymmetrical Timing
Post by: RaymondG on June 19, 2014, 04:13:00 PM
https://www.lendingclub.com/browse/cashBalanceAj.action?rnd=<some_random_number>

to get a JSON response containing your cash balance:
Quote
{"result": "success","cashBalance": "$xxx.xx"}

Is there a similar URL to run the number of notes on the platform?

https://www.lendingclub.com/browse/browseNotesAj.action?method=getResultsInitial&startindex=0&pagesize=1&r=14646
(the "&r=14646" must be a random number.)
At the end of of the content, read "totalRecords":980
Title: Re: Loan Information Asymmetrical Timing
Post by: Fred on June 20, 2014, 02:29:37 AM
My understanding is you need to be logged in to perform that action which requires screen scrape.

You can login to LC by simply sending a POST to https://www.lendingclub.com/account/login.action along with username & password as parameters.
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on June 20, 2014, 01:37:15 PM
I understand there are some upcoming discussions at LC about this situation. I suspect some time will need to elapse before there's a resolution and communication.

I'm sticking with it and hoping that it amounted to unintended consequences and that we'll see a more nuanced approach to loan loads in the future. I particularly like the gentleman's suggestion of allowing loans to be over-filled and then randomizing the fractions a couple hours after they hit the site. This would eliminate speed plays. It would also provide LC very good information about which loans may be priced incorrectly, which they could in turn use to improve their operations.

Title: Re: Loan Information Asymmetrical Timing
Post by: eastbiz on June 20, 2014, 11:57:23 PM
I sent LC a note tonite telling them what I think about the recent API/CSV/Web changes.

Titled "Lending Club ecosystem, battles with your customers, and the API"

I've attached a copy which is identical, except that I redacted my name.

Hope you guys back me up on this.  If you agree, send them a note telling them so.  If you think I'm full of crap, send them a note telling them so.

I sent it to my normal contact, and also Renaud Lapanche (CEO).

Very well written Fred93!!
Title: Re: Loan Information Asymmetrical Timing
Post by: rawraw on June 21, 2014, 07:33:06 AM
I sent LC a note tonite telling them what I think about the recent API/CSV/Web changes.

Titled "Lending Club ecosystem, battles with your customers, and the API"

I've attached a copy which is identical, except that I redacted my name.

Hope you guys back me up on this.  If you agree, send them a note telling them so.  If you think I'm full of crap, send them a note telling them so.

I sent it to my normal contact, and also Renaud Lapanche (CEO).
Have they replied?
Title: Re: Loan Information Asymmetrical Timing
Post by: lascott on June 21, 2014, 09:42:02 AM
I particularly like the gentleman's suggestion of allowing loans to be over-filled and then randomizing the fractions a couple hours after they hit the site. This would eliminate speed plays.
It could be useful if they figured out a percentage to do for the small funding to spread it out to more investors too.  1) Randomize for all investors investing $50 or less and satisfiy 70% of the loan.  2) Then randomize for all investors $51-$100 for 10%.  3) Then randomize for all investors $101-$200 for 10%. Then any amount. If some left, then back to 1).  It could be something more complicated than that but it illustrates my point.

If you do the quick match on size of loan / number of investor on many loans people are investing way more than $25 per loan so that limits the number of satisfied investors.  Maximize the number of investors.  As well this leads to a better experience as defaults hurt less for $25-$50 investment amounts as has been shown in many examples.

As I recall they were limiting the institutions to a certain percentage already.  This is a little like that but with subsets for the retail investors.
Title: Re: Loan Information Asymmetrical Timing
Post by: Rob L on June 21, 2014, 01:09:56 PM
The only way to really correct this problem is to remove the race condition.
To Lending Club this may not be problem at all. The changes may have had precisely the desired effect.

Title: Re: Loan Information Asymmetrical Timing
Post by: Fred93 on June 21, 2014, 05:19:08 PM
I sent LC a note tonite telling them what I think about the recent API/CSV/Web changes.
Have they replied?

My rep is arranging a phone call for Monday.   Don't know who will be on the call from the LC end.
Title: Re: Loan Information Asymmetrical Timing
Post by: lascott on June 21, 2014, 10:09:36 PM
I particularly like the gentleman's suggestion of allowing loans to be over-filled and then randomizing the fractions a couple hours after they hit the site. This would eliminate speed plays.
It could be useful if they figured out a percentage to do for the small funding to spread it out to more investors too.  1) Randomize for all investors investing $50 or less and satisfiy 70% of the loan.  2) Then randomize for all investors $51-$100 for 10%.  3) Then randomize for all investors $101-$200 for 10%. Then any amount. If some left, then back to 1).  It could be something more complicated than that but it illustrates my point.

If you do the quick match on size of loan / number of investor on many loans people are investing way more than $25 per loan so that limits the number of satisfied investors.  Maximize the number of investors.  As well this leads to a better experience as defaults hurt less for $25-$50 investment amounts as has been shown in many examples.

As I recall they were limiting the institutions to a certain percentage already.  This is a little like that but with subsets for the retail investors.
Just started looking at Prosper in hopes P2P-Picks ends up finding a partner.

Ran across this article related to limiting large investors ... sort like above point I was making.

http://www.orchardplatform.com/blog/availability-of-fractional-loans
Quote
In early February [2014], Prosper announced that it had asked investors to limit their investment in fractional listings to a maximum of 10% of any loanís requested amount.  The previous limit had been 50%.  This move was part of an effort to rebalance the marketplace on behalf of smaller retail investors, many of whom had begun to feel crowded out by larger investors who would snap up 50% of a listing immediately after it became available.  In todayís post, we examine the data to determine what effect the changes have had on the market for fractional loans.
Title: Loan Information Asymmetrical Timing
Post by: BruiserB on June 23, 2014, 06:59:51 PM

I sent LC a note tonite telling them what I think about the recent API/CSV/Web changes.
Have they replied?

My rep is arranging a phone call for Monday.   Don't know who will be on the call from the LC end.

Were you able to speak with anyone today?  Even if you can't go into detail of all discussed, it would be great to know if they are listening and understanding the issues caused by their recent changes.


Sent from my iPad using Tapatalk
Title: Re: Loan Information Asymmetrical Timing
Post by: sociallender on June 23, 2014, 07:51:00 PM
Received word today that LC will be updating the API at the same time the web site is updated.  No estimate of when the change will take place.  Not sure about the CSV.

SL
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on June 23, 2014, 09:27:29 PM
I hope the CSV just becomes part of the API. We should be able to request the API to reply in XML, JSON, or CSV. Glad to hear that they will synch again soon.
Title: Re: Loan Information Asymmetrical Timing
Post by: Fred93 on June 23, 2014, 11:44:56 PM
Had a nice call today with my regular contact + one program manager + 2 senior technical guys. 

They had read my writeup, and seemed to understand it.  I got no commitments of any kind from them.

They verified that the CSV is updated once per minute.  They say that the API is same deal, although at one point they said it was on "a different schedule". 

Regarding the difficulty of polling when the CSV is updated asynchronously, they suggested the following polling method:

Use the API.  Poll near the feeding times.  To poll, call the API BrowseLoans(false) method, which retrieves only "new" loans.  (New means most recent batch, even tho we don't know up front which batch that is.)  After downloading the list, look at the ListD field of each loan to see when these loans were listed.  Count the number of loans with ListD within the last 10 minutes.  That's the algorithm they say they use internally.  When a chunk of loans appear with recent ListD times, you have new loans.  As I understand it, one can't simply look at say the first loan's listD and know whether the loans are fresh, because a new batch will contain some loans that are old, for example loans that were released to the whole-loan buyers 12 hours ago, and are just now released to the fractional buyers.  However there will generally be some new loans, so you'll see those new listing dates.

They admit this is not excellent, but it is their recommended method.

I explained that the people use the CSV because it is so much faster than the API.  This was news to them, and caused some confusion.  Frankly they didn't believe me.  This evening I did a test, and forwarded to them the results.   In my tests, the API is 40x slower than the CSV.

I stressed my suggestion that they engage the API user community and discuss these issues, proposed changes, etc.  I make the same suggestion to you guys.  Call 'em and tell 'em what you think. 


Sent the following message to my contact a few minutes ago.

Quote
During the call today, the LC guys were incredulous when I said people use the CSV file rather than the API Browse Loans method because the CSV file is so much faster.  Frankly, they had difficulty believing me.  Please forward this to the group who was on the phone today.

I ran a little test tonite.  All I did was repeatedly download the same information by these two methods, while timing them.  Here are the results.

06/23/2014 19:23:40 API BrowseLoans(true) took 22286 ms
06/23/2014 19:24:03 CSV download took 583 ms

06/23/2014 19:24:03 API BrowseLoans(true) took 22222 ms
06/23/2014 19:24:26 CSV download took 515 ms

06/23/2014 19:24:26 API BrowseLoans(true) took 22079 ms
06/23/2014 19:24:48 CSV download took 490 ms

06/23/2014 19:24:49 API BrowseLoans(true) took 22980 ms
06/23/2014 19:25:12 CSV download took 505 ms

06/23/2014 19:25:12 API BrowseLoans(true) took 22094 ms
06/23/2014 19:25:34 CSV download took 478 ms

As you can see, the API appears to be slower by about 40x most of the time.  This is a huge difference.  There can be no doubt: The CSV file is much faster than the API!  This explains why your users prefer the CSV file to the API.

You can also see why I laughed when one of your guys said I should not execute the BrowseLoans function more often than once per second.  It takes 22 seconds to run!

I have a theory.  My theory is that the HTTP server used for your API is misconfigured, and not set to compress traffic.  My theory is that the time we observe is the time it takes the bits to move thru a limit-speed connection.  The API uses XML which is much more verbose than the CSV format.  This should be relieved considerably if you compress.

The above test was done from my home in San Diego, on a 30 Mbps cable modem.  You can't really expect most users to have connections any faster than that.

 I then copied fhe test program to my desktop at work, where we have a 1 Gbps Time Warner connection.  (I verified that the path to LC goes thru the TW connection)   I can't verify that I really had access to 1 Gbps of course, as there are shared 1 Gbps links within buildings etc.  Rerunning the test at work, I got these results...

6/23/2014 7:56:42 PM API BrowseLoans(true) took 4211 ms
6/23/2014 7:56:46 PM CSV download took 542 ms

6/23/2014 7:56:46 PM API BrowseLoans(true) took 1789 ms
6/23/2014 7:56:48 PM CSV download took 190 ms

6/23/2014 7:56:48 PM API BrowseLoans(true) took 6239 ms
6/23/2014 7:56:55 PM CSV download took 182 ms

6/23/2014 7:56:55 PM API BrowseLoans(true) took 6170 ms
6/23/2014 7:57:01 PM CSV download took 205 ms

6/23/2014 7:57:01 PM API BrowseLoans(true) took 7142 ms
6/23/2014 7:57:08 PM CSV download took 192 ms

As expected, the same program runs considerably faster here.  The API is still much much slower than the CSV file.

In any case, I hope this gives you some insight into why I told you today that the API is much slower than the CSV file.

Title: Re: Loan Information Asymmetrical Timing
Post by: Fred93 on June 23, 2014, 11:48:05 PM
Received word today that LC will be updating the API at the same time the web site is updated.  No estimate of when the change will take place.  Not sure about the CSV.

Is not what they told me.  I talked with them at noon. 
Title: Re: Loan Information Asymmetrical Timing
Post by: Fred on June 24, 2014, 12:56:47 AM
Had a nice call today with my regular contact + one program manager + 2 senior technical guys. 

Thanks for sharing.  It's nice to see some glimpses of LC inner workings.
Title: Re: Loan Information Asymmetrical Timing
Post by: Rob L on June 24, 2014, 10:14:21 AM
Lots of interesting information @Fred93; thanks.
Wonder where we go from here (if anywhere)?
Title: Re: Loan Information Asymmetrical Timing
Post by: sociallender on June 24, 2014, 11:00:52 AM
Quote
Is not what they told me.  I talked with them at noon.

I have a feeling now that maybe the rep that called me didn't have all the facts?  She did say that she wanted to dig deeper into my concerns which is why she took so long to get back to me.  Hmmmm...  seems like they need to make a public announcement so everyone is clear in their communication.
Title: Re: Loan Information Asymmetrical Timing
Post by: lascott on June 25, 2014, 01:41:31 AM
I just opened a Prosper account in hopes of P2P-Picks and some 3rd party note investing tool.  Related to this I started looking at NickleSteamroller and their blog.

Thought this was interesting. I realize it is over a year old but there are likely new folks like me how never have seen it.  Just a comparison to LendingClub's API.

Title: Prosperís new RESTful API
http://www.nickelsteamroller.com/blog/2013/03/prospers-new-restful-api/
Title: Re: Loan Information Asymmetrical Timing
Post by: hoggy1 on June 25, 2014, 10:26:06 AM
Regarding the shopping cart lock-out, I have tried this technique with little success because I have not found a mechanism to selectively delete loans that I subsequently decide not to purchase. I have to either place the order in full or abandon it. What am I missing?
Title: Re: Loan Information Asymmetrical Timing
Post by: Rob L on June 25, 2014, 10:36:35 AM
When "Viewing Loans", set the investment amount to zero for the loan to delete then hit enter.
Title: Re: Loan Information Asymmetrical Timing
Post by: Randawl on June 25, 2014, 12:07:30 PM
Regarding the shopping cart lock-out, I have tried this technique with little success because I have not found a mechanism to selectively delete loans that I subsequently decide not to purchase. I have to either place the order in full or abandon it. What am I missing?
When "Viewing Loans", set the investment amount to zero for the loan to delete then hit enter.

There is also a small trash can icon that appears to the right of each listing when you mouse-over it.
Title: Re: Loan Information Asymmetrical Timing
Post by: hoggy1 on June 25, 2014, 12:28:17 PM
Thank you both. Never saw that little trash can and never thought to set the amount to zero. They don't call us newbies for nothing.
Title: Re: Loan Information Asymmetrical Timing
Post by: Rob L on June 25, 2014, 12:57:34 PM
I ran a little test tonite.  All I did was repeatedly download the same information by these two methods, while timing them.  Here are the results.

06/23/2014 19:23:40 API BrowseLoans(true) took 22286 ms
06/23/2014 19:24:03 CSV download took 583 ms

I then copied fhe test program to my desktop at work, where we have a 1 Gbps Time Warner connection.  (I verified that the path to LC goes thru the TW connection)   I can't verify that I really had access to 1 Gbps of course, as there are shared 1 Gbps links within buildings etc.  Rerunning the test at work, I got these results...

6/23/2014 7:56:42 PM API BrowseLoans(true) took 4211 ms
6/23/2014 7:56:46 PM CSV download took 542 ms

As expected, the same program runs considerably faster here.  The API is still much much slower than the CSV file.


Actually I find this very surprising (not the API / CSV delta, but the home vs work one).
I would have thought the dominant download time factor would be the LC server response time, not the network speed.
Title: Re: Loan Information Asymmetrical Timing
Post by: sociallender on June 25, 2014, 03:52:24 PM
Quote
Actually I find this very surprising (not the API / CSV delta, but the home vs work one).
I would have thought the dominant download time factor would be the LC server response time, not the network speed.

With my tests of CSV vs API speed, I found the CSV to be considerably faster and with consistent response times versus the API.  The API ranged in response times from 2x to 30x.  Only tested on single network.  The other thing that I thought was odd was that the API seemed to be faster if I hadn't used it in a while.  I am wondering if there is any metering involved?  Could be a coincidence tho.   

Back to a previous question on this thread, does anyone know if the LC auto invest has the same delay as CSV or API?  Is there preferential treatment to auto invest?  It has been a while since I have looked at auto invest from LC but I thought they used to only auto investe periodically (If I recall, once a day).  However, from the latest communication from LC, it appears they auto invest as soon as the notes are released now.  Can anyone confirm?  If so, this seems like a significant advantage to using their platform for those seeking to purchase notes using their provided filters. 
Title: Re: Loan Information Asymmetrical Timing
Post by: TonySaunders on June 25, 2014, 05:08:09 PM

...

Back to a previous question on this thread, does anyone know if the LC auto invest has the same delay as CSV or API?  Is there preferential treatment to auto invest?  It has been a while since I have looked at auto invest from LC but I thought they used to only auto investe periodically (If I recall, once a day).  However, from the latest communication from LC, it appears they auto invest as soon as the notes are released now.  Can anyone confirm?  If so, this seems like a significant advantage to using their platform for those seeking to purchase notes using their provided filters.

That's quite interesting. I'm going to begin experimenting with it to compare with my automation and see what I can see.
Title: Re: Loan Information Asymmetrical Timing
Post by: Fred93 on June 25, 2014, 06:03:38 PM
Actually I find this very surprising (not the API / CSV delta, but the home vs work one).
I would have thought the dominant download time factor would be the LC server response time, not the network speed.

Agreed.  It is very surprising, and at this point doesn't make sense.  The amount of data transferred is so much smaller than the time it takes ... doesn't come anywhere close to using up my network bandwidth allocation ... and yet it goes faster when I measured on a faster connection.

There is something that we don't yet realize causing this.  I also wonder about some metering at the Lending Club server.  Don't have enough evidence yet.  I haven't put any work into this in the last day.  Might get back on it in the next few days.

This is a repeating pattern ... things are not as they appear.

Regarding the notion that it is occasionally faster, and gives one the impression that it is faster if you haven't used it for awhile, I have noticed that too.  I don't have objective evidence tho.  Just an impression.  Note that cable modems also do a thing where they don't clamp down their metering right away.  Time Warner calls this a feature, and has named it "speed boost".  It is designed to show great numbers for short file transfers, such as those you do with those "speed test" web sites.
Title: Re: Loan Information Asymmetrical Timing
Post by: GS on June 25, 2014, 08:25:10 PM
There was a NY Time article a while back that talked about LC putting "governors" on the high speed buyers... Not sure if it's true or not.

Quote
The fastest computer right now is getting the most loans,” says Peter Renton ...

Prosper and Lending Club have created speed limits, known as governors, to counter these moves, and they have instituted purchase limits to ensure that big buyers don’t hog all the loans. “It’s kind of an arms race,” Mr. Kassul says. “They put a governor on, but then everyone tries to trick the governor.”

http://www.nytimes.com/2014/05/04/business/loans-that-avoid-banks-maybe-not.html?_r=0
Title: Re: Loan Information Asymmetrical Timing
Post by: Fred93 on June 25, 2014, 08:56:29 PM
With my tests of CSV vs API speed, I found the CSV to be considerably faster and with consistent response times versus the API.  The API ranged in response times from 2x to 30x. 

It is incredibly helpful to me to have someone else report results similar to mine.  Gives me confidence that I'm not going nuts or making some stupid error.  (Both are still possible -- just less likely.)
Title: Re: Loan Information Asymmetrical Timing
Post by: hoggy1 on June 26, 2014, 01:23:52 PM

...

Back to a previous question on this thread, does anyone know if the LC auto invest has the same delay as CSV or API?  Is there preferential treatment to auto invest?  It has been a while since I have looked at auto invest from LC but I thought they used to only auto investe periodically (If I recall, once a day).  However, from the latest communication from LC, it appears they auto invest as soon as the notes are released now.  Can anyone confirm?  If so, this seems like a significant advantage to using their platform for those seeking to purchase notes using their provided filters.

That's quite interesting. I'm going to begin experimenting with it to compare with my automation and see what I can see.

I would think Anil Gupta from Peercube should be able to confirm? He is studying time-to-fund in another thread (topic=2379.15). Quoting him

 "PeerCube captures loan listings using API every minute at release times (-5 to +20 minutes) and then every few minutes rest of the day and night. This TTF analysis includes only the loan listings that were detected by PeerCube and the loans issued out of detected loan listings."

My guess would be the loans that were missed all together in their first look and those with high funded ratios were likely LC auto-invest early access? He might correct me.
Title: Re: Loan Information Asymmetrical Timing
Post by: Fred93 on July 03, 2014, 06:08:01 AM
With my tests of CSV vs API speed, I found the CSV to be considerably faster and with consistent response times versus the API.  The API ranged in response times from 2x to 30x.

Yep.  Apparently the LC team took my comments (and data I sent them) about API speed problems to heart, and made some adjustments.   A few days ago I got a brief email that said simply "try it now".  The 10x, 20x, 30x slower numbers appear to be gone, and now the API BrowseLoans function responds with similar speed as the CSV.  (API is maybe as much as 1.5x the CSV time in my recent tests.)  Still seems more variable than it should be, and not nearly as fast as it could be, so I believe more gain is possible.

Asked them what they changed, but so far they've been silent. 

Now loanBrowseLoans(all) takes me on the order of 1 second, instead of the 10 to 40 seconds I had seen in the past.  I've given them an argument for why it could be 200ms, (because its about 1.5 Megabytes of data, which should take 400ms to arrive over my 30Mbps cable modem service, and if they would turn on compression in their http server, the size would be cut in half, which computes to 200ms) but don't have any evidence that I managed to engage them on these points.

Meanwhile, I can't argue with the improvement so far.

Did get a brief comment today that they are working on some of the API deficiencies described in my memo.   No specifics.

Overall very positive, tho still a bit mysterious.
Title: Re: Loan Information Asymmetrical Timing
Post by: rawraw on July 03, 2014, 07:07:26 AM
I'm glad to see us actually talking to LC about issues now.  Hopefully it'll keep things moving forward
Title: Re: Loan Information Asymmetrical Timing
Post by: hoggy1 on July 03, 2014, 08:22:40 AM
Special thanks are owed to Fred of course but this is a great collective piece of work by many/most of the active board members. I think it unlikely Fred would be making progress were it not the other participants on this board. As a newbie I am very impressed and grateful to you all.
Title: Re: Loan Information Asymmetrical Timing
Post by: Rob L on July 03, 2014, 10:36:05 AM
... if they would turn on compression in their http server, the size would be cut in half, which computes to 200ms) but don't have any evidence that I managed to engage them on these points.
I think they turned on gzip compression of loanBrowseloans responses last night at the 9pm release.
Title: Re: Loan Information Asymmetrical Timing
Post by: kbenson99 on July 03, 2014, 01:37:56 PM
Nice!  Gzip compression has been enabled on the LoanBrowseLoans operation.  To get the LoanBrowseLoans operation to return the compressed loan data, I added 'Accept-Encoding': 'gzip' to my header.  Simple.
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on July 03, 2014, 04:26:17 PM
Good improvements. Now for the big kahuna. Is it synched with the website loan release?
Title: Re: Loan Information Asymmetrical Timing
Post by: Rob L on July 03, 2014, 06:34:51 PM
Sadly no, not at this time.
Loans are consistently available on the web site at 00:40.
The API and CSV are essentially simultaneous, as closely as I can tell, about 30 seconds later, at about 01:10. Perhaps a very slight edge to the API.
The API is updated every 30 seconds and the csv is updated once per minute.

That said, the kahuna kahuna is what happens during that 30 seconds head start.
Considerable improvements have been made by LC over the past week or two. Not as bad as it was, but not great.
Now that LC is giving us a peek every 30 seconds I can unequivocally say there is no feeding frenzy, mad rush, etc. at loan grades C5 and better.
The data are unambiguous. If you're using automation and C5's or better are your thing then that 30 second edge is a non issue.

Yeah, well so what. D's and up are what I want.
Well, that's basically me; so what have I been doing. Same thing as a lot of others I'm sure.
I have a simple stored LC filter that says all new D's and E's only please. With a little cash that ain't hard...
Last release at 2pm PDT I manually put 23 D's and E's into my cart. I bought 4 and pitched out the rest.
That 30 seconds is such an advantage I doubt I've missed putting any desired loans into my cart since being forced to be one of Pavlov's dogs (did I hear a bell? I'm hungry)
The good news (really good news) is that my chips are $25's and $50's.
From what I've seen recently folks using the cart are pretty much like me; the institutional money isn't taking advantage of the 30 second cart advantage at this time.

What do things look like at the time new loans are posted for the wonderful world of automation?
Many D's and higher are 25% or so invested. Lots thrown out of the carts later (just as I threw out 19 out of 23).
Meanwhile, you better be quick. I'm not making any quality judgements (Anil); you just can't buy what's "Already Gone" ( Eagles fan here).
Lots of heavy buying in the 30 seconds immediately after the loans hit loanBrowseloans or browsenotes.csv (or whatever it's named now).

For now, automation seems to have a shot at 75% or better of the D's and higher if it's very quick.
Meanwhile there are clever folks that will be compelled to insert automation into the 30 second shopping cart advantage.
No doubt it's being done on a small (individual) scale now.
Frankly, if it persists, I'll go there too. I would have done it months ago, but I've been able to get just about all the loans I wanted anyway so I didn't bother.
Then LC will find a way to kill it, since that can't be what they want. And the beat goes on ...









Title: Re: Loan Information Asymmetrical Timing
Post by: Rob L on July 11, 2014, 09:18:59 AM
The release of loans to the web site (browsenotes), the API and the CSV was simultaneous this morning (about 6:01 am PT 7/11/2014).
Perhaps a one-of-a-kind event. I've been out of town a few days so this could have been happening for a while.
Title: Re: Loan Information Asymmetrical Timing
Post by: mchu168 on July 11, 2014, 09:34:47 AM
Yes, but BlueInvestment Auto Invest appears to be down.  Anybody else having issues?
Title: Re: Loan Information Asymmetrical Timing
Post by: rawraw on July 11, 2014, 09:39:32 AM
The release of loans to the web site (browsenotes), the API and the CSV was simultaneous this morning (about 6:01 am PT 7/11/2014).
Perhaps a one-of-a-kind event. I've been out of town a few days so this could have been happening for a while.
I'm getting a little excited.  Nominations for Fred93 as the MVP (Most Valuable Poster) of the quarter!
Title: Re: Loan Information Asymmetrical Timing
Post by: lascott on July 11, 2014, 11:23:08 AM
Don't know if Anil's/peercube's times are for the API/CSV?
(https://forum.lendacademy.com/proxy.php?request=http%3A%2F%2Fi.imgur.com%2FpOK9079.png&hash=a66f9c243d236c4e4200f0176c5f59c3)
Title: Re: Loan Information Asymmetrical Timing
Post by: Rob L on July 11, 2014, 01:11:28 PM
Nice while it lasted (1 release).
The asymmetry is back; manual buying had a 30 second advantage at the 10am PT release.
Title: Re: Loan Information Asymmetrical Timing
Post by: AnilG on July 11, 2014, 02:02:56 PM
PeerCube tweet times are for API. PeerCube checks for available loans every minute. It doesn't care about little difference in seconds.

Code: [Select]
2014-07-11 10:01:07 --> LC Loans in Database = 1428
2014-07-11 10:01:07 --> LC Loans in Update = 1428
2014-07-11 10:01:07 --> LC Loans funded since last Update = 0
2014-07-11 10:01:07 --> LC Old Loans in Update = 1428
2014-07-11 10:01:07 --> LC New Loans in Update = 0

2014-07-11 10:02:12 --> LC Loans in Database = 1428
2014-07-11 10:02:12 --> LC Loans in Update = 1536
2014-07-11 10:02:12 --> LC Loans funded since last Update = 0
2014-07-11 10:02:12 --> LC Old Loans in Update = 1428
2014-07-11 10:02:12 --> LC New Loans in Update = 108

2014-07-11 10:03:26 --> LC Loans in Database = 1536
2014-07-11 10:03:26 --> LC Loans in Update = 1539
2014-07-11 10:03:26 --> LC Loans funded since last Update = 0
2014-07-11 10:03:26 --> LC Old Loans in Update = 1536
2014-07-11 10:03:26 --> LC New Loans in Update = 3

2014-07-11 10:04:39 --> LC Loans in Database = 1539
2014-07-11 10:04:39 --> LC Loans in Update = 1534
2014-07-11 10:04:39 --> LC Loans funded since last Update = 5
2014-07-11 10:04:39 --> LC Old Loans in Update = 1534
2014-07-11 10:04:39 --> LC New Loans in Update = 0

Don't know if Anil's/peercube's times are for the API/CSV?
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on July 30, 2014, 09:05:29 PM
And it's back. The website gets the loans earlier than the API/CSV.
Title: Re: Loan Information Asymmetrical Timing
Post by: RaymondG on August 23, 2014, 08:29:08 AM
It became worse starting at 08/21 20:59:59PM EST. The new loans were seen on LC website before the loan-releasing hour, such as 20:59:59 EST. But the CSV file for downloading was not updated until 08/21 21:00:58PM EST. That's one minute gap. It used to be about 30 seconds. If this become standard from now on, I have to re-write my code :(

From my log, it clearly show the following pattern:
* new loans appear on LC website seconds before [Release Hour, say HH:00:00]
* LC snapshot the loan list at about HH:00:45
* The snapshot is able to be downloaded as CSV file after HH:00:55

At the time new loans shown in CSV file that is one minute later than shown on website, quality loans are almost all gone.
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on August 23, 2014, 09:32:21 AM
It's all over the place day by day. I had to adjust my scripts to use the API and convert to CSV because now the API is 30-40 seconds past the hour. Web still may be faster.
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on August 23, 2014, 01:08:51 PM
Yep, still pretty intolerable:

10:00:00 - Web delivery
10:00:32 - API delivery
10:01:01 - CSV delivery
Title: Re: Loan Information Asymmetrical Timing
Post by: cfb on August 23, 2014, 03:40:17 PM
As I mentioned in a separate post, my use of prime/auto invest isn't giving me any better access to notes at feeding time.  I'm just getting the stuff that passes my filters but nobody else wants.  By doing the manual mad scramble I could usually get 4-8 notes a week and half or more didn't issue.  The autoinvest bought four and two were withdrawn.

I can also tell you exactly how a large/institutional vendor could have an advantage or use the delay to get notes.  Have a huge cash balance, wait until the drop, screen scrape to add every note to your cart, wait a minute for the csv data drop, download it, analyze it, drop everything that you don't like from the cart and buy.

All you'd need is enough in your account to cover your $25/$50/$500/whatever target per note times the max number of notes that drop.
Title: Re: Loan Information Asymmetrical Timing
Post by: Fred on August 23, 2014, 04:08:18 PM
I can also tell you exactly how a large/institutional vendor could have an advantage or use the delay to get notes.  Have a huge cash balance, ....

This does not have to be "huge",  just enough to temporarily hold ~200 notes (e.g., $5000 for $25/note).
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on August 23, 2014, 04:26:43 PM
...Have a huge cash balance, wait until the drop, screen scrape to add every note to your cart, wait a minute for the csv data drop, download it, analyze it, drop everything that you don't like from the cart and buy.

Notes that get fully funded don't appear in the CSV. I'm unaware of any way to get the extra credit variables for custom scoring. I've tried. Fractional market such a technique might work well, if there's at least $25 left by the time the API drops. Whole loan market is another story; cart the whole loans and you aren't getting the credit variables. Don't cart the whole loans and you aren't getting the loans. FML.
Title: Re: Loan Information Asymmetrical Timing
Post by: RaymondG on August 23, 2014, 06:02:42 PM
...Have a huge cash balance, wait until the drop, screen scrape to add every note to your cart, wait a minute for the csv data drop, download it, analyze it, drop everything that you don't like from the cart and buy.

Notes that get fully funded don't appear in the CSV. I'm unaware of any way to get the extra credit variables for custom scoring. I've tried. Fractional market such a technique might work well, if there's at least $25 left by the time the API drops. Whole loan market is another story; cart the whole loans and you aren't getting the credit variables. Don't cart the whole loans and you aren't getting the loans. FML.

Agree. The big problem is CSV snapshot starts at about 00:45, at which time many loans are already fully funded and they will not be in CSV download.
Title: Re: Loan Information Asymmetrical Timing
Post by: cfb on August 24, 2014, 03:04:35 PM
I can also tell you exactly how a large/institutional vendor could have an advantage or use the delay to get notes.  Have a huge cash balance, ....

This does not have to be "huge",  just enough to temporarily hold ~200 notes (e.g., $5000 for $25/note).

I was figuring the large/institutional investors were going for a lot more than $25/note or they'd never get capital deployed.  Seems like a lot of people are just fiddling with a few thousand in p2p, so $5K would be a lot for them to leave in an account uninvested.
Title: Re: Loan Information Asymmetrical Timing
Post by: rawraw on August 24, 2014, 05:24:28 PM
Yea I don't quite get the cart method either as described here.  If they only reserve $25 per loan, then they can only invest $25.  If they want to reserve larger blocks, well then they have to keep a much larger cash balance earning 0%.  I'm not quite sure I see how exactly this would work for an institutional investor.
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on August 24, 2014, 09:25:52 PM
Some institutional investors buy on the fractional platform as well. $500 or $1000 deployment can go pretty quick.
Title: Re: Loan Information Asymmetrical Timing
Post by: turing on August 25, 2014, 04:04:33 PM
After reading all the "Mad Rush" articles on PeerCube (https://www.peercube.com/blog/post/22), I've wondered why so many people are concerned about seconds and minutes.

Overall, there is no better return on the really fast to fund loans as their is with the slower ones.

I realize the P2P-Picks notes are probably gone really fast because there are quite a few people using that criteria, but in general this rat race doesn't make much sense to me.
Title: Re: Loan Information Asymmetrical Timing
Post by: Fred93 on August 25, 2014, 04:25:07 PM
After reading all the "Mad Rush" articles on PeerCube (https://www.peercube.com/blog/post/22), I've wondered why so many people are concerned about seconds and minutes.

Overall, there is no better return on the really fast to fund loans as their is with the slower ones.

Well, two things about that.

First, you have to trust his math.

Second, you must realize that he only showed that on the average one set of loans was no better than another.  That doesn't mean that the loans YOU pick will be average.  He certainly didn't prove that it is worthless to pick early.

There are some people who pick early but not wisely.  That doesn't mean I shouldn't pick early.
Title: Re: Loan Information Asymmetrical Timing
Post by: turing on August 25, 2014, 06:14:11 PM

He certainly didn't prove that it is worthless to pick early.

There are some people who pick early but not wisely.  That doesn't mean I shouldn't pick early.

There are definitely lots of good loans early, but his data also show that there are tons of good loans late as well.

In my opinion, it doesn't make sense to be involved in the timing rat race to get good loans if there are lots of good loans available later on too. 

If high frequency stock trading has taught me anything, speed is a very expensive game to play (tunnels underground to shave 3 ms off trading times (http://www.forbes.com/forbes/2010/0927/outfront-netscape-jim-barksdale-daniel-spivey-wall-street-speed-war.html))
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on August 25, 2014, 08:30:21 PM
I don't think Anil's analysis is convincing. Plenty of stupid people and entities move quickly.
Title: Re: Loan Information Asymmetrical Timing
Post by: DanB on August 25, 2014, 10:56:37 PM
I don't think Anil's analysis is convincing. Plenty of stupid people and entities move quickly.

Well then what would have to occur in order for it to be convincing to you??
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on August 26, 2014, 01:04:25 AM
One would have to prove that there is no subset of loans that, ex ante, would be expected to outperform a comparable index.
Title: Re: Loan Information Asymmetrical Timing
Post by: Fred on August 26, 2014, 02:48:33 AM
One would have to prove that there is no subset of loans that, ex ante, would be expected to outperform a comparable index.

Proving a non-existence would be very hard. ;-)

However, I am still convinced that speed matters in P2P lending platforms, just as it does in equities & commodities platforms.
Title: Re: Loan Information Asymmetrical Timing
Post by: rawraw on August 26, 2014, 11:13:27 AM
One would have to prove that there is no subset of loans that, ex ante, would be expected to outperform a comparable index.

Proving a non-existence would be very hard. ;-)

However, I am still convinced that speed matters in P2P lending platforms, just as it does in equities & commodities platforms.
I think it matters based on the current situation (4 load times, a lot of demand).  I think some structural changes (plus LC improving their score cards) could reduce the need for speed.  I still would be curious if Bryce graded all Anil's data and then we could have a more informed analysis on speed and if it matters.
Title: Re: Loan Information Asymmetrical Timing
Post by: cfb on August 26, 2014, 04:18:42 PM
One would have to prove that there is no subset of loans that, ex ante, would be expected to outperform a comparable index.

And do it for a long time.

As I've mentioned elsewhere, I think that 60-80% of defaults have to do with external, unknowable life events and not so much with people who just decide to not pay or scammmers.  Divorce, death, serious injury, sudden unexpected job loss and long unemployment/underemployment, other financial catastrophy forcing bankruptcy, etc.  You can't get those off any credit analysis or crystal ball.

So I think that you can improve your returns by a bit by choosing well, but I've had just as many blue chippy looking borrowers roll over on me as iffy looking ones.  What really sucks is that to get a lot of capital deployed I was pushing $150-500 into the better looking higher graded loans and enough of those defaulted to screw up my returns a little.  Still, I don't know where I could have made 10%+ anywhere else except equities or direct real estate investment over the last 5 years, and I have plenty of both of those going on already.
Title: Re: Loan Information Asymmetrical Timing
Post by: hoggy1 on August 26, 2014, 06:29:26 PM
As I've mentioned elsewhere, I think that 60-80% of defaults have to do with external, unknowable life events and not so much with people who just decide to not pay or scammmers.  Divorce, death, serious injury, sudden unexpected job loss and long unemployment/underemployment, other financial catastrophy forcing bankruptcy, etc.  You can't get those off any credit analysis or crystal ball.

I been concerned about this for a long time. If you are correct that the number are in the 60-80% range most of our analysis is almost useless as you describe. Those type of events will not distribute themselves evenly over incomes (higher income borrowers will weather a minor emergency better than someone living on the edge) but are likely to distribute themselves more evenly over FICOs (maybe - first thing to come to mind) or other discriminators. Let me put my statistician hat on and consult with a few colleagues to see if a better estimate can be calculated. And I hope the answer is not as high as you suggest or I'm wasting my time.
Title: Re: Loan Information Asymmetrical Timing
Post by: DanB on August 26, 2014, 06:58:50 PM
I would contend that "speed" doesn't matter much because "speed" is being used by the vast majority to chase a specific subset of loans that historically have shown themselves to outperform. Not necessarily outperform going forward. Remember a few years ago when people like Peter were heavily recommending "repeat borrowers" at Prosper & everyone & their cousin were piling into that category? Much anger & anguish occurred on this blog about the lack of these "good" loans & how they were all being gobbled up quickly in a variety of conspiracy theory type scenarios. Speed seemed real important then & much was made about it. I was in the minority back then in suggesting that people not be concerned with speed & not be part of the rat race in vying for those loans.  Well, how did those loans work out going forward?
Title: Loan Information Asymmetrical Timing
Post by: BruiserB on August 26, 2014, 08:55:17 PM

I would contend that "speed" doesn't matter much because "speed" is being used by the vast majority to chase a specific subset of loans that historically have shown themselves to outperform. Not necessarily outperform going forward. Remember a few years ago when people like Peter were heavily recommending "repeat borrowers" at Prosper & everyone & their cousin were piling into that category? Much anger & anguish occurred on this blog about the lack of these "good" loans & how they were all being gobbled up quickly in a variety of conspiracy theory type scenarios. Speed seemed real important then & much was made about it. I was in the minority back then in suggesting that people not be concerned with speed & not be part of the rat race in vying for those loans.  Well, how did those loans work out going forward?

So do you feel a random selection of notes is just as good as anything?  Or is the key to better performance to cull afterwards by looking for signs of weakness and selling on Folio? 


Sent from my iPad using Tapatalk
Title: Re: Loan Information Asymmetrical Timing
Post by: Fred on August 26, 2014, 09:17:06 PM
Speed gives you opportunities and selections.  You may not utilize the opportunities optimally, but that is a fixable issue.

On the other hand, those who without speed would have a smaller set of selections.
Title: Re: Loan Information Asymmetrical Timing
Post by: lascott on August 26, 2014, 09:45:53 PM
Speed gives you opportunities and selections.  You may not utilize the opportunities optimally, but that is a fixable issue.
On the other hand, those who without speed would have a smaller set of selections.
There is a comment that resonates with my way of thinking.  I've gotten in to P2P lending relatively slow so I can change criteria as I learn.
Title: Re: Loan Information Asymmetrical Timing
Post by: DanB on August 26, 2014, 10:12:12 PM

I would contend that "speed" doesn't matter much because "speed" is being used by the vast majority to chase a specific subset of loans that historically have shown themselves to outperform. Not necessarily outperform going forward. Remember a few years ago when people like Peter were heavily recommending "repeat borrowers" at Prosper & everyone & their cousin were piling into that category? Much anger & anguish occurred on this blog about the lack of these "good" loans & how they were all being gobbled up quickly in a variety of conspiracy theory type scenarios. Speed seemed real important then & much was made about it. I was in the minority back then in suggesting that people not be concerned with speed & not be part of the rat race in vying for those loans.  Well, how did those loans work out going forward?

So do you feel a random selection of notes is just as good as anything?  Or is the key to better performance to cull afterwards by looking for signs of weakness and selling on Folio? 


Sent from my iPad using Tapatalk


This whole "speed" thing presumes that a group of mimickers know what the best loans are going forward. Most of these people who believe they know have this confidence because they think they've isolated characteristics within the historical performance of certain loans & that this will insure that their picks will outperform going forward............or they're just following someone elses analysis, or recommendations.

Sounds real reasonable until you realize that the platforms are themselves actively working against this future & continued out performance by the very subset bandwagon jumpers have identified..............just like they did with the "repeat borrowers". In that example early lenders outperformed for the first few years & the followers & bandwagon jumpers under performed ever since. I have little doubt that at some point in the future that category will once again start to outperform again, .................after the vast majority of bandwagon jumpers have left out of sheer disgust at their under performance.
Title: Re: Loan Information Asymmetrical Timing
Post by: AnilG on August 26, 2014, 11:49:56 PM
DanB, I fully concur with you. You mentioned the key to lending that most lenders seems to ignore. The credit model and criteria used by platforms is not static. The platforms have much more data and resources to analyze them to nullify the pockets of out-performance that lenders may detect by changing the credit models going forward. Also, it takes a few years for these changes to reflect in past performance by that time platform has already moved the goal post.

IMO, the loan selection criteria should be automated based on a few "common-sense" attributes supported by data analysis but beyond that human should take over. It is okay to use any lists/criteria anyone published but then decide which loans in the list you would want to own. I still don't understand how lenders on this forum believe they can outperform by automation when banks after practicing consumer lending for decades and the vast resources at their disposal can't do that with automation.
 

Sounds real reasonable until you realize that the platforms are themselves actively working against this future & continued out performance by the very subset identified..............just like they did with the "repeat borrowers". In that example early lenders outperformed for the first few years & the followers & bandwagon jumpers under performed ever since. I have little doubt that at some point in the future that category will once again start to outperform again, .................after the vast majority of bandwagon jumpers have left out of sheer disgust at their under performance.
Title: Re: Loan Information Asymmetrical Timing
Post by: rawraw on August 27, 2014, 05:12:02 AM
IMO, the loan selection criteria should be automated based on a few "common-sense" attributes supported by data analysis but beyond that human should take over. It is okay to use any lists/criteria anyone published but then decide which loans in the list you would want to own. I still don't understand how lenders on this forum believe they can outperform by automation when banks after practicing consumer lending for decades and the vast resources at their disposal can't do that with automation.
 
I think it's because a lot of people on this forum may not have the most experience in lending/banking.  And websites make it extremely easy to create filter strategies.  Those two things make it really easy to fall for the "liars use statistics."  No credit score card remains fixed, they are constantly updated as a matter of practice.  That combined with the short time period of data creates a fair bit of noise.  If I remember correctly, DanB's investment strategy has always been "Can they pay?"

Now I will add a little caveat that some banks can outperform others in consumer loans via automation in underwriting.  I've seen wide differences in the effectiveness of pricing risk on prime and subprime borrowers.  But your general point I agree with.
Title: Re: Loan Information Asymmetrical Timing
Post by: Rob L on August 27, 2014, 08:56:22 AM
Echos of "What's This Best Loan Selection Business Anyway?":
http://www.lendacademy.com/forum/index.php?topic=1953.0 (http://www.lendacademy.com/forum/index.php?topic=1953.0)
Title: Re: Loan Information Asymmetrical Timing
Post by: AnilG on August 27, 2014, 10:30:11 AM
I agree with you about some banks outperforming others. Primarily, I notice that banks pick their playground wrt whom they want to lend by defining the credit boundaries, cut off points and risk banks are willing to take. Also a lot of automation at banks came as a result of avoiding discrimination complaints.

A few PeerCube users have been asking me to evaluate their portfolios. The common theme I am finding with exceptional portfolios are same investment strategy as DanB's "Can they pay or do they have incentive to pay?", manual review of loans before lending, less than 2,000 notes, active management, and willingness to get rid of loans quickly when they don't meet expectations. The performance of majority of automated hands-off portfolios don't seem to be far off from the averages.

I think it's because a lot of people on this forum may not have the most experience in lending/banking.  And websites make it extremely easy to create filter strategies.  Those two things make it really easy to fall for the "liars use statistics."  No credit score card remains fixed, they are constantly updated as a matter of practice.  That combined with the short time period of data creates a fair bit of noise.  If I remember correctly, DanB's investment strategy has always been "Can they pay?"

Now I will add a little caveat that some banks can outperform others in consumer loans via automation in underwriting.  I've seen wide differences in the effectiveness of pricing risk on prime and subprime borrowers.  But your general point I agree with.
Title: Re: Loan Information Asymmetrical Timing
Post by: DanB on August 27, 2014, 11:11:12 AM
I agree with you about some banks outperforming others. Primarily, I notice that banks pick their playground wrt whom they want to lend by defining the credit boundaries, cut off points and risk banks are willing to take. Also a lot of automation at banks came as a result of avoiding discrimination complaints.

A few PeerCube users have been asking me to evaluate their portfolios. The common theme I am finding with exceptional portfolios are same investment strategy as DanB's "Can they pay or do they have incentive to pay?", manual review of loans before lending, less than 2,000 notes, active management, and willingness to get rid of loans quickly when they don't meet expectations. The performance of majority of automated hands-off portfolios don't seem to be far off from the averages.

 
I think it's because a lot of people on this forum may not have the most experience in lending/banking.  And websites make it extremely easy to create filter strategies.  Those two things make it really easy to fall for the "liars use statistics."  No credit score card remains fixed, they are constantly updated as a matter of practice.  That combined with the short time period of data creates a fair bit of noise.  If I remember correctly, DanB's investment strategy has always been "Can they pay?"

Now I will add a little caveat that some banks can outperform others in consumer loans via automation in underwriting.  I've seen wide differences in the effectiveness of pricing risk on prime and subprime borrowers.  But your general point I agree with.

I also don't follow or even study other peoples filters. I don't eliminate categories from consideration & I don't care what the stated loan purpose is, since I don't believe what people say & doubt they can execute it in the fashion that they claim they will anyway even if I did believe. Why, because I know better. Unlike 98% of lenders here, I've spent 90% of my adult life living the life of the D-G borrower. Also, I don't have a minimum income requirement. ( I know, some of you readers are surprised by this one, but no I won't give you my reason :) )

I rarely if ever lend for 5 years. Why? Because  in my experience a substantial percentage of people in the D-G pool can't effectively plan for 5 years into the future. So, yes there are a number of smaller factors that I weigh & maybe even decide on emotionally.  And yes, after 5 full years of LC lending & 5 years of substantially outperforming the averages every single year..............I still review each loan manually & rapidly. And no I don't sign on 4 times a day. And yes I do make a concerted effort to look at what others aren't looking at in p2p & in life.  :)

PS.............One of the things I've found most amusing in p2p is how a group of lenders who are good savers/investors have some money to tons of money, have good to great credit scores etc etc have gone out of their way to lend to this other group of people who don't have money, have crap credit scores & are willing to pay interest rates they would never touch themselves............. And the kicker is that these lenders study, ponder & create models in the belief that they're remotely qualified to assess risk & behavior of this group of people who other than being air breathers, they have never had anything in common with. I'm sorry but that seems hilarious to me.
Title: Re: The end of BlueVestment?
Post by: rawraw on August 27, 2014, 11:13:45 AM
We've missed you DanB
Title: Re: Loan Information Asymmetrical Timing
Post by: lascott on August 27, 2014, 12:50:27 PM
I also don't follow or even study other peoples filters. I don't eliminate categories from consideration & I don't care what the stated loan purpose is, since I don't believe what people say & doubt they can execute it in the fashion that they claim they will anyway even if I did believe. Why, because I know better. Unlike 95% of lenders here, I've spent 90% of my adult life living the life of the D-G borrower. Also, I don't have a minimum income requirement. ( I know, some of you readers are surprised by this one, but no I won't give you my reason :) )

I rarely if ever lend for 5 years. Why? Because  in my experience a substantial percentage of people in the D-G pool can't effectively plan for 5 years into the future. So, yes there are a number of smaller factors that I weigh & maybe even decide on emotionally.  And yes, after 5 full years of LC lending & 5 years of substantially outperforming the averages every single year..............I still review each loan manually & rapidly. And no I don't sign on 4 times a day. And yes I do make a concerted effort to look at what others aren't looking at in p2p & in life.  :)

PS.............One of the things I've found most amusing in p2p is how a group of lenders who are good savers/investors have some money to tons of money, have good to great credit scores etc etc have gone out of their way to lend to this other group of people who don't have money, have crap credit scores & are willing to pay interest rates they would never touch themselves............. And the kicker is that these lenders study, ponder & create models in the belief that they're remotely qualified to assess risk & behavior of this group of people who other than being air breathers, they have never had anything in common with. I'm sorry but that seems hilarious to me.
I liked your insight and background. Gives me some things to ponder. I'm already trying to get out of some of the normal criteria but doing it in an automated way. Just want to stay above average and limit defaults.  Already starting some retirement mode thinking (balance % of stock vs bonds analogy to % D,Es vs A,B,C in my portfolio).

Re: Your "PS".
Various reasons people invest in P2P, IMO.
Topic: Which Type of Peer to Peer Lending Investor Are You?
http://www.lendacademy.com/forum/index.php?topic=2544.0
Title: Re: Loan Information Asymmetrical Timing
Post by: JoeB on August 27, 2014, 04:24:15 PM
I agree with you about some banks outperforming others. Primarily, I notice that banks pick their playground wrt whom they want to lend by defining the credit boundaries, cut off points and risk banks are willing to take. Also a lot of automation at banks came as a result of avoiding discrimination complaints.

A few PeerCube users have been asking me to evaluate their portfolios. The common theme I am finding with exceptional portfolios are same investment strategy as DanB's "Can they pay or do they have incentive to pay?", manual review of loans before lending, less than 2,000 notes, active management, and willingness to get rid of loans quickly when they don't meet expectations. The performance of majority of automated hands-off portfolios don't seem to be far off from the averages.

 
I think it's because a lot of people on this forum may not have the most experience in lending/banking.  And websites make it extremely easy to create filter strategies.  Those two things make it really easy to fall for the "liars use statistics."  No credit score card remains fixed, they are constantly updated as a matter of practice.  That combined with the short time period of data creates a fair bit of noise.  If I remember correctly, DanB's investment strategy has always been "Can they pay?"

Now I will add a little caveat that some banks can outperform others in consumer loans via automation in underwriting.  I've seen wide differences in the effectiveness of pricing risk on prime and subprime borrowers.  But your general point I agree with.

I also don't follow or even study other peoples filters. I don't eliminate categories from consideration & I don't care what the stated loan purpose is, since I don't believe what people say & doubt they can execute it in the fashion that they claim they will anyway even if I did believe. Why, because I know better. Unlike 95% of lenders here, I've spent 90% of my adult life living the life of the D-G borrower. Also, I don't have a minimum income requirement. ( I know, some of you readers are surprised by this one, but no I won't give you my reason :) )

I rarely if ever lend for 5 years. Why? Because  in my experience a substantial percentage of people in the D-G pool can't effectively plan for 5 years into the future. So, yes there are a number of smaller factors that I weigh & maybe even decide on emotionally.  And yes, after 5 full years of LC lending & 5 years of substantially outperforming the averages every single year..............I still review each loan manually & rapidly. And no I don't sign on 4 times a day. And yes I do make a concerted effort to look at what others aren't looking at in p2p & in life.  :)

PS.............One of the things I've found most amusing in p2p is how a group of lenders who are good savers/investors have some money to tons of money, have good to great credit scores etc etc have gone out of their way to lend to this other group of people who don't have money, have crap credit scores & are willing to pay interest rates they would never touch themselves............. And the kicker is that these lenders study, ponder & create models in the belief that they're remotely qualified to assess risk & behavior of this group of people who other than being air breathers, they have never had anything in common with. I'm sorry but that seems hilarious to me.


One word: BRAVO!

My feelings exactly.
Title: Re: The end of BlueVestment?
Post by: DanB on August 28, 2014, 02:21:39 AM
We've missed you DanB

Thank you. That's always nice to hear.
Title: Re: Loan Information Asymmetrical Timing
Post by: turing on August 28, 2014, 10:42:53 AM
PS.............One of the things I've found most amusing in p2p is how a group of lenders who are good savers/investors have some money to tons of money, have good to great credit scores etc etc have gone out of their way to lend to this other group of people who don't have money, have crap credit scores & are willing to pay interest rates they would never touch themselves

This criticism should be leveled at all banking and lending, not just P2P.  Credit cards, bank loans, home loans, pay-day loans, etc all do the exact same things.

They all loan money to people who don't have money (otherwise they wouldn't want the loans)
Most bankers and credit analysts have little in common demographically with the people they lend to.

And the kicker is that these lenders study, ponder & create models in the belief that they're remotely qualified to assess risk & behavior of this group of people who other than being air breathers, they have never had anything in common with. I'm sorry but that seems hilarious to me.

Yes, there is a place for knowledge that comes from actually experiencing something.  But this statement completely ignores the importance of data as well.  Both are valuable and can be helpful for decisions.

By your logic, all biology, zoology, oncology, etc is bunk.  How can we predict the behavior of a cancer cell when we are not cancer cells?  How can we predict the behavior of a lion when we are not lions? 

We look at data regarding past behavior and can make predictions on future behavior.  Not because we have lived it but because on aggregate the data does a good job predicting.

When we roll dice we only guess the correct number 1/6th of the time on any individual roll, but I can tell you with high confidence that after 10,000 rolls there will be roughly the same number of 1's as 6's.
Title: Re: Loan Information Asymmetrical Timing
Post by: rawraw on August 28, 2014, 10:56:22 AM
PS.............One of the things I've found most amusing in p2p is how a group of lenders who are good savers/investors have some money to tons of money, have good to great credit scores etc etc have gone out of their way to lend to this other group of people who don't have money, have crap credit scores & are willing to pay interest rates they would never touch themselves

This criticism should be leveled at all banking and lending, not just P2P.  Credit cards, bank loans, home loans, pay-day loans, etc all do the exact same things.

They all loan money to people who don't have money (otherwise they wouldn't want the loans)
Most bankers and credit analysts have little in common demographically with the people they lend to.
Banks don't use the principals money to lend.  It's not a good comparison
Title: Re: Loan Information Asymmetrical Timing
Post by: DanB on August 28, 2014, 03:18:23 PM
PS.............One of the things I've found most amusing in p2p is how a group of lenders who are good savers/investors have some money to tons of money, have good to great credit scores etc etc have gone out of their way to lend to this other group of people who don't have money, have crap credit scores & are willing to pay interest rates they would never touch themselves

This criticism should be leveled at all banking and lending, not just P2P.  Credit cards, bank loans, home loans, pay-day loans, etc all do the exact same things.

They all loan money to people who don't have money (otherwise they wouldn't want the loans)
Most bankers and credit analysts have little in common demographically with the people they lend to.

And the kicker is that these lenders study, ponder & create models in the belief that they're remotely qualified to assess risk & behavior of this group of people who other than being air breathers, they have never had anything in common with. I'm sorry but that seems hilarious to me.

Yes, there is a place for knowledge that comes from actually experiencing something.  But this statement completely ignores the importance of data as well.  Both are valuable and can be helpful for decisions.

By your logic, all biology, zoology, oncology, etc is bunk.  How can we predict the behavior of a cancer cell when we are not cancer cells?  How can we predict the behavior of a lion when we are not lions?

We look at data regarding past behavior and can make predictions on future behavior.  Not because we have lived it but because on aggregate the data does a good job predicting.

When we roll dice we only guess the correct number 1/6th of the time on any individual roll, but I can tell you with high confidence that after 10,000 rolls there will be roughly the same number of 1's as 6's.

So, you're objecting to the criteria my particular mind has in determining what I personally find amusing?? That's pretty amusing to me as well. Your attempt to use logic to analyse my sense of humor is also rather telling & amusing. You can't predict the behavior of a specific lion or any one specific animal. I thought that the loss of arm (& lives) has proved that over & over again. Sure after 10,000 rolls, but not after 1000 & no predictability of sequence within that 10,000. Incidentally, most people here never get to 10,000 rolls. I've been doing this 5 yrs, & I'm barely over 3000.
Thanks though for providing me even more amusement. You are a classic example of a person who can play the notes well while playing piano & yet never really be called an artist.:) Yes, I know, you don't really understand this statement either. Yet strangely enough I find that amusing as well & can live with the humor I feel. :)
Title: Re: Loan Information Asymmetrical Timing
Post by: hoggy1 on August 28, 2014, 03:59:37 PM

Peter,
We need a new General Category Board subdivision - an un-editable Hall of Fame for award winning post from other categories. Member nominate and second a posts, then the membership votes in a poll.

I nominate DanB's post below as the first Hall of Fame post. Any seconds?

I also don't follow or even study other peoples filters. I don't eliminate categories from consideration & I don't care what the stated loan purpose is, since I don't believe what people say & doubt they can execute it in the fashion that they claim they will anyway even if I did believe. Why, because I know better. Unlike 95% of lenders here, I've spent 90% of my adult life living the life of the D-G borrower. Also, I don't have a minimum income requirement. ( I know, some of you readers are surprised by this one, but no I won't give you my reason :) )

I rarely if ever lend for 5 years. Why? Because  in my experience a substantial percentage of people in the D-G pool can't effectively plan for 5 years into the future. So, yes there are a number of smaller factors that I weigh & maybe even decide on emotionally.  And yes, after 5 full years of LC lending & 5 years of substantially outperforming the averages every single year..............I still review each loan manually & rapidly. And no I don't sign on 4 times a day. And yes I do make a concerted effort to look at what others aren't looking at in p2p & in life.  :)

PS.............One of the things I've found most amusing in p2p is how a group of lenders who are good savers/investors have some money to tons of money, have good to great credit scores etc etc have gone out of their way to lend to this other group of people who don't have money, have crap credit scores & are willing to pay interest rates they would never touch themselves............. And the kicker is that these lenders study, ponder & create models in the belief that they're remotely qualified to assess risk & behavior of this group of people who other than being air breathers, they have never had anything in common with. I'm sorry but that seems hilarious to me.
Title: Re: Loan Information Asymmetrical Timing
Post by: turing on August 28, 2014, 04:03:41 PM
I thought that the loss of arm (& lives) has proved that over & over again. Sure after 10,000 rolls, but not after 1000 & no predictability of sequence within that 10,000. Incidentally, most people here never get to 10,000 rolls. I've been doing this 5 yrs, & I'm barely over 3000.

Predictability of sequence doesn't matter.  If you buy 100 notes it doesn't matter if notes 50-52 default, or if note 5, 20, and 67 default (random numbers) default.  It is still 3 defaults and 3% default rate.

I completely agree that we cannot predict behavior of one borrower (or one lion).  That is why we don't give loans to one single borrower.  We give loans to lots of borrowers and can sleep at night knowing the overall behavior is predictable.
Title: Re: Loan Information Asymmetrical Timing
Post by: DanB on August 28, 2014, 05:37:06 PM
I thought that the loss of arm (& lives) has proved that over & over again. Sure after 10,000 rolls, but not after 1000 & no predictability of sequence within that 10,000. Incidentally, most people here never get to 10,000 rolls. I've been doing this 5 yrs, & I'm barely over 3000.


I completely agree that we cannot predict behavior of one borrower (or one lion).  That is why we don't give loans to one single borrower.  We give loans to lots of borrowers and can sleep at night knowing the overall behavior is predictable.

Yes, but no one here wants "predictable". Because predictable is just another way of saying performing like the average. This conversation is about people who are trying to outperform the average. I've outperformed the averages for all 5 years combined & each of the 5 years individually.  How about you?  This conversation is also about whether automated can in the long term outperform the average. I say maybe, but only if the automation can somehow blend science with some "art"............& I haven't seen anyone here really show that's even possible.

Title: Re: Loan Information Asymmetrical Timing
Post by: rawraw on August 28, 2014, 06:29:11 PM
I use automation for note selection and FolioFN manually to "out perform" but I don't deal with a high mix of low grade notes. I'm over 50% in C or better, but still have returns similar to the low grade folks
Title: Re: Loan Information Asymmetrical Timing
Post by: cfb on August 29, 2014, 02:04:55 PM

I been concerned about this for a long time. If you are correct that the number are in the 60-80% range most of our analysis is almost useless as you describe. Those type of events will not distribute themselves evenly over incomes (higher income borrowers will weather a minor emergency better than someone living on the edge) but are likely to distribute themselves more evenly over FICOs (maybe - first thing to come to mind) or other discriminators. Let me put my statistician hat on and consult with a few colleagues to see if a better estimate can be calculated. And I hope the answer is not as high as you suggest or I'm wasting my time.

I sort of look at things from a top down vs data points up, because a lot of time meandering through the data points misses macroscopic evidence.

Most of my defaults are a result of bankruptcy vs "just stopped paying and disappeared".  The 5 top reasons (accounting for almost all) for bankruptcies are:

1) Medical expenses accounting for 40-62% of bankruptcies, depending on whose statistics you read.  Sudden high medical expenses can't be predicted from credit data.  Although I steer clear of loans for medical purposes because that seems to me to indicate that the person has had poor health or is about to engage in expensive medical care.

2) Job loss.  Not predictable using credit data.  I sort of liked it when they listed the company someone worked for.  I regularly saw "Company xyz to lay off 20,000 workers" and then a bunch of loan requests from people who worked there.  I'd pass.

3) Poor/excess use of credit, with the added interesting tidbit that most debt consolidation plans fail.  This one you might catch unless they've been holding it together and the LC loan is their last ditch effort to not go under.  Considering the high percentage of bankruptcies I've had within ~1 year or less of loan issuance, I'm buying into a lot of these last ditch efforts.

4) Divorce/separation.  The more someone makes, the less likely they are to divorce, so you could improve on this by only funding high income loans.  The more education someone has, the less likely they are to pull the plug, but you can't see education levels.  I think those two go hand in hand and the causative factor is too much money on the table to take a walk and lose it.  On the other hand, financial trouble is a leading divorce cause and points us back to #3, so you might imagine someone with a lower credit score, higher debt and marginal income to be ripe for a divorce. You might catch some potential divorce candidates with careful analysis, but we can't see marital status.

5) Unexpected high expenses.  Earthquakes, floods, tornado's, hurricanes, etc without the necessary or proper level of insurance, or such conditions causing at least a temporary loss of a job.  For example, many businesses closed, some permanently, after Katrina.  Loss of an underinsured car or to major mechanical failure can easily derail someone close to the edge.  While you could steer clear of places like Florida and California and tornado alley to mitigate the natural disasters, most of these situations aren't due to natural disasters.  While you might detect someone close to the edge or avoid low income borrowers who might not be able to power their way out of a $10-30k loss...still hard to detect stuff like "someone stole my car, I have no insurance, and I can't afford to buy another one" from a credit report.

Add to that about 1-1.5% or so of the borrowing age population in the US dies each year.  Unless they have a huge estate, since the majority of people have little net worth after secured debt is considered, those are almost automatic defaults.  The actual death rate is .8% but much of our population is <20 or over 65 and unlikely to be a borrower

There are a couple of other funny things about trying to reduce risk.  The first is what the demographic of an LC borrower is and if its changed over time?  Is the change sudden or a curve?  Are we getting trendy internet aware people who after careful analysis realize they can get a better deal through LC than a bank or using a credit card?  Are we getting people who realize that nobody who lends from a big building is going to give them a loan?  Are the people from the first few years of data we have drastically different from today's borrower?  Is the lender activity (immediate funding of many loans through feeding frenzies) changing the borrower type or behavior?  Is the long term data on consumer lending even applicable to the average LC borrower?

If todays LC borrower <> the average unsecured loan borrower, then data outside of the LC data we have in hand is worthless.  If the demographic for an LC borrower is different today than it was 5 years ago then all the data we have is worthless.  I'm going to guess than the first 2-3 years of LC borrowers were adventurous types who got looped in by marketing and only the better ones got funded.  I'm guessing current LC borrowers are more opportunistic and know they can get almost any loan funded at this point.

As an interesting aside point, I spend a fair amount of time online but have always used an ad blocker.  I also do an enormous variety and volume of investing and have done so for 20+ years.  I've never, ever seen anything about lending club until a couple of years ago when someone mentioned it as a joke on a web site for cheap deals on stuff.  Someone asked where they could park some money for a couple of months risk free and with high return and liquidity and someone said "lending club", followed by a lot of laughter.  Hence my feeling that much of the borrowing is online ad driven.  So your demographic is someone who clicks on an internet ad for borrowing money online.  Ick.

Only other point I'll make in my oh-so-brief post is the one about 5 year loans.  Certainly in a bare data evaluation, a 5 year loan is subject to a greater chance of striking death or one of the above 5 reasons to go bankrupt.  However in a rolling situation you'll have reinvested your 3 year loan and be 2 years into the next 3 year loan to match the 5 year term.  So you're going to have to deal with two borrowers, each with an unknown set of random potential problems.  Since we all know that the majority of defaults occur while a loan is 'seasoning', I think you're taking on more risk by buying 1.66 3 year loans vs 1.0 5 year loans.  I don't think I've seen a solid evaluation of 5 year loans vs 3 year loans with a tacked on 2 out of 3 year 3 year loan.  Since we also don't know how the demographics of borrowers has changed over time, we don't know if that longer term 5 year loan data is even valid for current borrowers.

Given that there are things I'm okay with lending about, but those sell out in 20 seconds and I'm not into trusting an automated process...this just doesn't work for me anymore.  But I think I'll make 9-10% on cash that would have been sitting in a checking account or money market account.  But I don't think I could invest $60k in a timely manner or get 10% out of it if I started today.

And I don't think the backward looking data analysis is worth beans.  Except for weeding out the really, really obviously dangerous loans that I don't think should even have been approved by LC.

Title: Re: Loan Information Asymmetrical Timing
Post by: AnilG on August 29, 2014, 02:50:34 PM
I don't have the default reason data from Lending Club. But the default reasons from Prosper are shown in the attached image. About half of defaults are due to bankruptcy.

Most of my defaults are a result of bankruptcy vs "just stopped paying and disappeared".  The 5 top reasons (accounting for almost all) for bankruptcies are:
Title: Re: Loan Information Asymmetrical Timing
Post by: cfb on August 29, 2014, 03:49:23 PM
I don't have the default reason data from Lending Club. But the default reasons from Prosper are shown in the attached image. About half of defaults are due to bankruptcy.

Looks like quite a few of those outcomes are good ones though.  Does 'delinquency' mean they're late but haven't fully defaulted/charged off yet?

The majority of my chargeoff/defaults are bankruptcy.  So I was relatively good at picking out the ones that weren't going to declare bankruptcy due to reasons you couldn't mine off an LC credit report.  Many are in as soon as 3 months, few went more than a year.  So it was their 'hail mary', knowing if they went bankrupt an extra $35k owed wouldn't matter since they wouldn't be paying it back anyhow.  Or maybe they partied their brains out for six months knowing it was going to be pretty sparse living for the next 7 years.
Title: Re: Loan Information Asymmetrical Timing
Post by: RaymondG on August 30, 2014, 03:07:49 PM
Speed gives you opportunities and selections.  You may not utilize the opportunities optimally, but that is a fixable issue.

On the other hand, those who without speed would have a smaller set of selections.

I totally agree with Fred. To me, the "Speed" is mainly about the size of available selections. I use automation so it is always better that more loans are available for examining. Automation takes much less time, effort while it achieves comparable result. It is coded to do what I would do manually and it did eliminate errors I often made when I hand-picked loans. It is also adjusted as new observations are found in periodical portfolio reviews. Why waste time to review loans manually on a smaller pool?
Title: Re: Loan Information Asymmetrical Timing
Post by: Fred on August 31, 2014, 01:27:24 AM
And I don't think the backward looking data analysis is worth beans.

How do you think people came up with the list of "the 5 top reasons for bankruptcies"?
Title: Re: Loan Information Asymmetrical Timing
Post by: rawraw on August 31, 2014, 06:59:31 AM
Backward looking data can be useful; however, I think the way most people do it with their filters is wrong.  Due to the underwriting changes that occur, what you really need to do is establish probability of defaults based on credit variables.  Then you subtract those against the current interest rate offered and choose the highest "return" notes.  This is much different than using the "net return" from filtering websites, which includes the interest rate historically offered which may be different from today.  In the filtering strategy, you are seeking what would've worked in the past but the underwriting of the past has changed.  And there may be several underwriting models contained in that small historical sample chosen. 
Title: Re: Loan Information Asymmetrical Timing
Post by: cfb on August 31, 2014, 12:13:02 PM
And I don't think the backward looking data analysis is worth beans.

How do you think people came up with the list of "the 5 top reasons for bankruptcies"?

Not by looking at lending club data  :)  IIRC the data was compiled by voluntary surveys of people when they were declaring bankruptcy.  Clearly nobody does this because they want to, circumstances just dictate that they have to.  While people bad with money might telegraph a likely impending bankruptcy via lates, inquiries and derogatories I'd imagine there are quite a few people who do whatever they can to keep paying the bills.  The problem for us is when an LC loan is the last thing they can try before the merry go round stops.

I have A's and B's who declared bankruptcy that are 700+, no defaults, no lates, no problems.

And again, it all hinges on whether borrowers and lenders are largely the same sorts of people now as they were in each year the data was collected.  My recollection is quite a few iffy loans didn't get funded 2+ years ago.  Looks like almost everything gets funded these days.
Title: Re: Loan Information Asymmetrical Timing
Post by: hoggy1 on August 31, 2014, 12:25:29 PM
CFB,

Thanks for the lengthy, thoughtful, and insightful post above beginning
Quote
I sort of look at things from a top down vs data points up, because a lot of time meandering through the data points misses macroscopic evidence.

Most of my defaults are a result of bankruptcy vs "just stopped paying and disappeared".  The 5 top reasons (accounting for almost all) for bankruptcies are:

However you go on in your subsequent post however to say this:
Quote
The majority of my chargeoff/defaults are bankruptcy.  So I was relatively good at picking out the ones that weren't going to declare bankruptcy due to reasons you couldn't mine off an LC credit report.  Many are in as soon as 3 months, few went more than a year.

As I pointed out earlier, bankruptcies and other defaults owing to "materially destructive life events" without "intent'"which you argue might constitute 60-80% of all defaults and which would be undiscernible from credit data, should certainly distribute themselves evenly over the life of all loans. If you are correct and the majority of defaults occur early in the life of loans (I think we all know that is true?) then it can't be that a large fraction of them are the result of sudden, unexpected, materially destructive life events that hadn't already happened or begun before the loan was taken.

Your thoughts?


Title: Re: Loan Information Asymmetrical Timing
Post by: rawraw on August 31, 2014, 05:41:09 PM
As I pointed out earlier, bankruptcies and other defaults owing to "materially destructive life events" without "intent'"which you argue might constitute 60-80% of all defaults and which would be undiscernible from credit data, should certainly distribute themselves evenly over the life of all loans. If you are correct and the majority of defaults occur early in the life of loans (I think we all know that is true?) then it can't be that a large fraction of them are the result of sudden, unexpected, materially destructive life events that hadn't already happened or begun before the loan was taken.

Your thoughts?
A person's financial situation impacts their ability to respond to "materially destructive life events."  For a G borrower, it could be requiring a new transmission.  For an A borrower, it could be requiring a new car.  Financial condition is going to be the factor that modifies the impact of normally distributed life events to impact different groups differently (non-normal distribution of adverse events to creditors)

Quote
I have A's and B's who declared bankruptcy that are 700+, no defaults, no lates, no problems.
We aren't talking anecdotes.  We are talking about probabilities based on the frequency at which it occurs.  Are you suggesting that A and B bankruptcy's with FICO of 700, no lates, "no problems" are as frequent as G bankruptcies?  Or are you just being difficult :)
Title: Re: Loan Information Asymmetrical Timing
Post by: hoggy1 on August 31, 2014, 06:02:20 PM
As I pointed out earlier, bankruptcies and other defaults owing to "materially destructive life events" without "intent'"which you argue might constitute 60-80% of all defaults and which would be undiscernible from credit data, should certainly distribute themselves evenly over the life of all loans. If you are correct and the majority of defaults occur early in the life of loans (I think we all know that is true?) then it can't be that a large fraction of them are the result of sudden, unexpected, materially destructive life events that hadn't already happened or begun before the loan was taken.

Your thoughts?
A person's financial situation impacts their ability to respond to "materially destructive life events."  For a G borrower, it could be requiring a new transmission.  For an A borrower, it could be requiring a new car.  Financial condition is going to be the factor that modifies the impact of normally distributed life events to impact different groups differently (non-normal distribution of adverse events to creditors)

I agree with you, but don't see what it has to do with what I was pointing out in my quote above. My point is that WHEN those event occur they should be EVENLY DISTRIBUTED in TIME over the life of the loans (with a big sample). We have all accepted that P2P loans default at a higher rate in the first 3-12 months. That is inconsistent with those events being largely "unsuspected life surprises". If they were then we should see just as many defaults in the second and third trimester as the first.


Quote
Quote
I have A's and B's who declared bankruptcy that are 700+, no defaults, no lates, no problems.
We aren't talking anecdotes.  We are talking about probabilities based on the frequency at which it occurs.  Are you suggesting that A and B bankruptcy's with FICO of 700, no lates, "no problems" are as frequent as G bankruptcies?  Or are you just being difficult :)

The second quote is not from me and precisely the post I was arguing didn't stand up to scrutiny.
Title: Re: Loan Information Asymmetrical Timing
Post by: rawraw on August 31, 2014, 06:43:28 PM
I only was responding to you in the first part. I didn't understand you meant in time
Title: Re: Loan Information Asymmetrical Timing
Post by: cfb on September 01, 2014, 02:36:33 PM
As I pointed out earlier, bankruptcies and other defaults owing to "materially destructive life events" without "intent'"which you argue might constitute 60-80% of all defaults and which would be undiscernible from credit data, should certainly distribute themselves evenly over the life of all loans. If you are correct and the majority of defaults occur early in the life of loans (I think we all know that is true?) then it can't be that a large fraction of them are the result of sudden, unexpected, materially destructive life events that hadn't already happened or begun before the loan was taken.

Your thoughts?

Why aren't they evenly distributed?

I'm betting a fair number of early loan cycle bankruptcies know they're in serious financial trouble and are planning on using the LC loan hoping to recover or at least put off the bankruptcy until they can think of something else.  So that tilts the table towards the early part of the loan.  Problem is, what's wrong or is about to go wrong may not show up in the data we get to see.  When you see lots of inquiries, very high revolving debt, extremely high dti or lates/defaults you've got the sign.  I don't buy those loans, and yet see perfectly good loans bankrupt within a year or so, often just a few months.

In medical cost bankruptcies they've been really sick or injured recently, probably with lost time on the job, or they're about to go into a possibly risky medical procedure.  So they might die, not be able to get back to work, or be overwhelmed by the costs.  Someone like that might present themselves as a credit card/debt consolidation vs medical, have good looking numbers, and just went to LC first so don't have the inquiries/lates yet.  Unexpected medical issues during the loan would of course be evenly distributed within a demographic class.  Health and hazardous behaviors tend to be more common within the low income/financially stressed end.  So a G should experience more of them than an A.  Its just a question of which G will experience it.  Which you won't figure out with backward looking data.

Divorces...most of them are due to money/financial/health.  I'd imagine that someone looking to borrow $35K is experiencing one of those problems.  If borrowing the money doesn't fix the problem and/or the monthly payments are burdensome then you have a  higher early term likelihood of divorce.  Much of the rest of the reasons have to do with discontent.  I'm thinking a four figure monthly payment digs into the lifestyle a little.  So of course the same lower income, higher financial problem folks will see more of this type of bankruptcy than high income, higher educated people.  The distribution within finance bands should be spread out but I'm looking at the loan as a correlated factor and perhaps causative. So avoiding low income loans helps cut into the curve.

There's also what I'll call the pseudo-bankruptcy.  Its when after paying their mortgage/rent, car payment and utility bills they find they don't have enough to meet their other obligations.  Their largest monthly bill that doesn't result in being homeless or having something turned off is their LC bill.  If they pay that, they have to skip paying 6 other people and getting 6 delinquencies or skipping the LC loan payment and getting 1.  In this case the borrower did the same things a bankruptcy would do: allocate assets towards secured loans and essential bills, stiffing the unsecured loan.  Another pseudo bankruptcy is from people who die and don't have much in the way of assets or immediate family that wants to process a bankruptcy or send in copies of the death certificate when a bill shows up.

I think if you avoid loans with low FICO's, low income, high inquiries and dump notes that experience sharp FICO drops you improve your chances of avoiding default.  I don't think you'll find enough high F's and G's that have a high FICO, high income and no inquiries to matter.  Aside from that small improvement I think your risk adjusted returns will be substantially similar.  The big improvement on returns is not buying A's, B's and most C's.

Getting slightly back to topic, we have a limited range of notes to buy.  Analysis of the notes is limited because filled notes don't make the data dump file.  Institutional investors can process and buy up the notes worthwhile to buy, up to the limit and deployment rate of cash they have.  Part of the problem is building filters against 100% of the data when we're only going to have legit access to <50% of the notes.  Filtering just shortens the time you have to buy and the longer the wait the poorer the quality of notes you'll be choosing from, simply because the hordes of borrowers didn't want them or thought they'd do better spreading out their deployment over longer periods of time.

Be interesting to see a returns analysis on just notes that made the data dump with and without the well accepted filters vs all notes.  I'd bet that there is quite a returns gap between what gets snapped up before any of us get a chance and the dregs left over, and not much delta in filtered vs non filtered.

I also did an analysis a while back and found that buying C4-E2 notes with no filters gave a very similar long term return to a filtered D-G strategy, (about 1% or less difference) but offered more than enough notes on a regular basis to build and maintain a portfolio.  So given the need for speed vs quality analysis, folks might look at that.
Title: Re: Loan Information Asymmetrical Timing
Post by: hoggy1 on September 01, 2014, 03:19:27 PM
I'm betting a fair number of early loan cycle bankruptcies know they're in serious financial trouble and are planning on using the LC loan hoping to recover or at least put off the bankruptcy until they can think of something else. 

I agree but as you point out at least there is a possibility of detecting such potential defaults in credit data.   I think these are the majority of defaults we see. I also agree that most defaults are NOT (though there clearly are some outright fraudsters) people who intended to default.  I think these are people living comfortably NEAR or just over the edge who thinking is "if we can just get over this hump, everything will be OK after we get our year end raises" but the new loan itself pushes them off the cliff. One reason I generally only fund debt or CC loans.

Why aren't they evenly distributed?

I hate to just repeat myself, but you yourself have pointed out here and elsewhere that they are not evenly distributed in TIME. I quote my earlier post changing one word:
Quote
My point is that WHEN those event occur they should be EVENLY DISTRIBUTED in TIME over the life of the loans (with a big sample). We have all accepted that P2P loans default at a higher rate in the first 3-12 months. That is inconsistent with those events being largely "unsuspected life surprises". If they were then we should see ALMOST as many defaults in the second and third trimester as the first.
Title: Re: Loan Information Asymmetrical Timing
Post by: cfb on September 01, 2014, 04:22:49 PM
I was just repeating your question (not asking it), and I thought I'd excessively explained why they would be front loaded rather than evenly distributed.

The short answer is if you're borrowing tons of money after clicking on a banner ad, there's obviously something less than great going on.  But your credit report may not reflect what it is.  And chances are either whatever is going wrong that required the loan or the burden of the loan itself is going to result in the borrower not repaying and perhaps going bankrupt.
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on September 10, 2014, 09:12:15 AM
Last night at 6 and today at 6 am the API saw loans around 7-10s after the hour. This is a nice improvement. I hope it's sustained.
Title: Re: Loan Information Asymmetrical Timing
Post by: Kombinator on September 10, 2014, 11:42:17 AM
Ah nice, so glad someone is tracking this, it would certainly be a highly logical and fair improvement to put the API and manual filter users on par.  Perhaps if you are tracking this can put up a delay graph over the last say 2 months and keep it up for the next two to see if the diff converges to hopefully 0?
Title: Re: Loan Information Asymmetrical Timing
Post by: lascott on September 10, 2014, 11:47:57 AM
Ah nice, so glad someone is tracking this, it would certainly be a highly logical and fair improvement to put the API and manual filter users on par.  Perhaps if you are tracking this can put up a delay graph over the last say 2 months and keep it up for the next two to see if the diff converges to hopefully 0?
I didn't think it was manual filter users that would be a big concern. I was thinking people using LendingClub "Automated Investing" (formerly PRIME) getting the multiple second head start.  As well as the screen scrapers auto populating their cart for later viewing.
Title: Re: Loan Information Asymmetrical Timing
Post by: MarinBB on September 10, 2014, 02:37:26 PM

On a relative basis, does the generic API BrowseLoans(true) action take much longer than downloading the CSV file for you guys? I recall that this was the case ~2mo ago when I last looked at the API; not sure if it has been cleaned up.
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on September 10, 2014, 02:38:58 PM
Manual buyers are a concern, too, because they can make a filter on the LC site, apply it within moments of a release, and then cart the result set. Even if it's $100 a loan, it only takes a few users to buy into a big chunk of loans.

Regarding the person asking for a chart, unfortunately I don't have time to do something like this for free.

Finally, today's 10am drop was 30s after the hour (back to unfortunate normal). Sigh.
Title: Re: Loan Information Asymmetrical Timing
Post by: rawraw on September 10, 2014, 02:41:41 PM
I still don't understand this chart stuff.  Have I missed the reply that explained it?  To have that much cash to lock up notes to only look for notes that qualify means you end up holding a significant amount of cash earning nothing, assuming you can only invest in what you carted.  I don't get how this is viable or why people are worried about it.  I guess if a lot of people were doing it, it could be annoying.  But at the same time, they'd be having huge cash drags that could potentially offset the decreased default rates.
Title: Re: Loan Information Asymmetrical Timing
Post by: MarinBB on September 10, 2014, 02:47:04 PM
Manual buyers are a concern, too, because they can make a filter on the LC site, apply it within moments of a release, and then cart the result set. Even if it's $100 a loan, it only takes a few users to buy into a big chunk of loans.

Regarding the person asking for a chart, unfortunately I don't have time to do something like this for free.

Finally, today's 10am drop was 30s after the hour (back to unfortunate normal). Sigh.

That's what I saw too, Bryce. New listings appeared on the website right on time at 10am sharp but did not show up in the file until 10:01:00 (60sec. later). By this point, two listings had filled and never appeared in the file: 27210183 and 26989832. I guess the current order of first appearance for new listings is 1) website, 2) API, and 3) CSV file though I've seen various permutations in the past week and I assume that LC is making changes on their side.
Title: Re: Loan Information Asymmetrical Timing
Post by: Rob L on September 10, 2014, 11:30:22 PM
I still don't understand this chart stuff.  Have I missed the reply that explained it?  To have that much cash to lock up notes to only look for notes that qualify means you end up holding a significant amount of cash earning nothing, assuming you can only invest in what you carted.  I don't get how this is viable or why people are worried about it.  I guess if a lot of people were doing it, it could be annoying.  But at the same time, they'd be having huge cash drags that could potentially offset the decreased default rates.

Think you made a typo and meant to say "cart" stuff, so here's the deal.

I have an LC browsenotes filter that says give me all the newly released 36 month D&E loans.
How many do you think is the max number of these loans at any one release? I can't remember ever seeing 25, much less 50.
At $25 a note and 50 notes max that's only $1,250 cash to cart them all.
At 10% APR that's a cash drag of only about $10 per month and I get first shot at every 36 month D&E loan released every time.
I can easily cart them all with a few point and clicks in less than 10 seconds. Then I can score them at my leisure, discard the ones I don't want and buy the ones I do.
Would you pay $10 per month for this privilege? Either a whole lot of people are doing this, or a whole lot of bots are.
Many D & E loans are already funded 50% to 95% when first seen in the loanBrowseloans API download (a full 20 seconds after I have carted mine).
There's no easy way to know if any loans are funded 100% and never even make it into the first API (or CSV) download file.
That's the story and that's why folks using automation see this timing advantage provided to web users and web bots as a very big deal indeed.
Title: Re: Loan Information Asymmetrical Timing
Post by: Kombinator on September 11, 2014, 09:35:31 AM
Rob, great explanation, thank you.  It clearly is a point to bring up, and an issue of asymmetrical information distribution to retail investors is not something a company wants to deal with as they get ready to price the IPO...if there was ever a good time to push LC to solve the issue this would really be it.  Perhaps making it more public may help them make the choice to put everyone on equal footing as opposed to just going with the flow...
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on September 11, 2014, 11:07:58 AM
...and an issue of asymmetrical information distribution to retail investors is not something a company wants to deal with...the choice to put everyone on equal footing as opposed to just going with the flow...

As I pointed out in the first post, the information distribution is equal. It's only the timing of the release by method that differs, and since everyone has access to every method, everyone is on equal footing. However, it really messes up our ability to automate and preferences manual buying. This makes less of a difference on the fractional side (although I and my subscribers have felt the volume decline some in the last few weeks) but makes a huge difference on the whole side (where it only takes one goofball to buy an entire loan).
Title: Re: Loan Information Asymmetrical Timing
Post by: Kombinator on September 12, 2014, 09:54:10 AM
The timing of information distribution is just as important as the content, think of all the reports issued and having the info 1 second ahead of the market...This is indeed much more of an issue on the whole loan side, but as you have pointed out even on the fractional side it presents problems.  It is quite puzzling why they would not address this, I can only hope that perhaps after the IPO and migration to the REST API they get their proverbial shit in order and just release everything simulatenously to manual, API and CSV file.
Title: Re: Loan Information Asymmetrical Timing
Post by: Rob L on September 12, 2014, 10:17:26 AM
If I remember correctly LC has publicly stated that this timing advantage is an intentional company policy to preserve a favourable experience for the manual user (or some such).
There's plenty of forum members that like it this way (present company not included).
Title: Re: Loan Information Asymmetrical Timing
Post by: Kombinator on September 15, 2014, 09:32:57 AM
Well, let us hope that this little disturbing quirk will go away ahead of the IPO and with the roll put of the new API.  I have full condfidence in the fact that the LC team wants to make a level playing field:)
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on September 15, 2014, 10:04:19 AM
I'm getting killed. Couldn't buy one loan in days via the API.
Title: Re: Loan Information Asymmetrical Timing
Post by: MarinBB on September 15, 2014, 10:18:29 AM
I'm getting killed. Couldn't buy one loan in days via the API.

Agreed, I added the API as a source of new listing information, but it's consistently getting beat by the website and the file.

Bryce, do you have any more info on the new REST API that is in development?
Title: Re: Loan Information Asymmetrical Timing
Post by: rawraw on September 15, 2014, 12:10:36 PM
Bryce, do you have any more info on the new REST API that is in development?
we do not speak his name
Title: Re: Loan Information Asymmetrical Timing
Post by: Joleran on September 16, 2014, 11:59:08 PM
The obvious problem is that as soon as the asymmetry goes away, manual buyers suddenly get fucked as hard as the API buyers are today, but worse and in reverse.  Today, a person frantically f5'ing the note listings can nab a good deal of high-desirability notes (that are funded in <5 seconds of listing) and filter through them at their leisure.  Tomorrow, this person will be 100ms behind a computer controlled program and will get nothing.
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on September 17, 2014, 11:45:49 AM
I think it's faulty thinking that investors _want_ to interact with the website in the long run. It's so tedious to log in 4 times a day and mash your F5 button. I've had to interrupt my day countless times to respond to my LC alarm. There is a non-trivial group of people who want automation but not LC's own automation. Right now that ability is totally broken and I'm going to need to start making withdrawal as cash piles up.
Title: Re: Loan Information Asymmetrical Timing
Post by: GS on September 17, 2014, 10:01:32 PM
Are we sure it's manual investors that are snatching up the new loans in 5 seconds, or could it be LC's Auto Investment users have an edge over everyone?
Title: Re: Loan Information Asymmetrical Timing
Post by: Randawl on September 17, 2014, 10:30:01 PM
Are we sure it's manual investors that are snatching up the new loans in 5 seconds, or could it be LC's Auto Investment users have an edge over everyone?

I can attest that this is not the case.  At least it is not the case for one particular account that has had it on since it allowed using of LC's (albeit lacking) filters.


Remember when saying manual investors that it also includes institutions that are manually investing as well.  Not sure if that distinction matters, just thought I might point it out.
Title: Re: Loan Information Asymmetrical Timing
Post by: brycemason on September 18, 2014, 12:40:13 AM
Certainly some of the loans I would buy are gone by the time the API sees them (they are 95% full according to the API download and by the time my order goes in it misses) or occasionally the API doesn't even see a particularly good one. Recent testing shows this isn't overwhelmingly the case.

Other reasons for the lower volume is that nothing is meeting my expectations (perhaps LC is setting rates closer to what my model suggests) or the overall type of loan (read quality according to my expectations) isn't available right now.
Title: Re: Loan Information Asymmetrical Timing
Post by: Lovinglifestyle on September 18, 2014, 01:30:28 PM

Other reasons for the lower volume is that nothing is meeting my expectations (perhaps LC is setting rates closer to what my model suggests) or the overall type of loan (read quality according to my expectations) isn't available right now.

Both of these reasons seem accurate to me.  My expectations are no longer being met by the LC rate settings, and what I am looking for is rarely available at any grade.  I've been getting the impression that the higher risk grades really are a higher risk now, as they fairly should be, imo.  I'm working on an attitude adjustment!
Title: Re: Loan Information Asymmetrical Timing
Post by: PeerSocialLending on September 23, 2014, 06:22:43 PM
For what it's worth I was able to invest in 57 notes (p2p picks top 25% CDE) between 2 release times today. I was a bit surprised to see them this high, but perhaps others are using top 10%?
Title: Re: Loan Information Asymmetrical Timing
Post by: Kombinator on September 24, 2014, 09:28:54 AM
Without a doubt it has gotten a lot better last week or so, hopefully this will continue at least till the IPO:)
Title: Re: Loan Information Asymmetrical Timing
Post by: Lovinglifestyle on September 24, 2014, 10:32:07 AM
If they aren't going to give back the not issued note money soon the inventory will grow even faster.  This is the first time I've been out of investment money (intentionally), so this is a new complaint for me.  I have $5, 275 of "Committed Cash", which I think is a personal record. 
Title: Loan Information Asymmetrical Timing
Post by: BruiserB on September 24, 2014, 12:24:21 PM
Lending Club's results happen when notes are issued, not when our cash is committed.

If people are seeing committed cash build up, then I would take that as a sign that LC has already issued their targeted amount of loans for the quarter and they are holding back issuances to give next quarter a jump start. 

If loans issuances continue normally and committed cash holds steady or drops through the end of the month, then I would see that as a sign that LC is having trouble meeting their numbers and wants to capture as much business in the quarter as possible.

Last quarter the slowdown in issuances started around the 24th of the month.


Sent from my iPhone using Tapatalk
Title: Loan Information Asymmetrical Timing
Post by: BruiserB on September 27, 2014, 11:14:21 PM
There's a post in the Lending Club borrower's forum where a borrower states they were fully funded and approved on 9/25 but told they will get their funds on 10/2. More evidence that LC is holding business till next quarter...they must have this quarter's numbers in the bag already.


Sent from my iPhone using Tapatalk
Title: Re: Loan Information Asymmetrical Timing
Post by: Joleran on September 28, 2014, 06:28:56 PM
There's a post in the Lending Club borrower's forum where a borrower states they were fully funded and approved on 9/25 but told they will get their funds on 10/2. More evidence that LC is holding business till next quarter...they must have this quarter's numbers in the bag already.

It will be interesting after the IPO if this behavior will continue - fresh issues throughout the last few days of a quarter could be a strong sell signal.
Title: Re: Loan Information Asymmetrical Timing
Post by: hoggy1 on September 28, 2014, 09:32:23 PM
I think we have wondered off topic here.
Title: Re: Loan Information Asymmetrical Timing
Post by: rawraw on September 29, 2014, 06:50:56 AM
There's a post in the Lending Club borrower's forum where a borrower states they were fully funded and approved on 9/25 but told they will get their funds on 10/2. More evidence that LC is holding business till next quarter...they must have this quarter's numbers in the bag already.

It will be interesting after the IPO if this behavior will continue - fresh issues throughout the last few days of a quarter could be a strong sell signal.
I don't know why everyone focuses on the IPO so much.  I doubt people will care.  It's not accounting fraud.
Title: Re: Loan Information Asymmetrical Timing
Post by: Joleran on October 02, 2014, 08:18:31 PM
I don't know why everyone focuses on the IPO so much.  I doubt people will care.  It's not accounting fraud.

The people that care are those who own LC stock.  If there's a reliable metric that gives insider information in an SEC-compliant fashion, this is a big deal.