P2P Lending / NFT Lending Forum

Over the last 5-6 months I've been working with my brother on a note selection tool to use for our own investing on Lending Club.  We decided to make the site public so others can use it too and we launched it yesterday.

PeerToPeerQuant.com

I would love to hear feedback.

Its free to make an account, free to see notes we recommend, and first 25 clicks on the notes are free ($0.15 after that).

We use a genetic algorithm to come up with our selection criteria and so far it has been working really well (15.98% ANAR and 11.4% XIRR).  We only have 3 months of data but we outperform most other accounts at that age and risk level.  Obviously we will need more time to know how it does in the long run.

Info on Methodology and Genetic Algorithm
Screenshots of results

Happy to see new players appear.  Wish you luck.

Average age of portfolio is very short for performance measurement purposes. 

from: Fred93 on October 24, 2014, 05:18:32 PM

from: turing on October 24, 2014, 01:56:24 PM

Are you actually charging retail investors?  If so, there are some legal hoops you have to jump through otherwise it is illegal.  Be sure to take care of that, if it isn't taken care of already.

from: rawraw on October 25, 2014, 06:34:05 AM

from: lascott on October 24, 2014, 07:17:03 PM

from: turing on October 25, 2014, 08:35:04 AM

I realized there is another issue if folks use other methods to pick notes already (IR, LR, LC, BV, BV+P2PPicks, NSR, etc).  What if they want to make sure they don't spend $0.15 to get a duplicate note that they already purchased?

In one theory, you could say that if I picked a note for $25 via BlueVestment+P2PPicksModel and if P2PQ picks in then it must be a really "good" note and may want to double down ($50 as $25 for each systems pick).

At least a few of these allow you to specify you don't want to buy already owned notes. (I'm pretty sure BV allows this stipulation).  But I don't know if it always works in the feeding frenzy of new notes being issued. I suppose more than one service could put the same note in your shopping cart before actually buying it...and at that time both services wouldn't think you already own the note. But like you say, I wouldn't be too upset if I ended up with a few doubles because two of my rules/services picked it.


Sent from my iPhone using Tapatalk

from: lascott on October 25, 2014, 12:33:50 PM

What time period was used for analysis?  Was there out of sample testing?  How often are variables checked for changing predictive power?  Is the LC grade a variable?

from: rawraw on October 27, 2014, 12:08:51 PM

So you're using LC grade, although they continually and unannounced change it. And when you say all notes, how can there be out of sample testing?

from: rawraw on October 27, 2014, 01:21:09 PM

It's okay, I'm typing from a cell phone so my responses may seem curt. What's your value proposition?

from: rawraw on October 27, 2014, 02:45:09 PM

I guess what I'm getting at is do you think you improved on what's out there already? Or just introducing price competition

I get your question now.  I think we do offer some benefits vs other models.

The biggest benefit is our model is not speed dependent.  There are lots of notes that fit the qualifications at all times of day.  People don't need to worry so much about API's and 3rd party services to make sure they get loans within first 5 minutes of being listed.

The other models out there are not inherently speed dependent but they have become a victim of their own success because so many people/institutions use them.  It makes it harder and harder to win the race to buy their recommendations.

Do you have the option to sort by probability of default?  That's what I'm really looking for in a model.  I could care less about the expected returns of the models that use only one part of a credit cycle but I am interested in calculating the probability of default and determining which loans I'd like to invest.  I'm not trying to maximize income for tomorrow.  I'm trying to hit a certain return benchmark with notes that I think won't be as volatile.

from: turing on October 27, 2014, 05:19:10 PM

Does the option of picking a particular note disappear once you've selected it once?
i.e. if I picked that 13.98% 36 month note mid-morning and went looking for more notes mid-afternoon would it still be there if it is was not 100% funded?  That is could I mistake it for a new note and pick it a 2nd (unwanted) time?

from: rawraw on October 27, 2014, 07:07:44 PM

from: lascott on October 28, 2014, 01:38:50 AM

Besides using the 1 month old results, do you release what the net return would be on already seasoned notes in the historical file?  Anyone can be in the top 99th percentile in the first month.

How often (when) are the notes listed updated?

from: lascott on October 30, 2014, 10:02:35 AM

No it's not.  All you have to do is invest in the highest interest rate notes possible and invest in nothing else.  As long as they all don't turn out to be first payment defaults, you'd be at the very top even of liked age accounts.  But I'm curious what the long term results are, so I'll keep posted

BTW you may want to request a subforum for your product.

from: rawraw on October 30, 2014, 07:05:27 PM

Quote"> from: turing on October 31, 2014, 11:35:03 AM

from: PeerSocialLending on November 02, 2014, 11:55:00 AM

Quote"> from: rawraw on October 31, 2014, 10:37:17 AM

Thanks Zach.

Ran across this old blog while searching for something else.  Thought those following this thread my find it interesting.

http://blog.dmpatierno.com/post/3161338411/lending-club-genetic-algorithm

Quote

Good find on that link.  I'm going to reach out to him to talk shop a bit to hear more about the methods he used and how he combated overfitting.

The biggest hurdle we found with genetic algorithms is making sure you build in methods to combat overfitting (regularization, cross sample testing, early stopping, etc).  Other wise the model starts predicting noise instead of predicting the underlying trend.

Thanks for sharing that link.

Pretty sure Bryce went head to head with that guy in a backtest white paper.  He took the posted strategy and backtested it compared to his

from: rawraw on December 25, 2014, 04:59:52 AM

from: turing on December 26, 2014, 10:41:37 AM

I included that genetic selection criteria in my portfolio comparison white paper. However, it wasn't a terribly fair inclusion as it didn't pick a ton of notes and the backtest period overlapped somewhat with the original author's testing period. I wound up taking it out in a later version. I also thought the best comparison for a product like mine was something that would pick at least 10% of 36 month loans (in the appropriate grades). The method of comparison was good, though, and I haven't seen anything like that since.

from: brycemason on December 28, 2014, 10:39:06 PM

from: turing on December 29, 2014, 01:22:29 PM

Thank you lascott.  Some great reading material for me to go over. 

I'll send you a copy, just shoot me an email.

from: brycemason on December 31, 2014, 03:10:03 AM