Loan to Listing Corelation — Ongoing Debate. | P2P Lending, Peer to Peer Lending, People to People Lending

That is exactly my point. Other people are already doing it, just not as blatant as my example. Eric looks up a lender’s loans and shows the status in the context of evaluating a lender’s performance.

Prosper gave me the data and I can tell you for all loans in the data(6393) I can identify the listings of 6335. 6335/6393 = 99.1%. If all the data they gave me is truly public, then it seems like Prosper has a fundamental flaw and a problem on it’s hands. I know I am a lender and must abide by TOS, but the data download is for anyone. Including others not bound by a TOS.

However, without an guaranteed unique matching criteria the limit as loan volume increases of the percentage of loans that can be identified is 0%. But right now I think it is over 98%.

What I am doing is not rocket science, but I did go to some effort to get almost all the loans.

And I am holding off publishing the sql, let alone the page of late loans to listings to give prosper a chance to respond. If prosper wants to see the sql I will be happy to show it to them, but my guess is they already have it.

On the other hand the data we currently get to evaluate loans is significantly less and wholly inadequate when compared to the institutions that normally lend money. We are all in effect gambling. The future default rate is an estimate at best, but probably more like a flat out guess across any data segment (let alone the extended credit most of our SO criteria fire on.)

I started a thread on the prosper forums to clarify an issue: Loan to Listing Sql, SQL to join loan to listing. I like my last post in the thread and am putting a copy on my blog for prosperity, please read the entire thread if you need more context…

Leave a Reply

Your email address will not be published. Required fields are marked *