Saturday, July 20, 2013

Peer to Peer Lending - 1.5 Years Later

So I obviously did not keep up with the monthly peer to peer lending entries. Life, work, lack or interest, etc.  But I never stopped the investments. Well, that's not entirely true, I decided to wind down my Prosper account and stick with Lending Club as my primary source. Primarily just because I seem to understand LC better.

So where do things stand these days? Let's review!

Propser


Please note that the status chart above looks worse than it actually is since I've been withdrawing all payments from the platform rather than reinvesting. If you kept the charge off in proportion it should be ~1/2 the size it appears.


Payments received: $3,783.33
Principal paid off: - $2,661.69
Payments in excess of principal: = $1,121.64
Principal charge-offs: - $793.16
Gain/loss to date: = $328.48

Principal value of active notes: $2,245.15
Total active notes: 130
Current: 120
Past due (1-30 days): 6
Past due (31+ days): 3
Payoff in progress: 1
Total charged-off notes: 37
Total notes paid in full: 48
Total notes sold: 0

Yes, you read that correctly, 37 charged off notes. Generally speaking Prosper's default rate tends to trend higher than LC's. This isn't terribly surprising since Prosper accepts higher risk (and higher interest) loans. (For comparison of default rates, check out Nickel Steamroller's chart.) But even then this rate is more than twice as high as my LC account, so I should, perhaps, chalk this up to less sound portfolio filtering and strategy on my part.

That said, my return is ~5.65% and I've still made a vastly greater return than if I'd left this sitting in my ING (Well... now it's Capital One 360 -- really? Man that's a... name.) account. I'm fully aware that I tried Propser first and that it was a learning experience. If I'd try LC first this whole write up may have looked very different and maybe I'd be winding down LC instead. Even so, I plan on continuing to wind down the Prosper account and focusing solely on LC.

Lending Club

LC has been very good to me these last many months.


My return is ~11.35% as of writing this. I've had 16 notes charged off total. I've also taken common P2P investing advice and begun selling late notes on FolioFN for appropriate discounts to par. Selling late notes is critical, it has already improved my return by 1%, and considering that's often better than the entire yield of a 'high-interest' savings account these days, that's nothing to sneeze at. 

Analysis

A number of websites have come and gone for analyzing the P2P loan base. Lend Stats is perhaps one of the most well known that comes to mind. The fact that both Prosper and Lending Club transparently offer information on their entire loan base means that anyone can try their hands at data analysis. The new space that's popping up, however, are "premium" (read: paid) analysis shops. There's Interest Radar and Nickel Steamroller Premium. I've been using IR for the last month or so. The site leaves much to be desired in terms of eye candy and web 2.0 functionality, but I do like the solid filtering and in-house scoring systems.  I have not used NSP yet, but I plan on it. From the descriptions they put forward it seems that the scope of what they're trying to achieve probably exceeds that of IR and will do so at a lower price point.

By the above I mean that NSP sounds like they are going to aim for fully automated investing, active management (perhaps automatically selling notes? that'd be nice), etc. IR is more about giving you the analysis tools that LC refuses to. Such as very specific filters (sometimes you don't want 20-25, sometimes you want 23.5, try telling LC that) and custom filtering options.

Investments

I'm coming up on two years in the peer to peer space, I'll hit that mile stone come December. At this point, I'm thoroughly sold. I plan on funneling my small savings account into LC and getting fully invested. (I should note that this is not my primary savings account, I do have a well-rounded emergency fund, and I'm not committing fiscal lunacy.) The risk to reward factor here is too compelling to ignore. Yes, the default rates could get considerably worse should we hit another economic downturn (just look at the default rates for LC during the 2007-2008 period) and so, in that respect, it is not at all comparable to a classic idea of savings. But I am young and can afford to take on more risk, and this is not my only venue of investment. My stock returns often beat my P2P returns, but I think from a risk and diversification perspective P2P lending has a solid place in anyone's portfolio.

Saturday, May 11, 2013

AnyBackup 1.0.0 Released

It's been about a year since I last did anything with AnyBackup -- that's not because I lost interest, but rather because it'd simply been working and did everything I wanted it to.

I finally mustered up some motivation to make some improvements, though. I spent the last week straight overhauling the codebase. I unit tested all the major non-UI modules and while I was at it I refactored all the code for sanity and for PEP styling issues. This took way, way longer than adding new features (not surprising).

Since it's been a year I figured why not do this in style and make this a milestone 1.0.0 release.

Changes:

  • Unit tests (hopefully less bugs and less breaks in the future)
  • Display transfer rates and total copy progress
    • It's a little rough, and transfer rates are calculated in a very naive way, but it's better than nothing
  • Warn users when a backup is going to delete old files from any backup drives
  • Proper threading model
    • Previously threads were killed on demand via a nasty kill/hack and it did not ensure consistent data, etc
    • New model nicely requests thread death and long running thread processes have been refactored to comply
  • Code clean up, PEP style conventions
  • Migrate from Pyro3 -> Pyro4
    • Pyro4 has async calls which is needed for the proper threading model mentioned above

Wednesday, May 1, 2013

Metro 2034 - English Fan Translation

Word on the street (or rather the Last Light forums) is that there's going to be an official Metro 2034 English translation released. Well... the rumor has it that the translation was already supposed to be released. Obviously that hasn't happened yet. This rumor began circling unbeknownst to me, otherwise I probably wouldn't have bothered finishing what I'm about to post here.

But without further adieu, I give you a cleaned up version of Metro2033Artjom's fan translation of Metro 2034!

There are some pretty big caveats:

  • It's no more accurate than the original fan translation (probably a good deal less so)
  • There were a few sentences I couldn't make heads or tails of so I scrapped them
  • I may have butchered a passage here and there thatI thought I understood
  • It's probably been colored/influenced/altered significantly by my style and humor
With that warning out of the way (and if you're still reading) I'll get to the good bits. I cut out a staggering amount of awkward phrasing and confused verbiage. (And probably introduced a few of my own, but shhhh... were't not talking about that!) I tried to throw in common (American) English idioms where I could and rearranged sentence structures that tweaked my brain. I may have had a little too much fun with Leonid's dialog in places, but what can I say? I like the character.

If you think my copy is crap (hey it's a free internet... or not, depending on your country) feel free to peruse the original fan translation here. I take no credit for this original translation, just the subsequent editing and interpretation I provide here. It was translated to English from the official German translation. If you're wondering what that can do to literature... I'll let this video do the talking:



Anyway, as I'm sure you've skipped everything above, here are the download links:


Let's all hope they actually do release an official English translation. I've got my fingers crossed!

P.S. To any lawyery-types. If there are copyright concerns here, please contact me, it was not my intention.

Friday, June 8, 2012

AnyBackup 0.9.6 Released

Hot on the heels of 0.9.5, I've released 0.9.6 to address some immediate deficiencies resulting from the major overhaul of the underlying application.
  • File cache layer between SQLite and AnyBackup
    • This addresses the lag the can occur when browsing an indexed drive, especially when you have > 100k files across all indexed drives
  • Improved search dialog making use of the new SQLite backend
    • Search by drive
    • Specify search type: Contains, Ends with, Starts With, or pass in a SQL string (i.e. %.mp3)
    • Specify file type (directory or file)
    • Search dialog is persisted so your input is saved (unless a drive is added or removed -- then the dialog is destroyed and recreated)
  • Clicking on a result file no longer automatically displays it in the content pane, instead this has been moved to a popup menu option
  • Changed remote indexing port to not conflict with standard XVNC port
AnyBackup 0.9.6 -- As always please note that these are all items I own or have bought in one form or another -- don't sue me. :)
The new search dialog in 0.9.6

The new 'Show In Content Panel' option

Tuesday, June 5, 2012

AnyBackup 0.9.5 Released

I've made a lot of changes since 0.9.4 in November last year. I put 0.9.5 up for download tonight, enjoy!

Major changes:

  • SQLite
    • Super fast
    • Less memory usage
    • Comparable disk usage
    • More flexible
    • Externally accessible
  • Many minor bug fixes
  • Complete / consistent icon pack
  • Backup sets
  • Thread crash detection
  • Logging to homedir so may work on UAC enabled machines
    • Super strict UAC settings seem to deny access to network drives, so if you use network drives like me you still might have issues

Not that I've heard from too many people on the app before, but as always, if you hit issues, don't hesitate to let me know or file an issue on the Google code page.

Sunday, June 3, 2012

AnyBackup Upcoming Release

For anyone out there that uses AnyBackup, if anyone out there does beside myself and family -- there have been over a thousand downloads but that says nothing about actual usage, I've not abandoned the project. I just haven't had time recently to do much. I spent a lot of time this weekend, however, revamping it. Some of the biggest changes are under the hood, I've totally changed how data is stored which necessitated moving and rewriting a great deal of code.

Major Changes:

  • Moved to SQLlite as a backend
    • Generally more performant than the proprietary mishmash of objects currently in play
    • Doesn't rely on pickle
    • More flexible
    • Data can be viewed / accessed outside of AnyBackup
      • Probably not super useful for most people but if you had a reason to want to get at this data, it's now very easy
      • It's also very useful for troubleshooting
  • Click & Drag functionality from AnyBackup
    • You can now click and drag files displayed in AnyBackup and drop them in folders, drives, your desktop, etc
      • The above is only true if the files reside on a drive that is connected, of course
  • Right click or double click on files to show in explorer
    • If you double click on file x.txt AnyBackup will launch an explorer instance with x.txt highlighted
  • Modified Time is now displayed in file tables in addition to size and name
  • Search now uses sql syntax instead of regex (necessitated by the switch to SQLlite)
  • Sets now supported
    • This was a feature that really didn't mesh well with the legacy code base so I hadn't gotten much further despite taking a few different approaches
    • Each set is completely independent of any other
    • You can have multiple identical sets if desired
      • i.e. if you want to do rolling backups with different sets of backup drives against the same content drive(s)
    • There's a new toolbar which lets you switch between sets
  • A few optimizations here or there to avoid slowing down on SQLlite queries
  • Thread crash detection
    • Previously thread handling really sucked (it's still pretty messy) but it was so bad a thread could crash and the gui wouldn't recognize it thus leaving the gui continuously displaying the scrolling progress dialog as if the process were still running
    • When a thread crashes now this is recognized and the error is popped up in a dialog box
Look for this revamped version to be up for download in the next week or so.

Thursday, April 5, 2012

Peer To Peer Lending - Month 3

This is my third monthly report for my peer to peer lending.

In terms of late and defaulting notes, Lending Club is, quite bluntly, kicking Propser's ass. Now I should point out that it's not exactly an apples to apples comparison as I have some riskier notes in my Prosper portfolio compared to Lending Club. That said, the notes that are likely going to default on Prosper would match my tighter filters on Lending Club. Let's review!

Prosper:
Note the build up of late notes!




Payments received:$693.89
Principal paid off:-$480.55
Payments in excess of principal:=$213.34
Principal charge-offs:-$0.00
Gain/loss to date:=$213.34


Principal value of active notes:$5,119.45
Total active notes:203 
   Current:196 
   Past due (1-30 days):
   Past due (31+ days):
   Payoff in progress:
Total charged-off notes:
Total notes paid in full:
Total notes sold:0  

Account value as of writing: $5,216.84


So how are we doing? Well, I've got four notes now that are 31+ days late. I'd be very surprised if any of them pay off. But surely all these notes that are delinquent must be high risk, right? Not really, and that's the rub.

The break down of these four notes: 1 B, 1 D, 1 E, and 1 HR. None of the notes had an delinquencies in the last seven years. One of the notes has a public record in the last ten years (which I've subsequently removed in my filter now). One common thread between these notes is that they all have 1 or more inquiries in the last six months. I've brought my Prosper filter inline with my Lending Club filter now and cut out borrowers with inquiries. This probably cuts out a portion of good notes, no doubt, but at the same time the largest chunk of loans are for consolidation and it makes a simple kind of sense that a good candidate for a consolidation loan should be looking to pay down their current credit lines, not take out new lines. (Yes, I realize things other than a credit line can register inquiries -- as I said, some good notes will get cut out with this broad sweep.) That said I'm only looking to reinvest payments, so I don't need a large volume of loans and can afford to be picky.

I fully expect these four loans to get charged off, they represent $150 in principal outstanding minus a small amount of payments received = $147.67 projected charge off. If this comes to pass my profit will be knocked down to below what my Lending Club account has brought in with less time under its belt.

Lending Club:




Account value as of writing: $5,136.79


My Lending Club account is doing great. All my notes are current. I have not had one note where a borrower has taken the money and walked away, which has happened to at least three notes on my Prosper account. If Lending Club offered an autoinvest option (based on your custom filters and with < $25k invested) I'd say there was no compelling reason to use Prosper based on my experience so far. Even so It's likely that even with a three week handicap losses will mean that my Lending Club account ends up coming out ahead of my Prosper account in earnings.


See the first update.See the second update.

Followers