/* ]]> */
Oct 312013

This year WinThatPool! is participating in NFLstatheads’ “Battle of the Bots” on reddit.

One of the things I like best about this competition is it explicitly measures WinProbability accuracy by applying an economic cost model. The results reflect a model’s accuracy better than a simple count of correct versus incorrect or a confidence pool point total. The lower the “cost,” the better.

Yesterday, a post updated the YTD rankings through eight weeks.

Follow the link, scroll to the bottom of the post and you’ll see the table below listing the year-to-date rankings:

Here are the standings after week 8 (only including those who had a week 8 entry):

Entrant Total Cost
WinThatPool – Human 22.352
Philosotoaster – Bot (Bot1) 23.327
steve496 – Bot 23.514
Rufus_Reddit – Human (Vegas Moneyline) 23.538
pudds – Bot (alg6) 23.562
Rufus_Reddit – Bot (Simple Modified Pythagoran) 23.63
Rufus_Reddit – Bot 23.859
Rufus_Reddit – Bot (Fancy Modified Pythagoran) 23.951
UncountablyFinite – Bot (Original Bot) 24.06
Wallamaru – Human 24.1
Rufus_Reddit – Bot (Gamescript Linear Model B) 24.105
pudds – Bot (alg7) 24.157
UncountablyFinite – Bot (Bot 2.0) 24.2
Philosotoaster – Human 24.547
Philosotoaster – Bot (Bot2) 25.38
k_Bomb – Human 25.981
redditcdnfanguy – Bot 26.521
BuckNewman – Human 26.855

Note that this competition didn’t begin until around Week 3 or 4 of the season, so these results are not as close to half a season’s worth of picks as you might think.

Don’t know how the rankings will hold up over the next nine weeks, but so far, so good!

 Posted by on October 31, 2013 at 12:10 pm

  No Responses to “WinThatPool! top NFL “bot” through week 8”