Wednesday, January 1, 2014

Turnover Index: Annual Letter to Shareholders

To the Shareholders of Turnover Indices Incorporated:

2013 was a disappointing year for Turnover Indices Incorporated (TII). After posting moderate gains in its first year of existence, the fund suffered significant losses in Week 4 of the 2013 season. Gains were modestly positive in subsequent weeks, but not enough to offset that initial misstep. See below for a weekly summary of fund performance:


weekbetswonstarting
bankroll
amount betprofitsending
bankroll
4 2 0 $1,000 $268 (26.8%) ($268) $731
6 1 1 $731 $32 (4.5%) $29 $761
7 3 3 $761 $130 (17.1%) $118 $879
8 3 1 $879 $70 (8.1%) ($12) $867
9 7 2 $867 $166 (19.2%) ($66) $800
10 4 1.5 $800 $95 (11.9%) ($45) $755
11 2 0 $755 $14 (1.9%) ($14) $741
12 1 1 $741 $37 (5.1%) $34 $775
13 4 2 $775 $46 (5.9%) $21 $797
15 1 1 $797 $18 (2.3%) $16 $814
16 2 1 $814 $11 (1.5%) $2 $816



In my review of last year's fund results, I speculated that the TII model may be placing too many bets late in the season, using its simple criteria of at least a ten turnover differential between the competing teams. By week 14, a 10 turnover differential may not be enough to skew the betting market's perception of team strength. To address this shortcoming, I laid out a new approach in this season's investor prospectus, which used per game turnover differential as the criterion for betting.

I also switched to a more sophisticated method of calibrating bet size, using the famed Kelly Criterion (aka Fortune's Formula) to increase bet size when the expected edge against the house is greater. Because fancier math always results in greater returns. Always.

Annual "Special Pleading" Section

Unfortunately, it appears this new methodology may have overcorrected its late season bias in favor of an early season one. The two losing Week 4 bets were based on the assumption that there was a ~59% probability of covering the spread, leading to a rather large percentage of the fund being invested (~27%). The first losing bet was Texans over the Seahawks, which was looking like a very smart investment, until Matt Schaub threw the quarterback's version of a hanging curve ball to the Seahawks' Richard Sherman.

The other week 4 bet was Steelers over the Vikings. This bet also looked promising late in the fourth quarter. Pittsburgh appeared to be on the verge of sending the game to overtime with a touchdown, having set up first and goal with 26 seconds left in regulation. Promising that is, until Ben Roethlisberger fumbled the ball away on third down. In short, this is all Ben Roethlisberger's fault. Screw that guy.

Week 4 results aside, we are modestly encouraged by the results using the Kelly criterion. From week 6 to 16, overall return on investment was a respectable 7.2%, even though the bets that were laid went 13-16-1 against the spread. This means that the model did a decent job of investing more money on teams with a greater probability of covering, and dialing back the investment when the probability of covering was smaller.

Corrective Action Plan

Rest assured that executive management here at TII has faced severe disciplinary action as a result of this season's disappointing performance (including, but not limited to, public humiliation and corporal punishment). We currently have an army of analysts crunching numbers, running code, browsing reddit when we're not watching, etc. All in an effort to return the fund to profitability. 

And we have an even larger group of analysts and marketers crafting ready-made excuses for why next year's fund performance fell below expectations. I don't want to be typing one of these damn things on New Years Day again (we had a hell of a New Years Eve party, courtesy of shareholder funds - I am quite hungover).

Sincerely,
Mike Beuoy
Assistant Janitor in Training
Turnover Indices Incorporated

Post Script

Weak attempts at humor aside, my initial thought is that week 4 may be too soon to start laying bets using this strategy, so next year's version will most likely start with week 5. In the meantime, I'll see if there is a less kludge-y way of addressing the early season issue.

3 comments:

  1. Seems like my comment about short and long runs still stands.

    I'm curious -- is it easy to check if the bets go profitable if you go to 1/2 or 1/10 Kelly betting instead?

    ReplyDelete
    Replies
    1. I ran a quick check. Fractional Kelly (either at 50% or 10%) would have just mitigated losses, but not eliminated them. The optimal fraction was 0%, unfortunately.

      Delete
  2. It was disappointing that the TO-Index didn't succeed, this year. Either that is due to yearly variance, or maybe the market corrected itself. I would like to see another year of this, and if possible, college football (the market isn't as sharp in college as it is in the NFL).

    ReplyDelete