I discovered a few blips in the system over the weekend when I was doing a small audit of the math - the Sharpe was using an improper array, and the Moneyline was reading from the wrong module. Neither of these things proved earth shattering for the week or the historical results, it was a matter of a few percentage points here or there.
What I have discovered, though, is that parlays are terrible. Parlays are like leverage in a hedge fund - they have the ability to enhance the returns by putting very little down, but the "very little down" can add up if the returns aren't there. The payouts are large for a reason: they are near impossible to predict. I've tried several permutations, and had little to no success on a consistent basis. And yet, when I do the historical review, had I chosen my top 5 spread picks ONLY, I'd be up huge this year.
Despite my discipline in terms of analyzing data, I can't seem to separate it into placing the bets. The parlays are fun when they hit. I'm going to spend some serious time this week exploring the parlay potential of each module and the combinations. Maybe if I combine Sharpe, stats, and rank by beta it'll work? Maybe the alpha and Sharpe is all I need? Also, I need to develop a teaser module - basically, I need to mirror the entire system and add the teaser points to each team to see what the best bets are. I don't know if they're any different, but we'll see.
Since I have no money on the game tonight, here are this week's results:
|Results for Week 12|
|Moneyline: SEA (-160) v. STL - $19.05 wins $30.96|
|Moneyline: CLE (-170) v. HOU - $5.00 wins $7.94|
|Spread: SF +10 (-110) v. ARI - $11.14 wins $21.27|
|Spread: MIN +7.5 (-120) v. NYG - $7.07 wins $12.96|
|Spread: STL +3 (-115) v. SEA - $11.48 wins $0.00|
|Spread: BUF +7.5 (-110) v. JAC - $6.25 wins $0.00|
|Spread: DET +3 (110) v. GB - $5.37 wins $0.00|
|Parlay: STL +3 (-115) v. SEA, CAR +3 (-120) v. NO, DET +3 (110) v. GB - $5.00 wins $0.00|
|Parlay: STL +9 (-110) v. SEA, WAS +9 (105) v. TB, BAL +15 (-110) v. SD, DET +9 (105) v. GB, PHI +28.5 (-110) v. NE, ATL +18.5 (-110) v. IND - $5.00 wins $0.00|
|Results for Week 12|
|%s: Week 12||-2.96%|
|Box Results - through Week 12|
Notice the teaser at the bottom of the bets that wasn't there in earlier posts. Yeah, that's called "5 minutes to game time, let's get crazy!" betting. I developed a slapdash teaser methodology that I didn't backtest that I decided to just go ahead an bet. Without that bet (or the parlay), I would have had a nicer week. Unfortunately, I am a moron.