So, upon advice of Capper, I altered the stats on which the financials are based, and now the results make a lot more sense. Now, the financials are figured thusly:
The beta is now figured using a ratio of points scored to statistical score - a scaled and redesigned version of DVOA from the geniusi at Football Outsiders. Basically, it's a ratio of points to team value over average. As an example, if NE is 40% above the average team (which scales to a score of 140), and they score 37 points, the ratio is 37/140 = 0.22. The baseline score is a score of 20 (the historical average) divided by 100 (the most average team possible), or 0.20. What that means is that, if NE scored the exact amount of points it SHOULD score, it would score 28 points. Because it scored 37, it has a higher point to score ratio.
This is a very effective way of describing whether or not a team is performing as expected or not. So when I figure the all the financials using these ratios, the rankings now look like this:
Sharpe | Alpha | Beta | |
Team | Rank | Rank | Rank |
ARI | 18 | 22 | 24 |
ATL | 30 | 28 | 27 |
BAL | 26 | 29 | 21 |
BUF | 27 | 25 | 17 |
CAR | 24 | 26 | 16 |
CHI | 23 | 17 | 23 |
CIN | 9 | 10 | 8 |
CLE | 4 | 2 | 30 |
DAL | 2 | 3 | 3 |
DEN | 29 | 30 | 4 |
DET | 7 | 6 | 18 |
GB | 8 | 13 | 20 |
HOU | 11 | 7 | 5 |
IND | 3 | 4 | 9 |
JAC | 19 | 23 | 2 |
KC | 28 | 27 | 19 |
MIA | 15 | 14 | 28 |
MIN | 14 | 21 | 29 |
NE | 1 | 1 | 14 |
NO | 12 | 9 | 25 |
NYG | 6 | 5 | 12 |
NYJ | 25 | 8 | 32 |
OAK | 20 | 15 | 26 |
PHI | 17 | 20 | 31 |
PIT | 5 | 12 | 11 |
SD | 10 | 11 | 6 |
SEA | 13 | 16 | 13 |
SF | 31 | 31 | 1 |
STL | 32 | 32 | 10 |
TB | 22 | 19 | 22 |
TEN | 16 | 18 | 15 |
WAS | 21 | 24 | 7 |
Almost the same as before, really. But now, I figured out HOW to use these scores. Or, more accurately, my lovely wife gave me the idea.
She suggested, "Why do you have to tell the Box how many games you want to pick? Shouldn't the Box tell you? Shouldn't there be a threshold?" I don't know why I hadn't thought of it, I thought that could be genius. But rather than letting the Box tell me what games to bet, (since that means there would be weeks I only bet one game, sometimes zero, and that's no fun) I'm having the Box now tell me what games to DOUBLE the bet. The beta does an AMAZING job of gaging volatility for spread bets, and a Sharpe/Alpha average picks moneyline winners like I pick my nose.
I'll get more into this during the week, but for now, let's just say I would have made FIFTEEN TIMES MORE on my bets if I'd been doing it all along. Pure semi genius.
Check back for results later today/tomorrow, it was another winning week! It was a little ugly, but still a winner!
1 comment:
I'm not following your logic on the Ratios.
1. Are the Sharpe and Alphas still based on scoring.... or did you change it to stats vs. the average team?
2. Beta is a measure of volatility... but it seems like you are projecting scoring with this ratio. Shouldn't this be used to gauge confidence that a team performs consistently rather than how many points you think they'll score?
3. How are you taking the ratios into account when you make your picks? It would seem that since the alpha/sharpe take into account performance and volatility, you should simply just pick the higher rated team in every instance. Is this what you are doing?
Post a Comment