
Originally Posted by
BettorBob
It's actually quite a bit more complicated than that...
For a given play (example: 2009-08-15 Cardinals Moneyline), I first look at all the handicappers who are on that particular play. I knock off all the handicappers who have 20 or fewer previous picks, and all handicappers with a negative ROI. This leaves me with only the handicappers who appear to be making money not simply through dumb luck.
For each of those handicappers, I then come up with an "experience rating", which I calculate by counting the number of picks he's made, subtracting 20, and then taking the fourth root (that is, x ^ .25). So, if a capper has 25 picks, his experience rating is (25 - 20) ^ .25 = 1.49. If he has 100 picks, his experience rating is 2.99. If he has 1000 picks, his experience rating is 5.6. The more picks he has, the greater his experience rating, but the rate at which the rating increases, decreases.
Next, I look at the number of units that the handicapper is putting on this particular play. Since he has already released more than 20 picks, I should have a pretty good idea of his average bet size, so I divide the number of units he's wagering by his average unit size. Therefore, if this bet is bigger than his average, then the result will be greater than one; if it is below his average, it will be less than 1.
I then multiply this number times his experience rating times his ROI. (I calculate his ROI as 100 * "units earned" / "units risked" - 100.) "Units earned" means all the money he has received back from his winning wagers (risk + profit).
Finally, I divide the resulting number by 10, and round it off to the nearest integer. That number ends up being the unit size for that particular capper. I then add up the unit sizes from all the other handicappers and end up with a total unit size for the play.
Though I will throw the whole bet away if I see that another handicapper with more than 20 picks and a positive ROI has a pick on the opposite side. In cases where the bet is on the ML or RL, I kill the bet if there's another bet on either the ML or RL of the other team (to avoid a situation where we might have, for example, a RL bet on the Cardinals but a ML bet on the Padres).
I plan to use this same strategy on football and basketball as well. When moving to another sport, I plan to only look at the cappers' records for that particular league. For example, if Wonderdog proves he is good at capping baseball, he will have to also prove he is good at capping NFL before his picks are calculated into the NFL consensus. Similarly, even if Anthony Redd proves he is good at capping college hoops, his record won't spill over into the NBA -- he will need to become profitable there as well (which, if you followed him earlier this year, you know he isn't). For football, though, I may need to lower the number from 20 to like 10, because there aren't very many football games compared to other sports.