Posted:
#1
I found a Pearson corelation of 0.753 between the over / under percentages and the ERA (earned run average) multiplied by the BA (batting average) for 16 teams, were over or the under percentage was above 0.55. That could mean the number obtained multiplying the team's BA by the team's ERA might be used a predictor for the over / under percentage. Of course my little study is not statistically significant but the correlation coeficient of 0.75 could be something. So I decided to put up a simple formula using those 2 numbers (BA and ERA) in order to predict the total of a game. I will test this formula on this thread because I think it cannot be backtested (the numbers are changing from one week to another and i think backtesting is not valid).
I will use only the batting average (BA) and the earned run average (ERA) for the last 10 games (since baseball is a "momentum" sport I think last 10 are more relevant than all games) and compute the total as bellow:
TOTAL ESTIMATE = (BAh + BAr) x (ERAh + ERAr) x 2,
where
BAh = batting avg for home team last 10 games
BAr = batting avg for road team last 10 games
ERAh = earned run avg for home team last 10 games
If my total estimate is 1.5 runs or more over or under the listed total I will take that bet.
Example:
Houston vs. Washington tonight:
Houston last 10: BAr = 0.237 ERAr = 3.60
Washington last 10: BAh = 0.261 ERAh = 4.19
Total estimate = (0.261 + 0.237) x (4.19 + 3.60) x 2 = 0.498 x 7.79 x 2 = 7.758
Total set by the bookie = 9.5, difference = 7.758 - 9.5 = -1.742,
I will take the under 9.5 for hou - was game.
GL Kemosabe......I'll be watchin' this.
Hot Forum Topics
