I have collected betting data on horse racing. My aim is to maximize my P&L.
I have been placing bets on 14 horses. The horses do not race in the same races.
There are variables that are active in the bets. Sometimes only a few variables are active and other times more are active. I am trying to establish the relationship between the variables and P&L. I think the analysis will find that certain variables, or combinations of certain variables, effect the likelihood of a profitable outcome.
The study may find that when certain variables are active, the results are worse. This would also help me because I would know to avoid placing bets when those variables are active, therefore minimizing losses and increasing my P&L.
I have placed around 1000 bets on 14 horses, and there are around 36 variables on each bet. Most of the time the bet only used around 10 of the 36 variables. I am aware the number of vairables : number of bets ratio is not ideal. However, there should still be enough data to draw some conclusions at the least.
I have attached a small version of the file so you can see how I have recorded the bets. If the characteristic/variable is true for the bet it is marked in the cell with a "1"
Do you think you can help me find if there is a relationship between a certain combination of bet variables and profitability, on each horse?
You may find there are variables that work best on every horse, and you may also find that each horse has an optimum set of variables. I do not know and would love to find out.
Thanks for you help!
Notes: Columns "V and W" are specific to each horse so will be exempt from the variable correlation study. I would like to see [separately] if there is a correlation between profitability and the P1 number being above OR below it's median for each time period. I also want to do the same but use the mean number instead of median, and compare the two sets or results.