This is something that I have been planning to do for a while now, public festival bets that is. While you can pay for some festival tips, for example Betfan is running a festival special for Chester, I still prefer trying to do it myself.
Anyway, for the first day I only have two bets, both 1pt win bets.
Chester 15:45 Mappin Time
Chester 17:25 Secateur
Both are result of my bayesian system, and while the system rates all of the days races and I could present a selection to each race, I do like to add a bit of human touch, namely to check that data looks consistent and that I can stand by the results of my might algorithm.
Guess this won’t be weekly review after all and after giving this a bit more thought I came to conclusion that monthly is better way to go.
Anyway, from the table below you can see results for bets for the month of April.
| ||All selections||Adjusted flat stakes||Adjusted variable stakes
Overall decent results and atleast for this sample it seems that my idea is holding true as return on turnover is best with my adjusted stakes and even to level stakes return is better when compared to just flat betting all selections.
But, one thing that I only now checked was what if I just adjusted stakes and didn’t filter the bottom end out (currently I drop out bottom of the adjusted stake so that four point bet becomes one point bet and anything else is discarded). Bets I do filter out are just about break even so it does make sense to drop them out, but what would have happened if I had just adjusted the stakes? Well, profits would have doubled but total stakes would have grown four fold at the same time so return on turnover was just little over 6%. Based on this I will move on to May without changing anything.