csewell1987 wrote: ↑Thu Jan 29, 2026 2:16 pm
Sounds like you have a similar sort of approach to me, except I chose the attributes. I use past form against todays race setup with a few other pointers to assign probabilities, and then price these up to a 95% book. I focus on the top race meetings only though as these are where form stands up best, which obviously has its limitations and requires higher staking. I have found without incorporating market data its best to get on as early as possible, which again has its limitations. Night before often not an option but morning of either on exchange or bookmakers (in shop) if liquidity isnt there. If you get closer to the race, without factoring market in, you often find market moves have taken you away from one's you have rated highest and would have previously backed. Obviously could go both ways of course but form tends to be used by tipsters and punters so often are subjected to moves.
Have you looked at Betfair historical data? The free data gives you 1 minute tick data, so it’s possible to check for compression and best entry points. I found it was mixed when I tested though there are an endless number of simulations to run.