Firstly thanks for the advise given. I have given it the respect it demands. I was off in another paddock. I have finally analysed my data (43000 races) and made 11 charts each with 9 price ranges. This will be the base for future testing. Each is set up so that I can test laying for liability, variable liability, for a target amount or a varying target amount. I can now test on individual favorite, some favorites or all favorites. On the first chart I played with the variables and found that laying some of the field to a fixed even liability worked best. I made the remaining spreadsheets set to this model and found some anomalies. In some price ranges there were major losses in select field sizes and little niches which gave excellent profit. I found that by dumping 17% of the races I can make an acceptable profit on turnover with a simple basic system run on BBP. By dumping another 5% of races more profit could be gained but that would mean 3 different systems to be run concurrently.
Beton