This is over 8 months (unfortunately my db with betair prices became corrupted before this point).

The methods use ratings which I have devised myself.

The ratings are based on horses who have run against each other previously, or whose form lines are linked through mulitple entry points from past races. For example if horse 1 raced horse 2 last start and Horse 3 raced horse 4 last start (in a different race) but sometime previously horse 2 raced horse 3 then I can produce a rating figure. I go many more levels deeper then that though (up to 10 levels of form comparison) and take into account a lot of factors. Its not something easily reproduceable by 95% of punters fortunately for me.

But enough of that ... I was interested in what peopl would take. Personally I take the 300 selections (ie the extra filters). The reasoning is as follows (feel free to disagree):

1. My risk tolerance is very low. I want to be 99% confident the horse I bet on will win and 99% confident the horse I lay will lose.

2. Methods/filters alwas tend to revert to some historic mean which is not apparent to us at the point we start betting. I would say they an swing by up to 10% in some instances. 300 bets is my starting criteria for a laying method and it needs to have less then 2 winners under $15. This gives me confident that even if its true strike rate was higher it would still be profitable.

It takes a lot of work to find these type of lay/bet systems and I don't think many people would put in the time necessary.