Enjay,

I hate to say it but I think there's a major glitch in your system. Your first post showed apparent results for the last 12 months and said 343 races and 417 selections. In another post you said you ran the results back to 9/99 yet the races were only 315 and the selections, actually 370, less than the 12 month period. This doesn't make sense???? If there were 417 selections for 12 months there should have been about tripple that for 3 years. The results also look a bit out of whack to me.

Testa..... I ran your system from Mar 99 to end June 2003. It had 1795 selections for 158 winners, a 8.8% strike rate and a loss at level stkes of 15.3% using NSW TAB dividends. Probably the worse thing though was the longest losing run which was a massive 72 selections.

While the idea may have merit, I can see that some of your rules are backfitted and thats why it doesn't hold up over unseen data. For example, what's the racing logic behind excluding 3YO races?

I analysed the results by favourite rank and they all lost pretty much the same amount. Interesting enough though, favourites who matched your rules only won 25% of the time for a 23% loss on level stakes.

It may in fact be a good system for identifying false favourites (rather than winners, regardless of the price).

Hope this helps.