Hi Duritz,

You say it's not backfitted but:
Was there a backfitting element involved when you developed the handicapping system? Like did you try 1 length = 0.5Kg, no that doesn't work how about 1 length = 0.4Kg, etc until it showed good results? Or have you proved the handicapping part works on fresh data not used for development?
Assuming your ratings are backfitting free did you take the ratings over the two year period and then try different filters over the top of them until you came up with a nice profitable outcome. This of course would be backfitting too.

I ask these questions not to try to shoot you down or anything of the sort and I hope your answers to both of them are in the negative. Just injecting an air of caution into the ebullient atmosphere your results have engendered lest you be carried away on the tide of expectation into perdition, depression and ultimate financial ruin.

Cheerfully yours,
KV