Wisdom suggests that you need many selections to prove a system, and I believe this theory has been generally accepted.

IF your base filter, eg; A Win S/R%, attracts a large amount of selections, then I believe this is good enough. How can one get 1,000 selections for a decent system, you'd have to go back 20 or 30 years in some cases, and racing will have changed so much that the systems is probs useless today anyway.

I disagree with those who contend that you need a large number of selections for your system to "prove it".

I believe you do need a large number of selections for your BASE filter only.

There needs to be a strong NEXUS between your BASE filter and all other filters, and the stronger the better. And one could then deduce that because all filters have a strong NEXUS to the base filter, and a strong NEXUS between all other filters, then the huge number of selections for the BASE filter could be "claimed" by the other filters.