Statistical question based on (sample, confidence intervals, financial markets)

Hi,

I work on financial markets,

We always do optimization for our trading strategy parameters on the past(in sample data) and then if its profitable we apply it on the future (out of sample data)

How to find the ideal sample size for an unknown number of participants to have a high level of confidence in our measurement (strategy out come)?

The objective is to differentiate between a resulted measurement based on luck and the one that has a statistical significance and can be trusted.

Example:
An unknown number of participants = Number of Trades.
A measurement = Win-Ratio.
And how to know the confidence rate for our optimization test?

Thanks,