Do you think fewer trades can make performance easier to evaluate?

A selective approach may reduce unnecessary exposure, but too few trades can also make performance data harder to interpret.

How do you find the right balance between selectivity and a meaningful sample size?