I know but if someone is going to be trying to brute force bitcoin private keys using whatever method, it's like you said, they'll want to be focusing on address types that are most in use obviously to "increase their chances". Thus not p2sh address types. Thus legacy addresses.
Actually, legacy addresses are the least "in use" at the moment, with around 24% (and declining) of outputs being legacy outputs. P2SH accounts for around 39% of outputs, and P2WPKH/P2WSH accounts for 37%. You can see this here:
https://transactionfee.info/charts/output-type-distribution-count/. P2SH has remained fairly static around 40% for around 2 years, but with segwit taking an ever larger share away from legacy addresses.
However, as I said above, I suspect most people trying to brute force random keys are focusing on legacy addresses, since legacy addresses still hold the most value once you account for the several million early mined and unmoved coins.