brucehoult parent
That's 1 byte smaller than `LDA #0`, but not faster. And you don't have enough registers to waste them -- being able to do `STZ` and the `(zp)` addressing mode without having to keep 0 in Z or Y were small but soooo convenient things in the 65C02.
You might like the PC Engine, a game console based on the 65C02*.
*Actually a custom chip also containing some peripherals.