I already knew about this but I'm personally against implementing it. Having no FD, FE or FF in the first round of numbers also means no FD*5+1 in the second round and so on. I'm against touching the rng at all unless we could somehow figure out or come up with every single detail. This in my eyes would also imply that the first 10 random seeds would have to be generated following the odds in the gameboy. And also that we simulate every single random number genration exactly as it occurs in the game boy (see my post 19). It's just not worth it imo.
Anyway I couple of things.
When it comes to critical hit calculation, after the random number is picked, rlc a instruction is executed thrice. This for example turns BF into FD, or DF into FE, so the chances would be different.
For the random damage factor on the other hand, there's a rrca instruction. For example, FB would become FD (though once again, getting FF would be impossible).
Either way, as I've said, I personally wouldn't bother simulating the rng since I don't think we would be able to make it 100% accurate anytime soon. I'd rather keep the current rng we have than implement some of the consequences derived from the game boy rng just because they are the easiest to figure out. Just my opinion.
Anyway I couple of things.
When it comes to critical hit calculation, after the random number is picked, rlc a instruction is executed thrice. This for example turns BF into FD, or DF into FE, so the chances would be different.
For the random damage factor on the other hand, there's a rrca instruction. For example, FB would become FD (though once again, getting FF would be impossible).
Either way, as I've said, I personally wouldn't bother simulating the rng since I don't think we would be able to make it 100% accurate anytime soon. I'd rather keep the current rng we have than implement some of the consequences derived from the game boy rng just because they are the easiest to figure out. Just my opinion.
