Training mode simulation on GGPO?


From what I’ve gathered, there are external tools that enable a training mode -esque environment in GGPO or other emulators. Can someone point me in the correct direction for getting those / how to use them properly?


Just download MAME and use cheats, problem solved.


To get almost the same functionality as OE training mode, you need FBA-RR, MacroLua, and cheats for infinite health/meter. You can use macrolua to save a series of inputs and replay them on a loop, including the save and load state buttons. So with that you can practice counterpoking any move, parrying or red parrying any move or blockstring, and defending against specific setups.

The only thing you don’t really get is a great random block for hit confirm practice, but that’s not really doable on any PC emulator. You could sit in crouching and randomly fluctuate between down and down-back I guess, but you still get the extra hitstun of crouching so it’s not as useful as random block you can find on PS2 or OE.

You could use MAME too I guess, but every version of MAME that isn’t a laggy junkpile doesn’t support MacroLUA AFAIK.


Komatik check out this thread MAME 3S Training Mode Guide


Few years ago, i used an AutoIt script to simulate random block, worked pretty great.
I’m pretty sure it would be quite easy to make a quick script to do the same in LUA on Fba-rr.


Ah I see. That seems worth investigating! I’ll ask @isotopez if he could do something similar.

I’ve had an idea in my head the last few months for hit confirm practice that I think would work on emulators that allow you to change the game speed. Basically the same thing they teach you when you learn technique on a musical instrument - slow the game down to 50%, low forward, confirm on hit only. If you mess up start over. Get 10 in a row correct, then bump yourself up to 51%. Etc until you finally make your way up to 110% game speed. Then you dial it back down to 100 and everything should suddenly feel easy to confirm.

Combine that with input display which allows you to make sure your inputs are clean as well.

If we can combine that with an effective random block simulation I think it would be pretty helpful for a lot of people.

Would be pretty useful for whiff punish practice too. Whiff punish chun low forward at 50-110% game speed the same way. Basically anything that requires reactions or for us to process information and output the appropriate response in a short amount of time would be helped by this.