CanadaHaha yeah it's pretty legendary!
POG THAT'S A FIRST EVEN SOME VETERANS HAVE TROUBLE
Yeet (and AS13B for the record) used a conventional method, playing with the game slowed down (and using savestates ofc, to retry jumps until they were satisfied). There is however another method, that is making an actual robot that plays the game for you. You can tell it where to place the mouse and it will do so every single time you run the script. The former method has the advantage of being quick and easy (relatively), but the latter is much, MUCH more precise. Instead of being limited by human precision (even if the game is slowed down, yes, that is still a thing), everything is available, you can decide the precise location of the mouse on every single frame. You can also easily compare two scripts, strats or routes, because due to this game (like most) being deterministic (AKA no randomness), you can replay as many times as you like all the scripts you've made, whereas with the former method, you basically need to stop when your guts tell you it's probably optimized.
This formula being necessary is so dumb when you realize FPS is calculated using IGT * 30 / RT xDD
Yes and no, it's made to show the fastest theoretical time, so even if this right now is indeed invalid in a way, it proves that the time is theoretically possible
It's indeed likely that sub 30 GOISE is possible but I wouldn't go as far as saying it looks highly unoptimized xD I would rather say it's certainly not optimized, like almost all TASes, but it's not obvious why or how, like most good TASes.
Does anyone know how much Mountain Skip has actually been attempted when TASing? It's definitely worth a shot. Even if it looks outrageously impossible for humans, that doesn't mean it can't be done with TAS, and that being said, it's even more likely to be possible because it does not look outrageously inhuman.
The GOISE TAS does so much crazy stuff that isn't RTA viable. The GOIWBF TAS does too but honestly, not as much. I don't see a reason why since GOIWBF is more extreme than GOISE on the unforgiveness (not just in casual gameplay but also speedrunning-wise) scale. This should mean the TAS has a bigger advantage over humans, whereas in more "forgiving" games (e.g. platformers, such as SMB), where physics and controls are less complex, humans can get much closer to TASes.
Hope you had a good read I guess :P
I wonder how much GOIWBF people know about GOISE on average xD
@ThisTextWasRemovedBy absolutely but there's still a few other factors. The method for TASing GOIWBF rn is (afaik) still the equivalent to what we would call Geo's engine in GOISE. The best that was done with that is 33.3s, and it wasn't even sub 36 until the new engine came out. I don't know how comparable the old times GOISE TASers were to Yeet, but despite the lack of info we could still definitely assume that the more precise engine would likely bring more potential.
Slipped at slope climb at 4s, failed gap jump at 7s, missed yolo pull at 11s so slow tower climb at 12s
