But they were done. Human skills can do them.

Actually, Somewes is working on 8-1 FPG, he's been having some success with it.

@LeKukie He's either the 5th or the 6th person who's done it. So several people have succeeded with it.

@Mars02 I think he’s the 8th

Wes said he's the 6th. But there could've been more people that've done it that we don't know about.

My slightly updated post from 8 months ago in the "4:54?" thread is below. Does any of this need updated? Has any human done 4-2 from a save-state or in any way?

As for the thread today, is 4:54 possible?, of course....anything is possible. But without new strats/discoveries, it is EXTREMELY unlikely.

Previous post: " The "sum of bests" ever achieved by humans, from save-states or in runs or whatever, is 4:55.213.

That includes the fastest 8-4 ever....which is more than 10 frames (0.167 seconds) faster than it's been done in any recent record run.

And as mentioned above, that also includes a framerule in 8-1 that has only been done from a save-state, and only by something like [6-8] people.

So those only get you to 4:55.213. To go 4:54, you'd need a faster 4-2 to save another framerule (0.350 seconds), which hasn't ever (yet?) been done by a human.

In short, for 4:54:

• Do all the tricks already being done.
• Add 8-1 that's only been done from savestates and only by approx [6-8] people.
• Add 4-2 that's never been done by a human in any way
• Run 8-4 about 2-3 frames faster than any recent record run. "

Well technically it's not even possible multi-segment so right now, yes, it is impossible

@grodney2000 if I remember correctly, the human SoB is now 4:55.096 with a better 8-4 than the one you mentioned.

Right now we believe the human limit is 4:54.98. But if we discover future glitches we could bring the time down lower.

I think if there was a glitch like that then we'd probably find it by now. But there could be. We are one of the most optimizing communities on SRC.

Thanks @raymondamantius , yes, it looks like the 8-4 IL is (7 frames?!) faster now!

Well SMB2USA took till this year to have a game end glitch discovered... so why not SMB1 (even though this game is super simple for that to happen... but just maybe :) )

4:54 is like a 0.001% chance of happening, and yes we said the same things with older times, but that was before we literally tore the game to shreds, the only sliver of chance of this working is working on your controller inputs, and how fast you can do them, now that is way easier said then done, all of us as community should accept how small of a chance 4:54 has of happening with all facts and ways we have tore the game a part, and just because someone can DO a clip that is for 4:54 doesn't mean they will get it every try, some people do these things once and can never do them again

It seems like you have little faith of humans acheiving 4:54. There could be other glitches that we have never discovered yet.

"and just because someone can DO a clip doesn't mean they will get it every try," Humans don't HAVE to perform a trick/clip every try they attempt to perform it to get it in a run. Runs are still capable of dying to FPG, 1-2G, and BBG. But we've still gotten a dozen runs with those tricks. Humans may not have gotten the new 4-2 clip yet. For several years the standard 4-2 clip was thought impossible (several years ago), so humans never tried to do the trick. But now (for the past 2 years), humans found a way to perform the 4-2 clip. Just because someone hasn't done the 'new' 4-2 clip YET, that doesn't mean they can't do it. <- I don't think you were trying to say that it's impossible, but I said it anyway.

There are 2 quotes from @Bismuth that I follow.

1. "The limits are never where we think they are."

2. "If it is possible, it will be done."

BTW., Maru is theorizing a way to do the game end glitch on SMB1 now, and it has to do with Manipulating enemy slots in a way that allows for running the code that reloads the title screen (as he said)

I think he'll be making a TAS of it too.