From the footage I have seen and placing it into excel for the fastest time possible in Mario Odyssey, It looks like the perfect fastest run possible would be 1hour 12mins and 36secs.Anyone else been working on this? / any views on the time?
I firmly believe that every world record can be broken with the exception of maybe a couple of misc. games.
The WR to SMB was beaten 13 days ago. SMB came out 32 years ago.
That's my answer.
Mario Odyssey is such a huge game that a “perfect run” can’t really exist for it. There’s always going to be new glitch discoveries, routing, optimisations and so on.
For an idea of how inconceivable an optimised run for Mario Odyssey is at the moment, take a game running at 60 FPS, which when run perfectly, takes exactly one minute to complete. That’s 3600 frames worth of perfect inputs. For an actual game of that length, it might take weeks or months to find the perfect run, depending on the complexity of the controls. Or just 60 seconds if the entire goal of the game is to hold right for a minute.
It might not look like it at the surface, but Mario Odyssey is an EXTREMELY complex game thanks partially to the vast choice of moves, partially to the 3 dimensions of the game, and partially to the fact that Mario Odyssey is incredibly non-linear. Perfecting an entire hour (although I guarantee that a ‘perfect’ run would last under 45 minutes) of Odyssey would take 3600 seconds of perfect inputs (or 216000 frames), with hundreds of input combinations available on each frame, and each input changing the availability of future inputs.
Even ignoring all the glitches that haven’t been discovered yet, finding the perfect hour-long string of inputs from that mess will take more than a decade, probably more than four or five, and that’s with a robot performing the inputs. Humans can’t even complete a game as simple as Super Mario Bros with the same speed as a computer, so the possibility of reaching machine efficiency with Odyssey is so close to 0 that it isn’t worth trying to calculate.
And again, that’s ignoring glitches, which could change the routing in who knows what ways, and are bound to be discovered routinely for the next decade at least.
TL;DR, Odyssey’s speedruns are as close to perfection as the Earth’s core is to receiving a breath of fresh air. Which is to say, not.
EDIT: And for an actual example: Did you know that spinning can be buffered from any animation? This could make it possible to quickly change directions while rolling, allowing it to be used in many more situations; and offer a way to do spin-throws frame-perfectly, and without risking the motion controls not working properly. That simple tech alone could possibly save minutes from a run, and it’s just a basic part of the control scheme.
Now imagine if a glitch was found that let you bounce off Cappy twice in one jump. Congrats, hundreds of new skips just became available, knocking the timer down a full 20 minutes. Cascade kingdom can now be done without even capturing an enemy, since you can just jump over the river and scale the cliff instead of capturing Chain Chomps. Et cetera.
Yeah, the original statement of 1:12:36 being the fastest is now confirmed bogus as Vallu just got a 1:11:48. There's so much potential to this game. The best possible time will not be closed in on soon.
I wonder if we could make some algorithm that takes all WR times on this site and can predict a limit for each game. Probably not too difficult to do using the API and some calculus.
You can’t accurately predict a limit because you never will know what will be discovered, that was the crux of the point made in this thread.
Even predicting the limit of a game with current known strats is impossible, that’s why people dislike sum of best as a comparison for doing runs - you can never play perfectly (unless the game is maybe minutes long) because things like stress, pressure, etc all impact the ability to perform.
Measuring the limit or perfect time for any game, any category is just a silly exercise. It always will be.
objectively wrong, if you can't accurately measure the best possible time for ur speedgame you're literally irrelevant in said game or you don't care enough bout ur speedegame...
To accurately measure the best possible time for a game, you'd only not only have to find literally the perfect permutation of some few hundred thousand frames (creating trillions of permutations that would take our best computers decades to simulate each possibility), but also somehow know in the future that no additional glitch, route change, skip, etc. will ever be found by anyone at any future time in all of human history...
tfw u can predict possible glitch discovery in the future n kno what a perfect time is before it's even close to possible, lern bout ur speedegame...
I've never been one to have an apoplectic fit because someone replies to a topic after some arbitrary phoney-baloney statute of limitations has passed (I've seen people have conniptions over "bumps" from as little as two months prior). If you can contribute to the topic, contribute. Not complicated?
But you gotta contribute ;)
