I'm trying to represent how the top runs are being accomplished, distilling it down to a graphic.
<edit: image removed because google changed their url scheme>
(I sure hope that link works.)
I know there are some TAS things not shown, and some backward fast accel, etc. But I'm really just trying to show the human-doable things that have been incorporated...beyond the 10ish-years-ago classic "glitchless"...in order to save frame-rules.
Thoughts? Or things I'm missing? I'd love to add a "number of frames" for 8-4, kind of to make the story complete, but I don't have that info.
p.s. I'm new here, and don't know that this is thread-worthy, so please lock/delete/ban as appropriate, with my apologies in advance.
I should call the 1-2 wall clip, the 1-2 pipe clip, or 1-2g frame rule.
Wall clip is a fine name, that's fine I would include 8-1 Fast accel + Mavpole, because someday someone will do it. No one has done it yet ever though. Only human theory
@Lul_ecks_dee I don't think that it will ever be done in a run, because not only is it a mavpole, but it is the fastest possible mavpole possible, and the fastest fast accel possible. It MIGHT be done in PRACTICE, but never in a run.
it might be in 10,000 years when humans are 100% perfect and society has absolutely no flaws.
You never know, maybe a setup for mavpole will be found, maybe someone will be crazy enough to do it,
The image in my original post has disappeared, so I hope this one works. I've updated the chart to contain details from Bismuth's awesome new explainer video.
Questions remaining: What was the 8-4 frame count on the glitchless record? What is the official time if 8-1 FR is saved....and I might as well include it in Human Theory now that it's been done from a save-state, right? Does lag frame only happen in 8-2, or can it happen anywhere?
Hi. I have somewes's new record being 13 frames faster than Kosmic. So that's one framerule (21 frames) faster, but then 8 frames slower.
If there was a lag frame (as there was for Kosmic), his 8-4 was 8 frames slower than Kosmic. If there was NOT a lag frame, his 8-4 was 9 frames slower than Kosmic.
Does anybody know if somewes had the lag frame? Or did he avoid it?
Thanks!
Okay, somewes says he didn't get the lag frame, so that makes his 8-4 9 frames slower than Kosmic....4 frames slower than darbian.
With the .245 now within a framerule, adding the 1-2 wall clip would be 4:55.xxx with a reasonably slow 8-4.
But also note somewes was 19 frames slower on 8-4 than the human best. He needs 15 frames to get to 4:55.xxx, so he (or anyone) could run the same framerule but do 8-4 at 15 frames faster ....i.e. 4 frames slower than the fastest a human has done it.
should human theory 8-1 have a coin in it? stuck in a plate has a video doing it on youtube.
Oh, huh! If it's been done, then I'd say both Human Theory and Human Best!
Original, Human Theory just meant what could be done without machine cheats like LR at the same time.
And Human Bests meant whatever had actually been done under any conditions.
I don't track it closely at all, so I'm happy to make any recommended updates!
I think it's only been done in practice but it is humanly possible.
Updated Human Theory and Human Bests to include 8-1 FA+FPG to save a framerule.....adjusted times accordingly.
- Human Theory is anything that's theortically do-able by a human.
- Human Bests is is the best that has been accomplished by a human under any conditions.
With kosmic's new record last night, we are now 1 framerule (in 8-1) and 20 frames (in 8-4) from the sum of best humans ever.
Oh geez, I didn't even see that kosmic tied somewes earlier in the night. I have added the tied run (and dates) for completeness.
Yeah. I used to have a column for human theory, as seen in the previous posts, but I couldn't find strong consensus on true human possible vs not.
I only see a giant minus where the picture should be Thanks for taking the time to update this as the WR progresses my lit b
Any% (NTSC) runs below 4:57.000 must now fulfill additional requirements in order to be verified.
- The run's full session must be included in the submission description.
- For emulator runs below 4:57.000, some form of input display must be visible for the duration of the run. A hand-cam or input
