Antarctica
@SupaYoshiBro @ScratchySpeedrunner There isn't another explanation. This code sets the inputs to null if the player has won and it runs every frame. This is unreconcillable. I have tried to find another explanation but the task was fruitless
@ScratchySpeedrunner The evidence for DapperScratchers TASing those other levels are the inputs after the end.
Does it though? what is the average frame rate for full game runs? I would be interested in a comparison to other runs
@SupaYoshiBro That gives way too much of an advantage. I agree with @pi314 when he says pause buffering is against the spirit of speedrunning.
wr :D
An additional thing to consider is whether or not older runs which used pause buffering should be grandfathered or not. On one hand if we grandfather them then that means that newer runs want to tie or beat these older runs will have a significantly harder time because the older runs are allowed to do something they can't. But on the other imagine working really hard on a run but then a year later, a run you sent got removed. That would seem really unfair. Enforceability is very hard though and truly impossible to detect cheating, but extremely repetitive inputs should be considered proof. We need stronger run verification due to the recent cheating scandal involved with @Sir_farcelot
the second apple you collect can be skipped by landing on the crumble block (but avoiding the apple) jumping to the right away and landing and then waiting and then jumping and landing on the crumble blcok just as its recovering before jumping and going to the other side of the crumble block and jumping to the next checkpoint
Oh so he is screwed. Even floating point can't save him.
Ok, so I previously thought that Sir Farcelot couldn't possibly be defended. But I found a minor discrepancy in programminguy's methodology. See, floating point implementations is dependent on the JS engine used. Note this is implementation we are discussing so the same mathematical computations could lead to different results in different browsers. So when Appel is moving on programminguy's machine, it might arrive at the flag earlier due to these differences. Although these differences are extremely minor, so I would suspect maybe one frame or so of difference between when Appel on Sir_Farcelot's machine touches the winning flag and when Appel on programminguy touches the winning flag.
Now I have a question for you @speeeedy At what time does Sir Farcelot's replay code let go of all the inputs. Is it above 15.2?
I use the same os and browser. Can you send the html file for analysis?
