[WR] Yiheng Wang 0.78 2x2 WR Average

[Filipe Teixeira]

This is proper time dilation in practice. Why didn’t I see this sooner?
Jperm forgot to mention this was another way to achieve a faster time.

T I M E M A C H I N E​

 

[qwr]

it would be disappointing if yiheng's 19 TPS was due to illegal starts. looking at the video without going frame by frame it looks fine. but if every other 2x2 top solver had to follow this rule then it would be only fair yiheng has too as well. I don't like the idea of seeing this disqualified because it looks like legitimate 2x2 solving.

 

[precompsubwr]

idea: everyone starts using sliding timer stops to protest against the wca


(jk please don't do this)

 

[qwr]

idea: everyone starts using sliding timer stops to protest against the wca


(jk please don't do this)

Even worse, everyone has to use 4-pad mode on stackmats

 

[Mr Cubism]

This interest of finding stuff for disqualifying is insane. I have an idea: from now disqualify every new WR. Why? The possibility of being A.I.-manipulated.

 

[NigelTheCuber]

people in the replies glazing yiheng

 

[kake123]

Moving forward (regardless of the outcome), every Delegate, Coach and Top Cuber around the world will have to take a free Master Class course on Timer Starts and Stops, Framecounting solves, Frame-by-Frame analysis and all possible Timer Start/Stop abuses under the sun.

Fail it and you are not allowed or at least not recommended to compete if you are a Top Cuber. If you still insist in competing despite the failure, you will be told that you risk potential disqualification. If not disqualified, you risk being regarded by notable members of the community as not legitimate.

This course should be included in all Cubing Courses/Academies around the world and translated to as many languages as practically possible.

 

[stoic]

The frame-by-frame analysis in the video posted above doesn't appear to take into account that the display may not accurately reflect the time on the timer itself at each point. It also makes an assumption about the FPS speed of the video itself - can this be taken to be accurate?

 

[Clock_Enthusiast]

The frame-by-frame analysis in the video posted above doesn't appear to take into account that the display may not accurately reflect the time on the timer itself at each point. It also makes an assumption about the FPS speed of the video itself - can this be taken to be accurate?

This is an extremely important point to consider. By WCA guidelines, the judge is even supposed to look at the timer itself when writing a result to reflect complete accuracy. The display does not reflect the exact time of the timer throughout the solve. In some very rare cases, it also sounds like the display doesn't shiw the correct time AFTER a solve either (I've never seen this, but the regulation must exist for some rare occurences.

 

[Clock_Enthusiast]

I understand that as a high-profile speedcuber, Yiheng will undergo a higher degree of critique over the verification of official solves. I don't see this level of scrutiny with other popular record holders. Yiheng reached a new level entirely for 2x2 and 3x3, and people are understandably surprised (especially given his age). I feel like many have villainized him to some extent now, which I don't believe is fair.

We need to recognize that this is an actual child who is not even ten years of age. Think to when you were this young--did you do every task perfectly? I could barely tie my shoes at that age (lol). The occasions where Yiheng has made errors may not from him attempting to sabotage results, but more than likely because he is a young child. It would make sense for him to be more comfortable with regulations given his experience and status, and any results which do not follow regulations should not count as they would(n't) for anyone else. However, the amount of effort this community puts into attempting to call him out and prove illegitimacy is alarming. There are cases where I agree that things have been questionable (including his first 2x2 0.93 average), but the second this kid does anything, everyone goes AFTER him.

 

[GloopyGum]

1. Displays don't update the time quick enough for it to be a reliable reference during real-time solves.
2. Sliding timer starts are not allowed under WCA regulations (because you start the timer with your palms on the timer, not your fingers). But in the past, the WCA has basically said that if the naked eye can't tell for sure if it was a bad timer start. This 0.53 2x2 single that the WCA allowed (with a very very bad timer start) is a good example.

It's very likely that Yiheng is doing this on purpose, and trying to see how far he can bend the regulations in his favor. In my opinion this is plain bad sportsmanship, and if we just allow these bad timer starts then the shorter events would just be about who can get away with the worst timer starts. The integrity of the WCA would be greatly damaged.

 

[OtaMota]

Displays don't update the time quick enough for it to be a reliable reference during real-time solves.

Instead, we can use this method that Ibrahim Khanani uses instead https://docs.google.com/document/d/1syPge3vcj1oMw7l6qaRPgFdXmUZnkf3U0jNWoBAPkVk/edit

Here, he measures the pure exec time and compares it to the recorded time. In the first solve of the 0.78, the pure exec time (0.76) was apparently longer than the recorded time (0.74) he got. This could be proof that Yiheng did a turn before/after the timer started/stopped.

 

[2x2MasterYT]

You guys are REALLY trying to find a reason why Yiheng can have another record canceled. He's 10, is he really trying to find a way to bend the regs to his favor? I think not. Timer starts! Come on

 

[OtaMota]

He's 10, is he really trying to find a way to bend the regs to his favor?

Intentional or not, disregard for the regs is not good.

 

[DuckubingCuber347]

You guys are REALLY trying to find a reason why Yiheng can have another record canceled. He's 10, is he really trying to find a way to bend the regs to his favor? I think not. Timer starts! Come on

If the excuse is that 10 year old children cannot be trusted to follow simple rules perhaps the following logic declares an age requirement be set in place to be able compete...
All competitors should be entirely expected to follow the rules carefully and thoughtfully set in place. Unless an exception is specifically made in those same set of rules Yiheng should *absolutely* not be exempted from scrutiny.

 

[stoic]

Instead, we can use this method that Ibrahim Khanani uses instead https://docs.google.com/document/d/1syPge3vcj1oMw7l6qaRPgFdXmUZnkf3U0jNWoBAPkVk/edit

Here, he measures the pure exec time and compares it to the recorded time. In the first solve of the 0.78, the pure exec time (0.76) was apparently longer than the recorded time (0.74) he got. This could be proof that Yiheng did a turn before/after the timer started/stopped.

I'm being pedantic, but it relates to my earlier point:

assumption about the FPS speed of the video itself - can this be taken to be accurate?

Unless I've misunderstood what Ibrahim has done here, he hasn't actually measured the pure exec time - he appears to have used a frame count as a proxy? And that relies on an assumption of various different cameras (or, more likely, phones) running at a constant and predictable frame rate which never varies, either before or after compression and upload to YouTube. In other words, he's done an experiment missing an important element of control.

If we assume an intended frame rate of 60FPS recording (and I'm not sure exactly if we should), what's the likelihood that someone with a brand-new iPhone is running slightly ahead at say 61 FPS, vs someone else's old Nokia which is nearly out of battery running at say 59 FPS? That would be enough to turn a 0.74 into a 0.76 based on frame counting alone due to hardware differences. I'm not a hardware engineer, so what am I missing?!

 

[baseballjello67]

I’ve heard this argument plenty of times - “You just don’t want Yiheng to have his record, he’s 10” or something along the lines of that. But, (at least from my own perspective) I’m incredibly happy when Yiheng gets a world record. I’m incredibly happy when anybody gets a WR, in any event! But the one thing that makes me incredibly mad is that argument. If Yiheng is willing to go to competitions and obtain these amazing results, he has to do it legitimately. That goes for any record for any person. Being ten does not have any difference on whether or not he should follow the rules.

 

[OtaMota]

I'm being pedantic, but it relates to my earlier point:


Unless I've misunderstood what Ibrahim has done here, he hasn't actually measured the pure exec time - he appears to have used a frame count as a proxy? And that relies on an assumption of various different cameras (or, more likely, phones) running at a constant and predictable frame rate which never varies, either before or after compression and upload to YouTube. In other words, he's done an experiment missing an important element of control.

If we assume an intended frame rate of 60FPS recording (and I'm not sure exactly if we should), what's the likelihood that someone with a brand-new iPhone is running slightly ahead at say 61 FPS, vs someone else's old Nokia which is nearly out of battery running at say 59 FPS? That would be enough to turn a 0.74 into a 0.76 based on frame counting alone due to hardware differences. I'm not a hardware engineer, so what am I missing?!

At least in Yiheng 0.78s case, Dylan Baumbach had both 30 and 60 fps footage, and found that the pure exec time was slower than recorded result.

 

[stoic]

That's interesting - thank you - but it hasn't actually fully answered my point. If the equipment used to measure the count of the frames hasn't been independently tested or verified, all it potentially shows is that something labelled as "30 FPS" is running half the number of frames as something labelled "60 FPS", not the ACTUAL verified frame speed.

 

[qwr]

That's interesting - thank you - but it hasn't actually fully answered my point. If the equipment used to measure the count of the frames hasn't been independently tested or verified, all it potentially shows is that something labelled as "30 FPS" is running half the number of frames as something labelled "60 FPS", not the ACTUAL verified frame speed.

Mobile phones almost always record at slightly variable framerate. However from my video editing experience this is very negligible. At least not noticeable enough to cause audio desync which would happen if it were I estimate 0.2 s off or more. You can go through the video with ffprobe to see the actual millisecond values for each frame.