[boole1854]

While interesting, this does not seem very convincing to me. They successfully show that Niemann was playing many games with a high percentage of "engine perfect" moves, but they do not do enough to show that this is inconsistent with what top players usually do. A whole distribution of scores is shown for Niemann but only limited summary statistics are shown for other top players. A proper comparison would involve showing the same type of data for both.

[VanillaCafe]

> They successfully show that Niemann was playing many games with a high percentage of "engine perfect" moves, but they do not do enough to show that this is inconsistent with what top players usually do.

I thought the video very much did make that case. A single known cheating game had a 98% correlation (Sebastien Feller Paris 2010), other GMs have generally at most 75% average correlation. The analysis had more than half a dozen games with Niemann at 100% correlation. If that's cherry picking, it seems like there are a lot of cherries to pick.

[roflyear]

Yah and Hikaru was able to find games of his that were 100% too, fairly quickly: https://clips.twitch.tv/FaintCuteKumquatPhilosoraptor-hDvbAj...

[usgroup]

Well she shows 6+ games where he has 100% correlation with the engine. What are the chances?

[aqme28]

> What are the chances?

We don't know! That's why this is an incomplete analysis. A comparison against other players of his caliber would answer that question.

[usgroup]

Hikaru tried to answer just that:

https://www.youtube.com/watch?v=qjtbXxA8Fcc

[boole1854]

Yes, that's the question that I wish she had tried to answer. What are the chances? Without checking for that pattern by other 2600+ GMs, we don't know the answer.

[goatlover]

Hikaru, who ranges from high 2700s to 2800s, said only one of his games is at 100%, which was his best game. Hans having multiple 100% is suspicious.

[8note]

How long was the engine run for those cases? Does the engine output change if you double or halve the depth?

100% correlation to an output that can be tuned doesn't seem that exciting

[johncessna]

And then explains the odds, to both this one and the parent's question

[__s]

https://imgur.com/a/JpJsMyI

[glumreaper]

OMG look at the Y axis scale of the graphs. Everyone else is 0-1000+ (often 0-2000), Magnus Carlsen's graph scale is 0-200.

[__s]

Yeah, because Carlsen's is only a sample of 426 games

The first 4 are the most interesting, having same sample size of 4000. But across the board players tend to have little distinction between choosing moves between 0.0 & 0.1, except one player