Hacker News Reader

Comment by 2bitencryption | Hacker News Reader

Can someone explain something -

Assume for a moment Hans did not cheat (aside from the times he admitted to cheating).

In the scenario where he did not cheat - how amazing is his performance? I.e. beating Magnus, Elo rating, etc. I know chess.com did an analysis, which I've read, but I'm hoping someone more impartial could explain it.

In the scenario where he did not cheat, is he really on a Bobby Fischer-like ascendance?

This is what I'm trying to understand - both scenarios seem so wild to me. Either (a) this guy is cheating a pro-level chess (totally wild) OR (b) he's on a crazy ramp-up of skill (totally wild).

(a) is pretty self-explanatory.

(b) is what I want someone to explain - if this is the scenario, just how crazy is the rate of his skill increase?

View on HN

He has had an impressive but not unprecedented ramp up, from 2400(ish) to 2700 in around 3 years. There are actually other young stars, in particular a trio of Indian prodigies who've had a similar surge at the same time. I recommend the work of Nate Solon, a chess master and data scientist, easy to find on Twitter, he's published some interesting analysis. It's surprising for Magnus to lose a White game at all, especially to someone not in the top ten, but really no more than surprising - these sort of things happen. Can Djokovic lose to the world #100? Absolutely, he can have a bad day. The world #100 (in tennis or chess) is 98% (ish) as good as the world #1. Did Niemann cheat against Magnus? Obviously I don't know but I do think extraordinary claims should require extraordinary evidence and I don't see it in this case.

Carlsen was also immediately worse out of the opening and the game went directly from the opening to an endgame where Carlsen(white pieces) was slightly worse.

His win over Carlsen wasn't considered to be superhuman. Neither player played perfectly. In his podcast, Fabiano Caruana (someone capable of judging this) said he felt very strongly that Niemann hadn't cheated in the tournament.

His ELO gains are impressive, but not unique. [1]

If he is cheating, no sensible hypothesis of how he's doing it has yet been offered. So it seems more probably he's a legitimately talented player whose merit's obscured by his past.

[1] https://twitter.com/ChessNumbers/status/1568340548490252288

What, the anal beads is not a sensible theory? /s

There was nothing particularly amazing. His rise is not extraordinary. The only part that is slightly unusual is that it came towards his mid to late teens, whereas such a rise is commonly seen in the early teens. However, even that can very easily be explained by the pandemic, where he used that time to dedicate himself to chess (which is something no one is questioning).

Further, in the particular game he beat Magnus in, he did not play an extraordinary game or anything. He simply did not make any major mistakes. Whereas Magnus did. If imagine the majority of GMs playing in Hans’s position would have won that game.

I‘ve been trying to draw parallels to other sports in this regard. Nobody accused Tiger Woods of cheating (at golf) when he redefined the sport. People DID accuse Lance Armstrong, and they were right (although in that case the whole field was cheating). New England Patriots? Kinda cheated. Michael Phelps has never been accused AFAIK.

As a non-chess player, I can’t help but wonder if this is just an entire sport refusing to accept that someone could come along and achieve this, or if it’s really just too good to be true.

Plus Hans did admit to at least some amount of cheating.

The comparisons are a bit extreme as Hans hasn’t become the world’s best chess player; not even close. He just beat Magnus once…

To be clear online chess legitimately has a cheating problem. I think it has gotten better but according to chess.com there have been a number of Grandmasters including some of the top 100 players in the world who have cheated on the site and admitted it. Cheating in over the board chess seems to be much less common but everyone realizes it is possible and if done well would be challenging to catch. I think this has primed the field for people to be suspicious.

I’m a bit skeptical of this figure that chess grandmasters have admitted it. From what I understand, if chess.com accused you of cheating, they tell you “you have 3 days to admit it or be banned”. So in the face of that, of course they’re going to admit to cheating (since admitting it leads to no other repercussions). To me it seems like a way for the anticheat team to farm credibility. This was my impression based on Maxim Dlugy’s statement

https://sites.google.com/view/gmdlugystatement/home

> To be clear online chess legitimately has a cheating problem.

Depending on how strict we want to be, Magnus himself has been caught "cheating" in online chess games before too:

https://www.reddit.com/r/chess/comments/xkn7ln/why_does_nobo...

Golf's a game of precision motor control, while insanely long drives might be evidence of cheating, it's hard to imagine a way to make yourself more accurate besides skill, at least without it being obvious on live television.

Biking obviously involves a lot of different things but it's much easier to see how doping is going to help.

Football has so many facets that it's kind of surprising there aren't more cheating scandals, but at the same time it's harder to isolate any one thing and say that's what's giving a team an edge.

Michael Phelps rose to prominence after most of the big doping scandals by which point anti-doping measures were well established. Besides doping, it's tough to cheat at swimming.

With chess, it's trivial to have a computer on your person capable of beating any human alive, it's just communicating with said computer that is a challenge. Given that chess masters are supposed to be extremely good at strategy and making clever moves people won't see coming, I'm sure there have been some ridiculous cheating attempts.

Legitimate curiosity from a non-golf person: how would one cheat at golf?

Mostly things like illegal equipment (besides blatantly breaking rules).

https://golfguidebook.com/best-illegal-golf-clubs/

“Improving your lie” is a common way, meaning altering the environment around the player, swing path, ball, or shot path in a way that yields some advantage.

This would include things like breaking off a twig in your way, removing leaves or grass around the ball, etc.

Steroids for longer drives.

And faster recovery allowing more training volume.

Cheating at golf is very common among recreational players. A common trick is to move your ball onto the edge of the fairway when nobody is looking.

> I can’t help but wonder if this is just an entire sport refusing to accept that someone could come along and achieve this

That's what Niemann wants you to believe, but it's not right. Carlsen is rated about 200 Elo points above Niemann. That makes a win in classical an upset, sure, but far from unprecented, and far from proof that he's a threat to Magnus' dominating position in the chess world the last decade. Weaker players have beaten Magnus, even with black.

Hans is currently rated around 2700 (at least his live rating after his performance at the US Chess Championship). A few other GMs around that rating have beaten Magnus in just the last week or two (Arjun Erigaisi and Gukesh to name a couple), so it's not out of this world. It is still certainly an upset, but not impossible.

Arjun and Gukesh haven't won classical games against him. Losses are normal in rapid, even for Magnus.

> In the scenario where he did not cheat - how amazing is his performance?

Not amazing at all. It’s a fairly standard performance which is why it’s very dubious he cheated OTB. The chess.com report has to stretch the truth considerably to try to imply there is something fishy OTB. Magnus was just playing aggressively to score against a weaker opponent and had a bad day.

His performance ramp up is not particularly impressive. He played less tournaments that people his age four years ago and his rating went up towards competitive GM level when he started seriously playing. That’s a typical trajectory for a young pro-player. To be clear, we are not talking about a Fischer-like, world champion contender path here.

His rate of improvement, assuming no cheating since his ban in 2020, is a faster ascendance than any other player in history, including Bobby Fischer. So yes, assuming he suddenly quit cheating cold turkey, his performance would indicate he is a once in a generation talent and could easily become the best in the world, according to the Chess.com report.

I'm a huge Hans fan. (F him for cheating, but setting that aside.) His play is objectively insane. Either he's somehow been cheating this whole time via shoefish, or he's the real deal.

GothamChess has had excellent coverage of his games.

Like seriously, look at this one: https://youtu.be/ainUWFboMLs?t=1490

Hans threw a masterful knockout punch at the end, one which his opponent completely overlooked, and was solely a single move. There wasn't cheating-like high level strategy involved -- it was a one-move fatality.

Whatever else the kid has done, he is one hell of a chess player. I find it hard not to root for him, even knowing his history.

I love that this video has over 300k views.

GothamChess is amazing in general. He’ll be casting an upcoming chess tournament (kind of like football commentators, but chess). So hyped for it. If you haven’t seen him before, check out some of his other vids too — he does a variety of styles, like “Guess the ELO” and 1:1 chess tutorials with other popular non-chess streamers.

I think the most impressive thing is that he’s posted a video every single day for… I don’t know how long. He makes it a point of pride not to miss one.

At some point I would expect the younger players to start doing this more often. After all, they have the ability to grow up learning to play from/like an engine. It just seems obvious that their play should diverge from what was traditionally canon.

> His play is objectively insane

But he doesn't do himself any favors with comments like "the chess speaks for itself", or offering what according to other GMs is a lackluster analysis of those astounding moves.

Whether it's immaturity or cheating remains to be seen.

Definitely. Though I wouldn’t have done any better at 19. Teens aren’t known for eloquence or having the proper amount of confidence.

Please listen to Duda,Keymer, Ali reza do post game analysis. They can speak very well about their ideas. It is not normal to be a 2700 and be unable to give your thoughts on the game you just played.

Yeah, because being a GM is a completely normal thing in the first place.

Bobby Fischer was one of the top chess players of all time and was a certifiably insane nut job. There’s nothing normal about being a GM in the first place.

Niemann literally just won the biggest game of his life. He did something none of Duda, Alireza, etc have achieved. His situation was not normal. Trying to judge his actions based on whether it’s what Duda, etc would have done themselves in a situation they’ve never found themselves despite having had many opportunities to do so is silly.

This whole idea of le crazy chess grandmaster is so silly. Fischer, even as he was declining precipitously, was producing cogent analysis and annotations on grandmaster level games.

>He did something none of Duda, Alireza, etc have achieved.

Only if you mean winning in Classical with Black.

Duda beating Carlsen in Classical

https://www.chessgames.com/perl/chessgame?gid=2010697

His interview. The quality of the broadcast is a bit bad. https://www.youtube.com/watch?v=fchM2Dcr5mU

-----------------------------------------------------------

Ali Reza beating Carlsen in Rapid with a beautiful checkmate https://www.chessgames.com/perl/chessgame?gid=2072900

-----------------------------------------------------------

Pragg beating Carlsen in Rapid at 16 https://www.chessgames.com/perl/chessgame?gid=2252714 Giving an interview at age 16 at 3 am after beating Carlsen https://www.chessbase.in/news/Praggnanandhaa-beats-Magnus-Ca...

His performance at the US chess championship with strict anti-cheating measures in place proves beyond a reasonable doubt that he's as good as his rating. This whole "improved too fast" as proof of cheating is nonsense.

He might have still cheated to beat Magnus but there's no proof.

As ridiculous as it sounds, did the US chess championship explicitly check to prevent the vibrating buttplug that has been jokingly thrown around as a theory?

It's an honest question. I didn't follow this closely and don't know whether a metal detector could find metal inside a human body.

Games were on a 30 minute broadcast delay with no in person audience. Vibrating butt plug works for receive but can't think of a plausible way for Hans to send the moves out.

> can't think of a plausible way for Hans to send the moves out.

He doesn't have to send the moves out, he just has to send them in.

How is his co-conspirator going to feed him moves without knowing the board state

Who said there has to be a co-conspirator? He could have a small computer inside him, if you really want to believe he's cheating.

To be clear, I'm not saying I believe this.

Still has to feed the computer moves

There are plenty of muscles you can use to give input.

So you'd rather believe he is a genius bio hci engineer? Which is more likely?

You don't need to be a genius engineer, you need to be a competent stage magician. And competent stage magicians are more common than 2700 Elo chess players, so it's not by itself unreasonable to suspect.

Clench in morse code?

I'll leave it as an exercise for the reader if morse code via butt clenching counts as "plausible"

I'll leave it as an exercise for the reader if morse code via butt clenching counts as "pleasurable"

If it can cure depression, I don't see why it can't also improve your chess game.[0]

[0]https://www.goodreads.com/en/book/show/331807.How_to_Good_By...

He writes with such conviction it's difficult to not believe him. Going to give this a go, wish me luck.

I knew what book that was going to be :D

No, not in morse code. Instead, the computer will (through vibration) cycle through a list of likely last moves, and you confirm the opponent's last move by clenching right after it comes up. This minimizes the considerable risk of incorrect move entry.

The list will also include a special entry indicating "none of the above", which will make the computer switch to the next set of less likely moves.

This scheme could be refined be having some move partitioning on top. For example, an initial list that has only ["pawn move", "light piece move", "heavy piece move"].

Yes, you can see it in the body scan video.

This isn't really true. He started later then most other high level players, and then the pandemic meant he was studying and playing online a lot without raising his OTB rating. He's a good player but I don't think there is any real argument that he might ever be world champion.

This is just not true. Check out the analysis here [0], for example. His rating improvement has been steep but he's also played a very large number of games, so his rating gain per game as opposed to per unit of time is not so crazy.

Also, no one seriously expects Hans to be a world championship contender.

[0] https://www.reddit.com/r/chess/comments/x9bgtx/how_quickly_d...

Every chess player plays a lot of games, Hans' have just been more public/documented ones. You really don't think that every 2400+ player is treating chess like a full time job, and Hans just simply is just outworking them?

Based on his performance at the US Chess Championship that just concluded, I would say either that, talent, or both.

>Assume for a moment Hans did not cheat (aside from the times he admitted to cheating).

as far as i can tell he's hasn't cheated since 2020. Ive seen several of his games though where he certainly has cheated.

>In the scenario where he did not cheat - how amazing is his performance? I.e. beating Magnus, Elo rating, etc. I know chess.com did an analysis, which I've read, but I'm hoping someone more impartial could explain it.

Magnus ultimately drew the match; the one he lost he resigned on move 2. So no remarkable performance.

>In the scenario where he did not cheat, is he really on a Bobby Fischer-like ascendance?

The chess.com allegations has him better than bobby fischer in terms of rise. Though I would say that's not accounting for elo inflation.

>This is what I'm trying to understand - both scenarios seem so wild to me. Either (a) this guy is cheating a pro-level chess (totally wild) OR (b) he's on a crazy ramp-up of skill (totally wild).

He's an admitted cheater; but what seems to have gone unsaid is that perhaps his cheating isn't actual cheating. Rather he was reading his stream comments and viewers were cheating for him. It's not quite the same.

Flipside, streaming is always going to take some of your focus and attention. Streamers are by default not going to be able to compete at the top most level. So the probability he's at a supergm level is basically impossible.

> He's an admitted cheater; but what seems to have gone unsaid is that perhaps his cheating isn't actual cheating. Rather he was reading his stream comments and viewers were cheating for him. It's not quite the same.

..what?! Your definition of cheating appears to differ from the accepted meaning of the word in quite a few ways

From everything said about Hans the thing that sticks out to me the most is his analysed centipawn loss over all his games. Typically, as someone improves, both their loss per move and loss-variance falls. For hans, there is limited improvement over time and a high variance in the quality of his moves. He effectively has played at a 2700 level for a couple of years... This seems to be unusal compared to compariable pros.

I have listened to some his interviews recently and he talks extensively about engine analysed lines. He talks about avoiding lines he hasn't analysed or is uncertain on. I wouldn't be surprised if he has optimised his development on engine analysis. It would explain a high number of highly accurute games. It would also explain the variance in his play where he forced into unknown lines.

Hans went from 2400 to 2700, so he obviously improved. Centipawn analysis is just youtuber snakeoil. Everyone at the top level has trained and analyzed lines with engines for the past 25 years.

> Assume for a moment Hans did not cheat (aside from the times he admitted to cheating). In the scenario where he did not cheat - how amazing is his performance? ...In the scenario where he did not cheat, is he really on a Bobby Fischer-like ascendance?

I think so, yes, and even considerably moreso.

According to Yosha Iglesias[1] and her anonymous mathematician friend, for 5 tournaments in a row starting at the Philadelphia Open in July 2021 Hans Niemann played at an average engine correlation of 73%, beating Bobby Fischer's average engine correlation of 72% during his 20 game winning streak between 2 Dec. 1970 and 30 Sept. 1971. Fischer's peak rating was 2895 in October 1971 at 27yo. Niemann's peak rating was 2699 this month, October 2022, at 19yo.

Also according to Yosha, in three years Niemann has played 10 games with a perfect 100% engine correlation and 23 games with engine correlations above 90%.[2]

My own chess layman analysis of the experts' analyses is that the best chess players in the world right now generally play games with engine correlations ranging game to game from the mid-60%s to the high-80%s, with much rarer games in the 90%s, and games 100% engine correlated being a once or twice in a lifetime event. Magnus Carlson at his best has an engine correlation of 70%.[1]

Unless he cheated, it is quite clear to me, at least, that Hans Niemann is the singular greatest chess player of all time.

[1] https://www.youtube.com/watch?v=jfPzUgzrOcQ&t=15m11s

[2] https://twitter.com/iglesiasyosha/status/1574374153997647874

Yosha has admitted that their analysis was flawed… https://twitter.com/IglesiasYosha/status/1574308784566067201...

Only the ROI of the probability of 6 consecutive tournaments was miscalculated, not the engine correlation percentages.

Fischer at his best during 20 consecutive winning games at his peak performance had an average engine correlation of 72%, while Niemann's average was 73% across 5 tournaments. I just saw this: Magnus Carlson's average engine correlation across 5 tournaments in 2021 (roughly the same period as Niemann's 5) was 66.48%[1].

Niemann conspicuously appears to be a superior chess player at 19yo than both Bobby Fischer at 27yo during the peak performance of his career and a much better chess player than Magnus Carson at 30yo while he was and is reigning World Champion.

[1] https://twitter.com/ty_johannes/status/1574780445744668673?t...

I thought the consensus was that Magnus played badly and Hans played adequately but also made clear mistakes himself.

It's a very impressive rise, but in part because it started later in life than other young GMs. In absolute terms, he's still on par with about a dozen young 2700s who are about the same age. It's just that he caught up really fast.

The velocity is the issue not the skill level.

If I never bowled in my life, started, and two weeks later claimed I was getting 270 or higher on a consistent basis you wouldn't say "well there are pro bowelers his age that get similar scores "

The problem is Hans is still playing OTB at a 2700ish level. If Hans is managing to cheat in OTB chess with so many eyes on him, with all the security measures in place, and doing it so he's able to fool all of the anti-cheating statistical analyses, then honestly he should just be robbing banks or something because he could make way more money with those skills.

With regard to having many eyes on him, that would make it easier for a confederate to be inconspicuous.

Robbing a bank is high risk and illegal. Developing a cheating method for some competitive game sounds clever

What I don't think most people are appreciating in this thread is that at this level of play the cheating can be extremely subtle and yet extremely effective. None of the players in this dramas are talking about a cheater having a hidden mind link to a chess engine. These games typically hinge on one or two crucial moments. GMs need only one bit of information at the right moment to play several hundred points above their normal rating. A cough at a point where there's a win opportunity or a potential trap is all that's needed.

I think most people realize this. But, how would this even work with the time delay and a closed environment?

Last time Magnus lost a classical game with white was in 2018

You can prove statistically that he used Stockfish in some critical games; 100% same moves. In other games he was at Magnus' level, around 75%, and in some on his normal level.

Yes, Stockfish is really good. But not crazy. Just call him Hans Shoefish.

> In the scenario where he did not cheat, is he really on a Bobby Fischer-like ascendance?

From what I understand, if he's innocent then he's beyond Bobby Fischer.

Reply on news.ycombinator.com