Kramnik has a point

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Rebel, chrisw

Peter Berger
Posts: 664
Joined: Thu Mar 09, 2006 2:56 pm

Re: Kramnik has a point

Post by Peter Berger »

Alexander Schmidt wrote: Fri May 03, 2024 1:11 pm This thread is about Hikaru, and the only evidence of cheating is his strength, and that's just nonsense. Just watch his videos where he explains all of his moves. How the hell can someone seriously claim he is cheating?
Basically, because his streaks are a little too good to be true mathematically.
Of course, we are talking about one of the very best chess players on the planet here, and no one ( also not Kramnik) is claiming otherwise, so it comes as no surprise that Hikaru can do most amazing things other people can't.
How could such a scenario look like theoretically? (And of course, I don't claim anything like this at all, just to give a general idea on what +could+ happen in a parallel world). Let's imagine Hikaru sees the eval bar while playing ..
User avatar
towforce
Posts: 11751
Joined: Thu Mar 09, 2006 12:57 am
Location: Birmingham UK

Re: Kramnik has a point

Post by towforce »

Peter Berger wrote: Fri May 03, 2024 1:35 pmBasically, because his streaks are a little too good to be true mathematically.

As I said before, the maths is flawed:

1. The data is cherry picked

2. At heart, the evidence relies on statistical distributions that require the events to be independent. I gave three reasons why this would not be true of the games under review

What you really need here is for someone to take a bloody good, long look at this whole scenario in a disciplined and impartial way (ideally someone with proven ability to assess evidence accurately: my ideal person would be a superforecaster). I'm not going to volunteer to do this job myself, I'm afraid.
The simple reveals itself after the complex has been exhausted.
Alexander Schmidt
Posts: 1228
Joined: Thu May 10, 2007 2:49 pm

Re: Kramnik has a point

Post by Alexander Schmidt »

towforce wrote: Fri May 03, 2024 1:17 pm I'd love to watch that video - but it's nearly 6 hours long - so it's a non-starter for me.
The point is: Everybody who wants can watch Hikaru playing. You don't need to watch the whole video, but you get an impression how he plays.
Peter Berger wrote: Fri May 03, 2024 1:35 pm Basically, because his streaks are a little too good to be true mathematically.
I didn't watch the last Kramnik video, but I watched some others. In statistics you have to draw the whole picture. If you play 100 games, statistically there are 50 chances of a winning strike of 50 games. That's what Kramnik doesn't understand. Beside that, playing chess is something else than rolling a dice. If you are 500 ELO weaker and you play against the best, you get nervous. You will find many games in Hikarus strikes where an opponent throws away a won position because of that.

Simple question: Since the rating is based on the results, how can the results be to good for the rating?
Uri Blass
Posts: 10424
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: Kramnik has a point

Post by Uri Blass »

Alexander Schmidt wrote: Fri May 03, 2024 4:29 pm
towforce wrote: Fri May 03, 2024 1:17 pm I'd love to watch that video - but it's nearly 6 hours long - so it's a non-starter for me.
The point is: Everybody who wants can watch Hikaru playing. You don't need to watch the whole video, but you get an impression how he plays.
Peter Berger wrote: Fri May 03, 2024 1:35 pm Basically, because his streaks are a little too good to be true mathematically.
I didn't watch the last Kramnik video, but I watched some others. In statistics you have to draw the whole picture. If you play 100 games, statistically there are 50 chances of a winning strike of 50 games. That's what Kramnik doesn't understand. Beside that, playing chess is something else than rolling a dice. If you are 500 ELO weaker and you play against the best, you get nervous. You will find many games in Hikarus strikes where an opponent throws away a won position because of that.

Simple question: Since the rating is based on the results, how can the results be to good for the rating?
simple answer.

rating is based on previous results.
later results can be too good for the rating that you have.

If it is a small number of consecutive games then no problem with it but if it is a big number of consecutive games then it is logical to suspect cheating by some way.

I do not claim that cheating is the reason and there may be other reasons(for example maybe the player is unstable and there are times that he play when he is tired and lose rating and times when he is not tired and earn rating).
swami
Posts: 6647
Joined: Thu Mar 09, 2006 4:21 am

Re: Kramnik has a point

Post by swami »

Firouzja Alireza had similar streak in lichess bullet. It's possible to have a steady streak - some of the reasons could be:

Not getting paired with similarly strength rated player because second strongest bullet player is still 100 points lower. (It happened with Hikaru, Carlsen and Alireza)

Some players are pretty fast at solving timed puzzles - almost always retain the top position in number of puzzles solved within 3 minutes (Puzzle racer/Puzzle storm - which is available in Lichess and Chess.com)

Some are good with time management. Nakamura had authored the book specifically on tips to optimize the game play in bullet - premove, forced move, sequence, recapture, time management, openings in Bullet

There are other physical aspects - for example, why Asians are good at Table tennis, relatively? and Brazilians at football? This is also applicable in timed games.
swami
Posts: 6647
Joined: Thu Mar 09, 2006 4:21 am

Re: Kramnik has a point

Post by swami »

Uri Blass wrote: Fri May 03, 2024 6:59 am
CornfedForever wrote: Fri May 03, 2024 3:47 am I've seen so many attempts by Kramnik to smear people (like he attempts to with Nakamura here) and justify his math...it's just makes me sick. I could not watch more than 20 min of this one. Sorry. You know what's really funny is that Naka can do what he does while talking to others about the stock market...football...etc. The guy has a gift. Kramnik struggles to even mumble his way through an online blitz tournament.

So, lets think about bullet for a moment. A middling GM named Daniel Naroditsky who has basically retired from OTB chess can win countless bullet games and tournaments against players who are FAR better than him when it comes to normal chess. He goes on many streaks (that's what Arena is largely about) How can he do this? Is he cheating? I don't think so. Of course, Kramnik doesn't play Bullet Brawl and such so he's go not reason to talk about Daniel I guess...unless maybe it would run counter to what he's dishing up (?).

Anyway...Naroditsky plays 1-0 better than about anyone else on the planet consistently online. He uses all the tools and tricks he has to do this...and they do not include engines, just psychology and 'mouse skills'. Online blitz chess is just a step removed from bullet, similar strategies - ESPECIALLY as clocks start running down and they always start to run down. Nakamura has just finished a hair from qualifying for a World Championship match a 2nd straight time. He's really that good. Daniel though...has never sniffed those lofty heights. I just have to wonder why Kramnik speaks openly (now) about the success of Nakamura and Martinez with his 'wink and a nod' pony show and not about the feats of Naroditsky which are arguably far more impressive?
It seems the reason he does not talk about bullet is that in bullet moving the mouse fast is even more important and he prefers to talk about blitz that is closer to normal chess.
Popular format for online Blitz is 3+0 or 3+2. It's more closer to Bullet, than it is to normal (classical) Chess.

It's not just about moving the mouse fast. There are lots of Tactics, Sacrifices and missed opportunities.
Naturally someone who is fast at solving tactical puzzles is also good at Bullet/Blitz, and to the extent FIDE Rapid and Blitz. Ever wonder why certain GM's perform poorly in FIDE rapid in relation with their Classic rating?
Alexander Schmidt
Posts: 1228
Joined: Thu May 10, 2007 2:49 pm

Re: Kramnik has a point

Post by Alexander Schmidt »

Uri Blass wrote: Fri May 03, 2024 4:52 pm rating is based on previous results.
later results can be too good for the rating that you have.
The rating of Hikaru didn't change in the last 4 years. If the rating increases now significantly we can talk about cheating suspicion.
User avatar
AdminX
Posts: 6363
Joined: Mon Mar 13, 2006 2:34 pm
Location: Acworth, GA

Re: Kramnik has a point

Post by AdminX »

Sounds like Kramnik is saying 'If I Can't Do It, It Cant Ne Done'. :D
"Good decisions come from experience, and experience comes from bad decisions."
__________________________________________________________________
Ted Summers
chesskobra
Posts: 215
Joined: Thu Jul 21, 2022 12:30 am
Full name: Chesskobra

Re: Kramnik has a point

Post by chesskobra »

swami wrote: Fri May 03, 2024 6:17 pm Some players are pretty fast at solving timed puzzles - almost always retain the top position in number of puzzles solved within 3 minutes (Puzzle racer/Puzzle storm - which is available in Lichess and Chess.com)
I once saw Nakamura doing puzzle rush or something. It was at lightning speed. Even before you see where is what piece, he moves and a new puzzle appears. Kramnik should watch Nakamura's puzzle rush sometime, and then try to do the same, to get an idea.

Nakamura's defense is also very active. When he is on the backfoot, he keeps creating a lot of activity and problems for the opponent.
Uri Blass
Posts: 10424
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: Kramnik has a point

Post by Uri Blass »

chesskobra wrote: Fri May 03, 2024 7:24 pm
swami wrote: Fri May 03, 2024 6:17 pm Some players are pretty fast at solving timed puzzles - almost always retain the top position in number of puzzles solved within 3 minutes (Puzzle racer/Puzzle storm - which is available in Lichess and Chess.com)
I once saw Nakamura doing puzzle rush or something. It was at lightning speed. Even before you see where is what piece, he moves and a new puzzle appears. Kramnik should watch Nakamura's puzzle rush sometime, and then try to do the same, to get an idea.

Nakamura's defense is also very active. When he is on the backfoot, he keeps creating a lot of activity and problems for the opponent.
puzzle rush is different from chess games and you do not have a new position when you need to see where are the pieces so I do nnot think that it is relevant for rating in games,