Computer game analysis

A tilt? Could be. Still, the leading move 42. d6 is an atypical blunder, there is 3 countermoves to miss (not counting the option of making intermediate checks) before making that move. It looks like white expected black to play Kxe6, are there really players at that level capable of disregarding the opponents option not to accept an offered exchange? Or is there a more subtle explaination as to why a good player would believe that move to be good?

Edit: "It is most likely the case that white was not using computer assistance, and that white missed that after d6 Bxd6 that Rxb6 was impossible because of Bc5+"

Still, if you missed that you would believe Kxd6, Bd8 or a check to be the optimal respons, and that's still not looking good for white.

One thing you have to remember, though, that I heard somewhere that most GMs these days are quite capable of beating engines with time controls of at least 3 days per moves, which is one of the reasons the ICCF allows engine use.

Common, please make the distinction between suspecting cheating and paranoia! What is an atypical blunder? A blunder is just a blunder, period! We hang pieces, we allow mate in one (ask Kramnik about this, he seems to be an expert), we make elementary mistakes in pawn endings. Because we are humans, not machines! It does not mean that we used an engine up to a point and then we started to play alone.

I went through the game using Fritz, allowing it to go to a depth of 18 or so.  White plays good moves but only first/second choice 60-70%, black plays first or second choice 90% of the time, indeed your partial matches were regarded as best by Fritz.  Having said that white was, as has been said very much in the game (-0.26) until he played 42d6, after that he was lost and probably lost the head, maybe he was setting a trap we don't spot :-) 

I notice that different engines sometimes have different views on positions though.  I had  2 games against MirceaH who was later banned.  In the first I was crushed, in the second he got a little overambitious and went for a risky tactical sacrifice, I found my way through it and then he went downhill, eventually resigning.  I got the chess.com comp analysis on both games.  In the first he was perfect, I had 5 or so inaccuracies.  In the second we both had inaccuracies in my opinion but the interesting thing was that in the final position where he resigned the chess.com comp gives him only at a minor disadvantage (edit +0.50)! It also says I made 4 inaccuracies whereas he made none.

 I stuck it in Fritz then and Fritz gave it as something like +2.6 for me....so different engines give different results. He mentioned he was having difficulty accessing the site at the time and was arguing with chess.com

I'd like to see what your Fruit engine has to say about the game.

Of course engines vary. You cannot be sure that one has 100% agreement with an engine unless you find exactly the same engine and run it at exactly the same depths. That's why, if someone wants to analyze cheating, he/she ussually considers the first 3 choices of an engine and then makes the statistics.

It is also not uncommon to have 60-70% agreement with the first move of an engine. A reasonable player can achieve this even in live chess. Sometimes it is possible to have an even higher percentage, if the game has a forcing character. But to me it is unreasonable to play like B in this game. It was a slow, positional game and the vast majority of B's moves were not forced. For many moves there were like 3-4 good alternatives. Yet, amazingly, B almost always chose the best one, even if the differences between them were insignificant.

As for the game posted by MM78: for a human it is quite obvious that the end position is playable, ...Bd6 is a must. Fruit evaluates the position as better for white, starting from +0.90 at depth 12 to +2.23 at depth 19. I don't know what is at depth 20, it takes too long. I also don't have the time to analyze the entire game.

The average correspondence player plays only marginally better than they do OTB; but true correspondence often plays at least a class level better because the make effective use of the time and resources available.  A master level player can play a flawless or even perfect game without "technical assistance" given the days for analysis and the opening and endgame libraries available.

The World correspondence champions from the 70 played much worse than the top players at chess.com. You can find their games and compare. These are the facts, and I can think of just a single explanation. Can someone give another sound explanation? I would be really interested.

Endgames libraries (tablebases) are forbidden at chess.com

it's a good one

To be honest, I am very sad about this issue. It is likely that many people who see such a level of play, and, like myself, think that it can only be achieved by a computer, will either start using themselves an engine or they will switch to other sites. Fortunately, there is a site where, if someone plays a single game like B, he will not play other games.

I would have liked somebody, after reading this topic, to come and say: "your engine is wrong, look, here and here and here how many mistakes B did". Unfortunately, no such thing :( Since B has been checked many times for computer usage and nothing clear has been found, I really begin to ask myself if it is worth playing chess here...

Instead of finding someone in particular to get worked up about, why not focus on all that this site has to offer? Look at the lineup of columnists they've added, and don't you think that this site has a very bright future? It can cater to all levels of player, especially people like me who very much like to listen to what people like IM Pruess have to write about.

I think it is an unrealistic expectation that they will have a completely working anti-computer detection system. The site historically listens to its users, and develops features based on that. Perhaps you're right about Mr. X who plays a pretty darn flawless game as black in the above game. (To anybody who thinks that white uses a computer too, I'm sure white would be extremely flattered).

If you take a long term view on the site, in the big picture, they've made extremely impressive improvements to the site functionality. Know that computer detection is one small part of one small problem of the site. It doesn't affect everybody, and of course there are other ways to cheat than by using computers. In the world of chess website sins, it is not #1. For the record, I would consider the #1 bug in a chess website to be a change to a user-entered move. (Can you imagine?). Perhaps #2 would be dropping a user-entered move, and #2 would be a clock error in turn-based chess.

According to the site operators, for every computer user, there are 99 honest players. Myself included. Perhaps we should play a game? :-)

Do you know this for a fact?  Have you reported B?

I don't think that public discussions of reporting abuse are appropriate.

I disagree.  Naming and shaming in the discussion forums is inappropriate.  There is nothing inappropriate about discussing the reporting of abuse in a thread where it's already been implied throughout that the alleged perpetrator used engine assistance.  The identity of the subject has been obfuscated here so I think that either the entire discussion, from perpetration to punishment, including the reporting, is fair game or the entire discussion is inappropriate.

In any case, I wasn't so much concerned with whether the report was made or not, but rather the apparent assumption that this player had in fact been checked and the implications drawn as a result of that apparent assumption.

Yes, I agree, the entire discussion is inappropriate. Apparently they just don't like us talking about cheating in the forums.

Fair enough then -- if that's the case, this thread should be locked.

You are right Ozzie, sometimes I tend to use a negative tone :(

I do think this site is good. In fact I believe it will soon become a very serious concurrent for ICC and I'm really glad about this. It has great features, like the tactics trainer, which is by far the best tactics trainer. But, yes ... it is more tolerant with the cheaters. That is my opinion. There might be also the problem of lacking human resources to analyze the games.

I also agree that, for 1 cheater, there are 99 honest players. Maybe for 1 cheater there are like 400 honest players. So, overall, the playing experience is quite good.

I have not reported B. But B has been a top player here for a long time (let's say top 30, as a vague estimation). The staff say that they regularly check the top players, so it is impossible that B has not been checked and nothing suspicious has been found. Then why bother to report B again?

Just locking a thread will not make me think for instance that I was wrong and in fact B is a great player. I would say B is at least GM level, maybe >2700.

That's all the purpose of this thread: I'd like to know what is wrong with my analysis. If there is nothing wrong with it, then why B has not been banned? I just picked a game of B, but this player constantly displays a playing strength well above his opponents and I find this playing strength amazing (if it is not an engine).

OK, let's try to turn this into something constructive. Maybe chess.com did not check B. Either by mistake or because of the lack of human resources, even if they claim to periodically verify the top players. I will report B. I am convinced that White played against a computer, since in B's games there are more such examples of his/her amazing playing strength.

Let's see what happens. To be honest, right now I am quite disappointed and I am considering not playing here anymore. Since - at this moment, to me - it looks that, if you do not have 100% agreement with an engine, you can cheat as much as you wish and nobody will bother you. Maybe the staff will prove that I am wrong about this last claim...

This is correct. Just for the record, you should always report a player you suspect to be cheating.