Gary Kasparov 58 Years old is no longer 2812 use your engine

[Pedro]

Could it be that if Kasparov did a training camp he would do better, would be in the middle of the leaderboard?

Just as a hobby, if he hired some strong GMs for sessions of two weeks to a month of intensive training, playing blitz all day etc, wouldn't he be in better condition for this tournament in Croatia?

[lkaufman]

Today he scored 2 out of 9. His fide blitz rating dropped 157 Elo points to 2644. Aside from being rusty, it also shows just how much the level of chess has advanced, especially at blitz due to online play. Now i don’t feel so badly about my own fide dropping 200 points. It is another world now.

So his Blitz rating went down to 2655 approximately, that is average for inactivity and aging

Actually...we are dealing with a very small sample size (and against a small group of highly rated players). If he played another 30 or 50 games...I can't project the trend but it would likely fall a lot further than 2655.

First point is that 2655 is a wrong number, his blitz rating dropped from 2812 to 2801 in 2018 from a high level blitz tournament, indicating that he still played at top level then. But this time he dropped another 157 to 2644 now. He had already been retired for well over a decade in 2018, and played in further FRC events after that, so it's quite strange that he suddenly got rusty and old at 58. I don't have an explanation, and indeed his performance in these latest 18 games was way below even the 2644 rating. It partly shows that blitz greatly exaggerates rating differences, but it also shows that the other players in this event must be really strong blitz players by historical standards, as I'm sure that Kasparov is still quite a strong grandmaster in blitz. I suspect that if there were as many FIDE over the board blitz events for GMs as there are online blitz events, the top guys would have ratings like the 3200 or so seen on chess.com. Kasparov might now be "only" 2900 or so on such a scale. Computer evaluations of blitz game scores show that Carlsen for example is far more accurate than Bobby Fischer was, and he was in a class by himself in blitz half a century ago. Such is the progress of chess with the aid of computers and internet play.

So h0w did you do against Stockfish 14 in a few games of Knight Odds better than with Kom0do MCTS 2 ?

I played a blitz (4' + 2") game at knight odds with SF 14, but I won too easily, it just let me trade off all the pieces. I have almost no chance in blitz with Dragon at knight odds, even GM Lenderman struggled to win under similar conditions. I suppose an engine just can't give knight odds or more without a high Contempt setting. SF 14 doesn't even have a Contempt option, so it's useless for this purpose.

I think that even with default setting there is no reason that the engine is going to allow trading pieces when it is knight down because the evaluation should discourage it and also it is better if the search is tuned to disourage it (meaning the search based on hoping that the opponent may miss some winning lines that are not easy to find in case that it see that the score with normal search is below some number).

contempt is something that should help in equal positions against inferior opponents.
Not something that should help with knight odds because the best practical chance is not to trade pieces regardless of the opponent.

You are talking about the way an engine should play, whereas I was talking about the way Stockfish 14 actually does play. Both Dragon and Lc0 seem to have some understanding that it is better not to trade pieces when you are down a piece, but Stockfish 14 doesn't seem to know this. Regarding the search, MCTS search (in both Lc0 and Dragon) does indeed help when playing a piece down, alpha-beta is bad at this.

[kiroje]

Today he scored 2 out of 9. His fide blitz rating dropped 157 Elo points to 2644. Aside from being rusty, it also shows just how much the level of chess has advanced, especially at blitz due to online play. Now i don’t feel so badly about my own fide dropping 200 points. It is another world now.

I've seen some things happening in openings that were considered to be "bad" about 15 years ago; some profilactic moves, such as moving a knight to a6 instead of c6, to be able to attack a queen coming to b3, with Na6-c5. 15 years ago, nobody would even have considered 'developing' a knight to a6 in the opening.

It just shows how much opening theory changed over the years. Some of the old but revered chess pedagogic books from the 1950's - 1970's (many of them written by Max Euwe, on which many Dutch chess players were trained) could be so far out of date nowadays that they're possibly irrelevant.

I'm not a super strong chess player (FIDE 1800-2000 depending on how much effort I put into a game, with the average opening knowledge of a strong club player), but even I can see that openings and even middle game theories have changed a lot since I studied chess somewhat seriously in the 80's and 90's as a kid and teenager. If I can already see this, it will have a MASSIVE impact at the top level and how chess is played there.

I believe the main difference from when Kasparov was at his prime and now is that the chess engines developed tremendous strength and taught new chessplayers(and of course some older ones) the strength of defensive skills(this i believe was improved immensely) and that openings and middlegames have to be calculated much more concretely than before. The latter is most apparent when really strange openings lines are played, just becuase it works tactically even though they look positionally "hopeless" ( I hope i make sense - else i can elaborate of course )

[jefk]

considering these young GM's being busy with chess
almost all of their time (except Anand maybe),
Gary's performance to me actually looks very good.

Fun to watch btw, the youtube reports from this
Croatian tournament (at least some parts, which i watched,
looking at all the stuff is too much for me).

[lkaufman]

considering these young GM's being busy with chess
almost all of their time (except Anand maybe),
Gary's performance to me actually looks very good.

Fun to watch btw, the youtube reports from this
Croatian tournament (at least some parts, which i watched,
looking at all the stuff is too much for me).

No way to call his performance "good", just compare with how he did in a similar event 3 years ago. He clearly was not happy with his performance; he practiced with Peter Svidler before the event, and clearly did not expect to finish at the bottom. But he says he'll do much better in the chess960 (FRC) event coming up, and I believe him.

[jefk]

LK

No way to call his performance "good"

ok, i now see the final results, apparently he collapsed
in the 2nd half of the tourn (my comments were based
on intermediary results, and as a reaction to the comparison
with the Stockfish analysis in a game or two).
yet overall he remained on par with Jordan van Foreest,
the winner of this year's Wijk aan Zee tournament
and such a level of chess is not nothing; i simply defer
that this was a tourn with strong young blitz players.

[Chessqueen]

considering these young GM's being busy with chess
almost all of their time (except Anand maybe),
Gary's performance to me actually looks very good.

Fun to watch btw, the youtube reports from this
Croatian tournament (at least some parts, which i watched,
looking at all the stuff is too much for me).

No way to call his performance "good", just compare with how he did in a similar event 3 years ago. He clearly was not happy with his performance; he practiced with Peter Svidler before the event, and clearly did not expect to finish at the bottom. But he says he'll do much better in the chess960 (FRC) event coming up, and I believe him.

I strongly believe that any player over the age of 55 will do much better playing 960 (FRC) since they will not have to memorize all the latest opening innovations and only will have to keep a sharp tactical mind, whereas, with regular chess old citizens (over 55 years) have to be very well informed of all the latest chess opening variations.

[lkaufman]

considering these young GM's being busy with chess
almost all of their time (except Anand maybe),
Gary's performance to me actually looks very good.

Fun to watch btw, the youtube reports from this
Croatian tournament (at least some parts, which i watched,
looking at all the stuff is too much for me).

No way to call his performance "good", just compare with how he did in a similar event 3 years ago. He clearly was not happy with his performance; he practiced with Peter Svidler before the event, and clearly did not expect to finish at the bottom. But he says he'll do much better in the chess960 (FRC) event coming up, and I believe him.

I strongly believe that any player over the age of 55 will do much better playing 960 (FRC) since they will not have to memorize all the latest opening innovations and only will have to keep a sharp tactical mind, whereas, with regular chess old citizens (over 55 years) have to be very well informed of all the latest chess opening variations.

That is true at top level, but at amateur level you can play ancient or unsound openings and it will hardly matter.

[Cornfed]

considering these young GM's being busy with chess
almost all of their time (except Anand maybe),
Gary's performance to me actually looks very good.

Fun to watch btw, the youtube reports from this
Croatian tournament (at least some parts, which i watched,
looking at all the stuff is too much for me).

No way to call his performance "good", just compare with how he did in a similar event 3 years ago. He clearly was not happy with his performance; he practiced with Peter Svidler before the event, and clearly did not expect to finish at the bottom. But he says he'll do much better in the chess960 (FRC) event coming up, and I believe him.

I strongly believe that any player over the age of 55 will do much better playing 960 (FRC) since they will not have to memorize all the latest opening innovations and only will have to keep a sharp tactical mind, whereas, with regular chess old citizens (over 55 years) have to be very well informed of all the latest chess opening variations.

That is true at top level, but at amateur level you can play ancient or unsound openings and it will hardly matter.

Even at the top level in traditional chess...if the time control is fast enough - and especially online, suspect/unsound openings get trotted out. Of course it helps if you are a Carlsen and you have the ego to think you can get away with it!

[carldaman]

I watched live as GK had Black vs MVL and he had White busted right out of the opening. He just lacked the technique to convert at this faster time control, which can easily induce blunders, spoiling a great position in just a few quick moves.

Vachier-Lagrave is certainly up to date with current theory, so this may also be a matter of quickness and stamina that diminish with age, and not just an issue of staying abreast with opening theory.