< Earlier Kibitzing · PAGE 58 OF 60 ·
|Nov-01-10|| ||Bridgeburner: <boz>
When I first encountered engines, I was furious that they deconstructed favorite masterpieces. Since then, paradoxically, engines have imbued me with deep and enormous respect for the masters of the past and present.
|Nov-01-10|| ||alexmagnus: <I recommend people try my method on at least one game to see what's involved, as this would inform the debate.>|
<Bridge> On one game I did it once (took me 6 hours of analysis - and I did it <only> forwards, running 16 ply "stadard" and longer if the evaluation was "borderline". Since my own chess level is quite low (1600), I couldn't contribute with much human analysis, but I think I would recognize a "draw mistaken by the engine for a win" in those endgames (the game I chose ended in a middlegame though, so no problems there).
BTW how comes that if the player makes the same move as the engine suggested, the evaluation remains the same in your analysis? I mean, if players did ten "perfect" (in these terms) moves in a row, the eval must change, as you end up not on the same ply as at the origin of the sequence of perfect moves (be it forward or backward).
|Nov-01-10|| ||Bridgeburner: <alexmagnus>
I'd be interested to know which game you analyzed and what your results were!
<BTW how comes that if the player makes the same move as the engine suggested, the evaluation remains the same in your analysis? I mean, if players did ten "perfect" (in these terms) moves in a row, the eval must change, as you end up not on the same ply as at the origin of the sequence of perfect moves (be it forward or backward).>
If a player makes the engine’s preferred move immediately after a 16 ply analysis, the evaluation of that move will usually remain the same because playing the best move maintains the engine evaluation. But this is not always the case, as using a 16 ply evaluation on the next move is effectively turning it into a 17 ply evaluation of the original move.
The further you move along the game, the more likely the engine's evaluation of the "best move" will change as you are expand the engine’s “horizon” by bringing into “view” variations that were previously over the calculation horizon. Quite often, when a large additional amount of information from later analysis is brought back to the original move, not only will an evaluation change, but the engine might change it's mind about its preferred move.
That’s why I end up reversing the sliding process from the last move, so that each subsequent move can take advantage of the information that’s built up in previous evaluations both on the initial forward slide, and from the further information built up on the reverse slide.
The reverse slide will change just about all the initial evaluations from the initial forward slide.
Even these aren’t necessarily conclusive, if the evaluations don’t remain consistent throughout…that’s when variation/subvariation analysis may become necessary so that additional hash tables built up from analysis of variations between the actual moves made in a game can provide the necessary information to fully inform an analysis and therby stabilize evaluations.
|Nov-01-10|| ||alexmagnus: <I'd be interested to know which game you analyzed and what your results were!>|
I don't have that analysis now, I did it just for fun. It was one of Hammer's games from that tournament in Poland where he strongly underperformed.
|Nov-01-10|| ||PinnedPiece: My prediction:
Magnus World Champ by 2015.
|Nov-01-10|| ||Bridgeburner: <alexmagnus>
I mentioned in my last post that <[t]he further you move along the game, the more likely the engine's evaluation of the "best move" will change as you are expand the engine’s “horizon” by bringing into “view” variations that were previously over the calculation horizon.>
A simple well known example of this is that an engine may give you a favorable analysis of a move but it stops just short of a lethal knight fork at the top of a variation tree just over the move "horizon", regardless of whether it's a 16 ply or 25 ply evaluation - no matter how deep the ply, there is always a horizon with who knows what dwelling just beyond.
As you slide the engine along the game's moves, that fork (and other previously unexamined moves) will be eventually factored in and will change the existing evaluation. It can only change the earlier evaluations once you slide the engine back to re-analyze and re-evaluate the earlier moves.
The value of the sliding process is that you're constantly updating the engine's calculations with more information on a cumulative move-by-move basis. The value of the reverse slide is that it has all the information from the initial slide and uses the information gathered all the way from the end of the game to update its original evaluations, and sometimes its preferred moves.
The reverse slide also provides information additional to that acquired on the initial forward slide because the engine's heuristics are constantly gathering fresh information additional to that which has already been acquired.
It's critically important that the engine remain on for the whole process until game analysis and evaluation is complete.
|Nov-01-10|| ||suplexer: Continued motivation not the age determines the age of decline for chess players. kasparov was 2813 rating when he retired at 42. and he only retired not because he was declining but he was bored of chess and couldn't have the world championship match he wanted quick enough. Do you think kasparov would have retired if he was world champion? i doubt it. Even botvinnink was world champion in mid 40's and also Emmanuel lasker. I can assume decline only stars when the world champion of the day stops put in the collusal ammount of practice that they used to, or gains other interests off the 64 squares.|
|Nov-01-10|| ||researchj: Check out this one: www.magnuscarlsen.com|
|Nov-02-10|| ||anandrulez: I think Kasparov would have retired if he was word champ actually , he did have some other interests outside chess - like Russian president .|
|Nov-02-10|| ||Kazzak: www.magnuscarlsen.com
Good thing his chess isn't as prone to errors as the "Norglish."
|Nov-02-10|| ||lostemperor: I read something funny at the chessbomb live chat. Two chess players were chatting at a Carlsen game in Nanjing. One player said to the other: "Carlsen looks like Justin Bieber!" Said the other: "Who is Justin Biber?" :)|
|Nov-02-10|| ||lostemperor: What is also striking is that Anand 40 and Carlsen 19, not coincidently #1 and 2 of the world, are known to have a photographic memory. A significant advantage. Learning new variations in hours instead of days. I don't know which other chess players, except for Morphy, have a photographic memory also? Anand could be playing top chess for a long time still.|
|Nov-02-10|| ||mkrk17: <lostemperor: What is also striking is that Anand 40 and Carlsen 19, not coincidently #1 and 2 of the world, are known to have a photographic memory. A significant advantage. Learning new variations in hours instead of days.>|
Your last two sentences contradict themselves. "Learning" new variations has got nothing to do with photographic memory. It might have something to do with "remembering" new variations, but i am sure it has nothing to do with "learning" new variations.
|Nov-02-10|| ||nimh: <And how exactly did they decide that in his 40 decisive games against Kasparov, Karpov played worse than "the average grandmaster"? That smacks of utter BS to me and calls the whole method into doubt.>|
They did not 'decide', it's the engine that said so. :)
Note also that decisive games usually have lower quality, since to have a decisive result either side has to make a mistake first.
Actually I'm as puzzled as many why it was necessary to make a distinction between results with decisive games included only and those with all games included.
|Nov-02-10|| ||lostemperor: <mkrkr17> right, remembering I mean.|
I hope I remember correctly Anand and Carlsen have photographic memory since I was looking where read it. Anybody knows for sure? Fischer probably had photographic memory too since he could reproduce every single move from a simultanuous display once even after some time.
|Nov-02-10|| ||plang: I suspect most strong GMs have unusual memoty skills. That is a useful tool but a great player needs a lot more than that. Koltanowski set the record for blindfold simultaneous exhibitions. He was a strong OTB player but not an elite one.|
|Nov-02-10|| ||The Rocket: <"Let's assume for a second that computers are strong enough to understand everything world champions have played and can judge them based on their inaccuracies. Well... something isn't right with that technique when it ranks Karpov as more accurate than Kasparov, despite Kasparov winning far more tournaments, getting a much higher rating, and beating Karpov in 5 consecutive matches. Clearly, the computers are missing something.">|
You seem to forget the fact that all of these misstakes kasparov makes which karpov does to a lesser extent is not as evident to a human being as it is to a 3200 elo engine.. in fact kasparovs more risky play though less acurate in theory can be harder to play against than anatolys more sound but less challeninging play, in particular when karpov had the black pieces compared to kasparov(who probably picked up aloth more points than karpov in that aspect)
Rybka is not the truth in chess.. the rybka analyzis is not the final word of the accuracys of world-champs.. if rybka did indeed know this for an answer then it would not lose any games.
when rybka loses a game it sometimes judges that the opponent made a misstake(somewhere in the middlegame) and rybka then plays to refute it. but the fact that the engine still loses games proves that it does not evalute errors and other factors flawlessly.
|Nov-02-10|| ||alexmagnus: <3200 elo engine>
3200 Elo is a myth. It's actually 3200 Swedish national rating. Or do you really think Rybka would beat Anand 9:1?
|Nov-02-10|| ||The Rocket: its 3200+ on this site as well.. and that has nothing to with the swedish site you are referring to|
<"do you really think Rybka would beat Anand 9:1?">
Would be really interesting to see!
Magnus carlsen for instance with his 2826 elo.. would he really beat fritz 8(2800) in a match? of course not.
|Nov-02-10|| ||rapidcitychess: <The Rocket>
The ELO sytem, devised by Arpad Elo, caculates to:
-200 = 25%
In English, that means some one 200 points below you hasa 25% chance of winning. Off the top of my head, 400 points goes to 17%, 600 goes to 9% and zips down to 1%. The upper ranges makes it harder to caculate, because it is almost a point of perfection, but perfection is impossible to achieve until ~30 piece tablebases. Then opening theory can be deemed to a couple exchanges and wind down to a draw.
Off the original subject, the number of positions in chess is ridiculuos. Uncaculable. Finite beings like us will not grasp it. Amazing.
And yet somehow people think that man is all-powerful.
|Nov-02-10|| ||Kaspablanca: lostemperor: Fischer also had a photographic memory, better that Anand`s and Carlsen`s.|
|Nov-02-10|| ||keypusher: <Kaspablanca: lostemperor: Fischer also had a photographic memory, better that Anand`s and Carlsen`s.>|
And Pillsbury's was better than Fischer's. And Morphy's was better than Pillsbury's. And [insert further meaningless, unprovable claims here].
|Nov-03-10|| ||siamesedream: <Magnus Carlsen`s Blog
Nanjing Pearl Spring 2010 finished -Tal WC Blitz n
Nanjing Pearl Spring 2010 finished, Tal WC Blitz next The early 10 am start of the last round of the Nanjing tournament didn't stop the players from showing excellent fighting spirit. Anand outplayed Bacrot with the white pieces to take clear second (with Bacrot clear third), while Wang Yue came very close to his first Nanjing victory. He outplayed Topalov, won a piece, but squandered it all in the time trouble and lost. I had white against Gashimov and didn't have much of an advantage from the opening. Allowing him to capture on e4, the position opened up and his queenside pawns on dark squares became vulnerable. He lost (or sacrificed) a pawn to gain counterplay on the kingside. I missed an opportunity to enter a promising rook and knight ending a pawn up, and instead he traded rooks to reach a drawn but somewhat tricky ending. The queens and opposite-colored-bishop-ending a pawn down proved manageable for him, especially after I wasted a tempo with Qc7+ and a draw was reached after 71 moves. Clear first with 7/10 and a 10 rating point gain made the tournament highly successful for me, and the 24 hours travel back home went smoothly. In many respects the Nanjing tournament was a top-level event of the absolute highest class and I'd like to thank the organizer for the great event. In November I look forward to the Tal Blitz World Championship 16th- 18th in Moscow as well as possibly some pleasant activities with my main sponsors Arctic Securities and Simonsen lawfirm. Finally I'm happy to present my brand new official website, www.magnuscarlsen.com. Hope you will check it out☺. Magnus Carlsen, Haslum, November 1st, 2010
|Nov-03-10|| ||lostemperor: <And Pillsbury's was better than Fischer's. And Morphy's was better than Pillsbury's. And [insert further meaningless, unprovable claims here].> why do you belittle comments on this subject. In chess it is ever more aimportant to have a photographic memory. Of course every top-player needs to have excellent memory today and it is never measured. But sometimes it is documented or backed by eye-witnesses. I found the article on Carlsen photographic memory http://www.chessbase.com/newsdetail... <Magnus' memory is said to be photographic. His coach did a little stunt for some journalists: on TV he showed the boy a diagram from a position in a chess book Magnus immediately replied which game it was, from which book, and roughly how it went. His father also has stories of five year old Magnus reciting the name, size and population of all the 430 counties in Norway. This ability is undoubtedly very useful when keeping up-to-date on modern opening theory. One might assume this has been vital when building up his amazing opening repertoire>|
|Nov-03-10|| ||alexmagnus: <In English, that means some one 200 points below you hasa 25% chance of winning. Off the top of my head, 400 points goes to 17%, 600 goes to 9% and zips down to 1%.>|
Don't know where you get those numbers but they are not true.
200 pts is indeed 1/4, 25%
To get each next 200 step you have to simply multiply with 1/4.
< Earlier Kibitzing · PAGE 58 OF 60 ·