Engines will always play objectively the best move available in the position (not the most interesting move or the move that causes the most confusion/ practical problems for the opponent played by the great human players like Tal & Kasparov).
Can chess engines calculate Mikhail Tal's sacrifices?
Sort:
Interestingly, one particular Tal sacrifice is given by Houdini as the best move in the position. Read this Naroditsky article for more:
https://www.chess.com/article/view/tals-sacrifices-explained
"There are sound sacrifices, and then there are mine." - Tal
"There are sound sacrifices, and then there are mine." - Tal
Yes, at least those based on tactics. However, strategic chess is different. Imagine a diagonal line of pawns. We intuitively know to attack the base pawn. Once it is gone , the next pawn is undefended...etc. However, to think tactically involves way too many calculations. If I do this and black does this...https://en.wikipedia.org/wiki/Decision_tree The power of exponential growth then kicks in. Suppose you put one grain of wheat on square one, two on square 2,4 on square 3, 8 on square 4... " On the entire chessboard there would be 264 − 1 = 18,446,744,073,709,551,615 grains of wheat, weighing about 1,199,000,000,000 metric tons. This is about 1,645 times the global production of wheat in 2014 (729,000,000 metric tons).[8] " FROM https://en.wikipedia.org/wiki/Wheat_and_chessboard_problem Computers can beat us at tactics but not strategically.
When playing a computer make it a strategic game, not a tactical game. Computers can calculate far more moves ahead then we can. But they cannot calculate the amount of possibilities as 1,645 times the world's wheat production!
NMinSixMonths wrote:
Tal couldn't even fully calculate his sacrifices. And I'm sure some of them baffle the engines as well unless they are given ample time to think.
Neskitzy wrote:
"There are sound sacrifices, and then there are mine." - Tal
In other words, Tal himself could not calculate all the possibilities. However, he understood intuitively ( strategically ) that they would work!
Yes, chess is about calculations but it is also art! The greatest players were not only smart they are also artists!
https://sports.ndtv.com/chess/im-improvising-constantly-says-anand-1617586
I have lost many games. However, I also have recorded some of my games that take my breath away. I will never be a GM. So I concentrate on the joy of chess!
[COMMENT DELETED]
[COMMENT DELETED]
doctorpsb wrote:
yes they can. many of Tal's sacrifices are inaccuracies according to top engines.BUT that takes nothing away from the champ.To face them over the board and specially in the pre-engine era was really really tough and that made him the most exciting player ever.But the fact remains that today's top engines would have refuted many of the sacrifices of Tal's.
Can you give examples? Even then that is unrelated to my contention that many combinations require strategic thinking.
Jul 18, 2017
0
#16
In this game I played, the computer thinks the line I played is good for black, but I knew it was good for white. Go on, plop the pgn into a computer and see how it evaluates Ng4. Follow the main line as far as it goes and I bet you it will switch to favoring white from black if it doesn't at first. The specific line I played is incalculable, but it is curious that the computers considers it good for black when it must be good for white. Somehow, Tal's sacrifices are not unsound but TOO COMPLICATED for computers to understand. Maybe understanding how Tal played is the key to beating computers.
BobbyTalparov wrote:
Chess engines make terrible positional moves. They are simply tactical calculators.
This is not true at all. Computers have highly tuned evaluation functions that consider dozens of variables in the positional value of a given static position, and, by crunching many millions of nodes, they can use this accurate per-position evaluation to make good *strategic* moves, not just good tactical moves.
Most of the time, computers are much better positional players than you.
The only remaining weakness they have is that they might not see positional "ideas" that are far beyond their horizon. A human player might have an intuition for the positional advantage a move has 35 moves from now, while the computer has to node-crunch to get there and see the idea pay off, so it might not see if it can only make it to, say, depth 20 in a timely fashion (or much less in the opening). It might also prune positionally strong moves from its search tree too early on because they weren't paying off at the shallower depths of search.
So while it's true they have weaknesses in positional play compared to the intuitions of top human players, it's not at all true that they're "just tactical". Most of the time their positional decisions are actually quite good, because their evaluation function is so well tuned.
If you want to see all the positional variables an engine considers in action, this is a sort of simulator:
https://hxim.github.io/Stockfish-Evaluation-Guide/
Bent Larsen claimed that Tal never even would have played for the World Championship if the players had one hour more on their clocks. Not sure he was right about that, but he meant that his sacrifices were good enough when you didn't have enough time to calculate them. Engines are much faster and better at calculating, but then Tal didn't play engines (or with an extra hour on the clocks).
Jul 18, 2017
0
#19
fabelhaft wrote:
Bent Larsen claimed that Tal never even would have played for the World Championship if the players had one hour more on their clocks. Not sure he was right about that, but he meant that his sacrifices were good enough when you didn't have enough time to calculate them. Engines are much faster and better at calculating, but then Tal didn't play engines (or with an extra hour on the clocks).
Tal was a psychic. He is one of three chess players to exhibit psychic abilities in their play. The other two were Capablanca and Carlsen. If Tal were alive today he would do the exact same thing he did back then. Humans don't stop making mistakes that you can psychically predict just because they're 2800 now.
I always wondered. When i take a look at my chess engine it always recommends good positional moves. But it doesn't even attempt to do sacrifices unless it is for a immediate tactical combination. Can it spots tal's sacrifices? or pretty much any sacrifice, even one that kasparov, fischer or even the few recent games of i have seen of wesley so.