What do you think?
Rate this book
Chess and Computers
A history of the relationship between computers and chess from man's first attempt at mechanized chess to today's programming of computers.
- GenresChess
145 pages, Paperback
First published January 1, 1976
12 people want to read
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 of 1 review
September 25, 2014
Computers have become so depressingly good at chess that it's nice to see a machine getting it completely wrong every now and then. There was a fine example last night in the game between Carlsen (world #1) and Aronian (world #3) at the Sinquefield Cup in Saint Louis. After 51 moves, the players reached the following position, where Carlsen has just moved his king to b4.
The computers providing instant commentary on the various chess sites assessed the position as winning for White. He's three pawns up, right? But every strong human spectator quickly saw that Carlsen had blown it, and that the game was now a dead draw. Black just needs to keep his king in front of the h-pawns and his rook attacking the a-pawn from the side, preventing White from activating his rook. If White, as he is doing here, moves his king to protect the a-pawn, Black checks from the side until the White king heads off toward the kingside; then he goes back his first plan of attacking the a-pawn from the side. White's king has nowhere to hide from the sideways checks, since his pawns are all on the edge.
Carlsen, stubborn as ever, made Aronian play on for another 33 moves, presumably hoping that he would get sleepy and make an elementary mistake; but nothing happened, and they eventually shook hands. The computers continued to insist the whole time that White was winning.
[The Chessbase site has now posted a detailed analysis of the drawing manoeuver here.]
________________________________
Another interesting Chessbase article about chess and computers. As the author says, there is a widespread belief that the incredible advances in the strength of chess engines are solely due to faster hardware. He found an ingenious and dramatic way to test this hypothesis; he took a state-of-the-art program from 2014 and used it to play a short match against a state-of-the-art program from 2006. In order to make the test as tough as possible, he ran the new program on a smartphone and the old one on a fast multicore desktop machine.
The difference in processing speed between the two hardware platforms was a factor of 50 in favor of the desktop - but the new program, running on the vastly slower smartphone, trounced the old one 5-1. Impressive! Evidently, it's not just the hardware.
The computers providing instant commentary on the various chess sites assessed the position as winning for White. He's three pawns up, right? But every strong human spectator quickly saw that Carlsen had blown it, and that the game was now a dead draw. Black just needs to keep his king in front of the h-pawns and his rook attacking the a-pawn from the side, preventing White from activating his rook. If White, as he is doing here, moves his king to protect the a-pawn, Black checks from the side until the White king heads off toward the kingside; then he goes back his first plan of attacking the a-pawn from the side. White's king has nowhere to hide from the sideways checks, since his pawns are all on the edge.
Carlsen, stubborn as ever, made Aronian play on for another 33 moves, presumably hoping that he would get sleepy and make an elementary mistake; but nothing happened, and they eventually shook hands. The computers continued to insist the whole time that White was winning.
[The Chessbase site has now posted a detailed analysis of the drawing manoeuver here.]
________________________________
Another interesting Chessbase article about chess and computers. As the author says, there is a widespread belief that the incredible advances in the strength of chess engines are solely due to faster hardware. He found an ingenious and dramatic way to test this hypothesis; he took a state-of-the-art program from 2014 and used it to play a short match against a state-of-the-art program from 2006. In order to make the test as tough as possible, he ran the new program on a smartphone and the old one on a fast multicore desktop machine.
The difference in processing speed between the two hardware platforms was a factor of 50 in favor of the desktop - but the new program, running on the vastly slower smartphone, trounced the old one 5-1. Impressive! Evidently, it's not just the hardware.
Displaying 1 of 1 review