“Next month, there’s a world chess championship match in New York City, and the two competitors, the assembled grandmasters, the budding chess prodigies, the older chess fans — everyone paying attention — will know this indisputable fact: A computer could win the match hands down.”
Wouldn’t the same be true here: spellingbee.com/
Or here: jeopardy.com/
You could also say the same sort of thing about any sprint, marathon, etc. – gee, a car could win that hands down. So what? A contest between two humans is still meaningful.
Lately, I’ve been reading posts (mostly on Facebook) by those seizing on this perspective and making bold (I would call them uninformed) claims that chess is on the verge of being solved. I point out that even the best computers, when subjected to look-behind analysis, still make lots of mistakes (mostly errors in judgement), play some types of positions poorly, and have only been able to ‘solve’ endgames with just a handful of pieces/pawns on the board. While computers continue to get better, they are really not that much closer to ‘solving’ chess than they were, say, 10 or 20 years ago. We are long way from that point in time - but anybody with a passing knowledge of chess and some more advanced knowledge of computers makes the regular claim that the great leaps of computing power mean solving the game is just a few years away.
Dream on.
There are a finite number of moves in chess. That said, the old saw is that there are more possible moves in chess than atoms in the Universe. Computers could “solve” chess but we wouldn’t be around to see the day. It’s estimated that it would take at least 10 to the 90th power years with a 1MHz processor. Even with a quantum leap in computing power, it’s probably best not to hold one’s breath.
The size of a computer large enough to consider all possible moves and to “solve” chess would exceed the size of the universe. The length of time to send a single signal across that processor is in the range of 90 billion years - which far exceeds the current age of the universe.
Numbers, numbers, numbers. Whether chess is a sport, science or art, it is still a fascinating game to play. The scientists trying to kill the game with their emphasis on speed and memory have yet to do so. I say “yet” because their efforts in backgammon and other games have been successful in diminishing those games as competitions. DeepMind has turned from chess to Go. As resources by Google and Facebook are applied for their purposes to learn how to effectively search and evaluate information, they will likely make a bigger impact on Go for a time. Whether they are successful at killing Go and chess as games and almost spiritual exercises for its players will depend on us. If we think the games are “solved” because of statistical analysis, then they will diminish in popularity. Whether it is a positive or negative trait, human beings are stubborn and refuse to give up in the face of overwhelming force or evidence. Resistance to the futility of playing chess will remain among the few anarchists that have passion.
"…He has solved checkers and has painstakingly verified that none of the data were corrupt or inaccurate. Vasik Rajlich, an international chess master [profiled in ”Dream Jobs,” IEEE Spectrum, February 2007 described the accomplishment as the latest in a line of games that have been solved computationally. ”Every now and then some game is solved,” he says. ”And now we can ’check this box’ for a rather major game.” So far researchers have solved Connect Four, Qubic, Go-Moku, Nine Men’s Morris, and Awari.
As for the question of solving a game like chess, which people suspect will also result in a draw, the amount of data is even more monstrous. The number of positions in checkers is thought to be roughly the square root of the number of positions in chess. That’s somewhere in the order of 10 to 40th power to 10 to 50th power positions. Schaeffer says that even with the two-pronged technique he used in solving checkers, a breakthrough such as quantum computing would be needed to even attempt to solve chess. But he isn’t quick to rule out the possibility. ”The one thing I’ve learned in all of this is to never underestimate the advances in technology,” he says."
spectrum.ieee.org/computing/soft … ers-solved
It is really going to be hard for the government to provide bread and games once all of the “boxes are checked” and millions of jobs (about 30-40%) are automated, computerized, and robotized out of existence in the next two decades. Elephants are fast disappearing, so no more circuses. Gluten free bread is not very tasty either. The seething mass of discontented unemployed people will find little solace in that so many “problems” have been “solved” by our gleeful computerists. Education or re-education is no panacea. For some of us, it is easy to say “Aprez moi, le deluge”, but the wee ones among us will have a hard time making ends meet while finding joy in a decreasing array of sports and games that will come under the knife of technological progress.
Should that not be “Après nous le déluge”?
I expect that a heuristic for “solving” chess will be refined, partly by loosening the exact meaning of the term ‘solving’.
I think we should be open to such a loosening.
What if by the year 2099 a fellow claims the following: If you give him any position, he will tell you whether White (or Black) can “force a win” from the position, or instead that White cannot force a win. The fellow cannot produce a tabia of all variations from the position, and his computer has never looked at those variations before. The fellow argues the tabia is not the crucial test of having ‘solved’ chess.
Instead, the test is…
Suppose the fellow offers to defend his claim for the position by offering $1,000,000 to any challenger who can avoid defeat by the fellow’s computer from the challenger’s position even with the challenger allowed to use any and all chess computers as assistants. All challengers fail.
Does that qualify as something kinda like “solving chess”? I might say Yes.