Consequences of Stronger Hardware and Chess Eng...
04 Feb '08 04:512 editsChess engines today are as strong and probably stronger overall than the best humans. It's true that there are still many positions where engines are nowhere near as good as a GM, but it's inevitable that within the next few decades, they will dominate every aspect of the game. Rybka has already reached the 3200 ELO level today (and it is probably accurate for humans not just with other engines).
Perhaps, even further down the line, their strength will be so great that a chess computer the size of a wristwatch could play better than the best humans. This is a bit embarrassing at face value, but more importantly, it opens the way to a whole slew of cheating applications. Perhaps, if there is significant money in chess, a device can be exactly for this purpose (but of course marketed under a different purpose). Maybe the Topalov chess chip some suspected is not so far from reality. Most tournaments wouldn't have the resources to find such devices and even if they did, what would it come to? Would one have to submit to a strip search and a walk through a metal detector? Would one need to be supervised in the restrooms? Even if virtually everyone is honest and it doesn't become a real problem, the suspicion poisons the atmosphere and great games may be marred by accusations.
Moreover, this will undoubtedly have a great impact on opening theory and general play. If you think engines play a big role now, imagine when they're far, far stronger than any human. Now, imagine hundreds of clusters of computers with hardware tens of times more powerful than today's supercomputers studying a position of a topical opening for several months nonstop. I suspect that some openings thought unplayable will be considered playable again due to some novelty verified or produced by engines, but in general, this will make far more openings almost unplayable at the high levels. With the internet and easy access to information from databases, perhaps it will affect even the Class B and up levels. One might think that at the human level such differences would not be so huge. However, the human has knowledge from ELO levels probably around 4000 and he can practice the positions against such opposition. Humans are already far stronger today than they were a century ago and this trend will continue. Memory and time are the major limiting factors.
Chess won't be solved and it will certainly still be considered a draw with best play but the opening moves will be played with far greater accuracy than ever before and deviations will be punished more severely at high levels. A large part of the mystery will be gone and the scope for creativity diminished. This may lead to perhaps only several dozen openings being fully and consistently playable. In the absence of large variety and with the increase in accuracy, even more draws will arise at the high levels.
Its effects on the general game may also be vast. What about chess style? With engines so strong, it will probably be determined what the "best" one is and humans will try to improve themselves and emulate this. It's already clear that an excessively defensive style like Petrosian's would no longer be competitive at the highest levels. Soon, variety in style will essentially disappear because after all it is just a distortion of the reality on the board. Moreover, computer play will become ever more sophisticated and natural (partly because humans will become used to it and partly because engines will become stronger positionally) and it will be indistinguishable from extremely strong and inspired human play. Even if it weren't a human could select the second or third candidate moves or try different engines every few moves if he were attempting to cheat (a worrying subject). Such extreme analysis also suggests that spectacular and completely unpredictable creative feats may be rarer, and when they occur, it will soon be discovered why they don't completely work. While still interesting, most people appreciate truth above all else and so this fact would detract from the beauty.
What about chess's image and prestige? It was once called the touchstone of the intellect and raised on a pedestal so high that God himself could not reach it. However, that was when humans thought only they could master it. Chess hasn't become easier, but like any magic trick, once you see more of how it works- it's less amazing. Titled players will probably not be as respected as they once were and that isn't just because chess will be understood better, but also because they wouldn’t be as indispensible and useful in analysis. Don't know what to do? Turn on Fritz. Don't understand? Let it show you and explain plans. This is already in Fritz 11, albeit rudimentary, but it will become much better.
This is of course just my spur of the moment rambling (only took 45 min) nightmare, but it feels too real- too plausible. Some may say that just as marathoners don't stop running because cars can go much faster and for longer periods, the fact that chess engines will completely dominate should not affect human play. But as I showed, it has and will continue to. The car doesn't teach the runner and running only occurs between humans. Moreover, unlike the car, the computer is easy to hide and seems too similar to what the human is doing. This nightmare is so scary to me that I am concerned about chess's future even for the next century.
I realize that this is exaggerated and there are some benefits that engines have brought about. I myself have gotten a lot of joy out of a better understanding of chess and being able to create and rely on my own ideas after verification. They have only enhanced my appreciation and love for the game thus far and I can't imagine never playing chess again. However, the possibilities are still alarming and not just for chess. Computers will reshape and eventually dominate all games of perfect information (Go included). Humans have always attempted to better understand everything about the world and it's only a matter of time before our curiosity and analysis exhausts our creativity.
Would it bother you if engines completely dominated chess? Do you think it hurts chess? How realistic do you think my fears are?
Thanks for bearing with me and sorry for being so wordy!
Perhaps, even further down the line, their strength will be so great that a chess computer the size of a wristwatch could play better than the best humans. This is a bit embarrassing at face value, but more importantly, it opens the way to a whole slew of cheating applications. Perhaps, if there is significant money in chess, a device can be exactly for this purpose (but of course marketed under a different purpose). Maybe the Topalov chess chip some suspected is not so far from reality. Most tournaments wouldn't have the resources to find such devices and even if they did, what would it come to? Would one have to submit to a strip search and a walk through a metal detector? Would one need to be supervised in the restrooms? Even if virtually everyone is honest and it doesn't become a real problem, the suspicion poisons the atmosphere and great games may be marred by accusations.
Moreover, this will undoubtedly have a great impact on opening theory and general play. If you think engines play a big role now, imagine when they're far, far stronger than any human. Now, imagine hundreds of clusters of computers with hardware tens of times more powerful than today's supercomputers studying a position of a topical opening for several months nonstop. I suspect that some openings thought unplayable will be considered playable again due to some novelty verified or produced by engines, but in general, this will make far more openings almost unplayable at the high levels. With the internet and easy access to information from databases, perhaps it will affect even the Class B and up levels. One might think that at the human level such differences would not be so huge. However, the human has knowledge from ELO levels probably around 4000 and he can practice the positions against such opposition. Humans are already far stronger today than they were a century ago and this trend will continue. Memory and time are the major limiting factors.
Chess won't be solved and it will certainly still be considered a draw with best play but the opening moves will be played with far greater accuracy than ever before and deviations will be punished more severely at high levels. A large part of the mystery will be gone and the scope for creativity diminished. This may lead to perhaps only several dozen openings being fully and consistently playable. In the absence of large variety and with the increase in accuracy, even more draws will arise at the high levels.
Its effects on the general game may also be vast. What about chess style? With engines so strong, it will probably be determined what the "best" one is and humans will try to improve themselves and emulate this. It's already clear that an excessively defensive style like Petrosian's would no longer be competitive at the highest levels. Soon, variety in style will essentially disappear because after all it is just a distortion of the reality on the board. Moreover, computer play will become ever more sophisticated and natural (partly because humans will become used to it and partly because engines will become stronger positionally) and it will be indistinguishable from extremely strong and inspired human play. Even if it weren't a human could select the second or third candidate moves or try different engines every few moves if he were attempting to cheat (a worrying subject). Such extreme analysis also suggests that spectacular and completely unpredictable creative feats may be rarer, and when they occur, it will soon be discovered why they don't completely work. While still interesting, most people appreciate truth above all else and so this fact would detract from the beauty.
What about chess's image and prestige? It was once called the touchstone of the intellect and raised on a pedestal so high that God himself could not reach it. However, that was when humans thought only they could master it. Chess hasn't become easier, but like any magic trick, once you see more of how it works- it's less amazing. Titled players will probably not be as respected as they once were and that isn't just because chess will be understood better, but also because they wouldn’t be as indispensible and useful in analysis. Don't know what to do? Turn on Fritz. Don't understand? Let it show you and explain plans. This is already in Fritz 11, albeit rudimentary, but it will become much better.
This is of course just my spur of the moment rambling (only took 45 min) nightmare, but it feels too real- too plausible. Some may say that just as marathoners don't stop running because cars can go much faster and for longer periods, the fact that chess engines will completely dominate should not affect human play. But as I showed, it has and will continue to. The car doesn't teach the runner and running only occurs between humans. Moreover, unlike the car, the computer is easy to hide and seems too similar to what the human is doing. This nightmare is so scary to me that I am concerned about chess's future even for the next century.
I realize that this is exaggerated and there are some benefits that engines have brought about. I myself have gotten a lot of joy out of a better understanding of chess and being able to create and rely on my own ideas after verification. They have only enhanced my appreciation and love for the game thus far and I can't imagine never playing chess again. However, the possibilities are still alarming and not just for chess. Computers will reshape and eventually dominate all games of perfect information (Go included). Humans have always attempted to better understand everything about the world and it's only a matter of time before our curiosity and analysis exhausts our creativity.
Would it bother you if engines completely dominated chess? Do you think it hurts chess? How realistic do you think my fears are?
Thanks for bearing with me and sorry for being so wordy!
Back to Top