andrew writes: "Over the last decade, computers have been able to dominate human chess players. in that time attention has shifted from creating anti-computer strategies to creating computer-resistant chess variants. The inventor of one such game, Arimraa, has an interesting article on Chessbase.com about what it takes to make a board game in which it is still possible for the best human players to remain competitive against computer software."
This discussion has been archived.
No new comments can be posted.
Computer-Resistant Chess Variants
|
Log In/Create an Account
| Top
| 32 comments
| Search Discussion
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
(Score: 5, Informative) by drgibbon on Saturday February 22 2014, @03:15PM
Arimaa is a great game. I do prefer chess myself, but the advantage of Arimaa is that you can learn the rules very quickly. All the pieces move the same, it's just a matter of relative strength. Kids can pick it up really fast and start playing meaningful games more or less straight away. Chess takes a bit more time, as the goals are more subtle and harder to comprehend for a rank beginner. There's already some interesting strategy [wikibooks.org] developed for Arimaa too.
Certified Soylent Fresh!
(Score: 2, Interesting) by Spottywot on Saturday February 22 2014, @03:43PM
Looks like an interesting game, I'll try it. My boy is interested in chess but he's 5 and finding it tough to grasp the principles. I'll learn Arimaa with him and see how he takes to it.
It is an interesting goal to find a game that humans can be better at than computers,(more like current software finds it hard, or doesn't exist)but chess computers have been around a long time. Can't imagine it would take long for a determined developer to devise something to start the ball rolling. Then again I've not played the game yet.
(Score: 5, Informative) by drgibbon on Saturday February 22 2014, @03:55PM
Kids will probably appreciate the proper board [arimaa.com] (with the animal shaped pieces). They sell it through Amazon [amazon.com] for a reasonable price. I'd like a larger board and pieces myself (and some are a bit hard to distinguish from one another), but they're weighted, and not bad at all.
You could play with a chess set, but that can make playing chess a bit difficult later on (once you start to see chess pieces as Arimaa pieces).
Certified Soylent Fresh!
(Score: 3, Interesting) by panachocala on Saturday February 22 2014, @04:03PM
(Score: 3, Funny) by Daniel Dvorkin on Saturday February 22 2014, @03:29PM
Write a program that can beat a human at chessboxing, and I'll be very impressed.
Pipedot [pipedot.org]:Soylent [soylentnews.org]::BSD:Linux
(Score: 2, Funny) by drgibbon on Saturday February 22 2014, @03:34PM
I get the feeling that it shouldn't be too difficult [youtube.com]. Just a few minor adjustments and some special hardware ;)
Certified Soylent Fresh!
(Score: 1) by Daniel Dvorkin on Saturday February 22 2014, @03:57PM
Ah. Perhaps I should have specified "running on commodity hardware" ...
Pipedot [pipedot.org]:Soylent [soylentnews.org]::BSD:Linux
(Score: 2, Funny) by drgibbon on Saturday February 22 2014, @04:02PM
In that case I guess chess-boxing could be considered a computer-resistant chess variant par excellece!
Certified Soylent Fresh!
(Score: 1) by Daniel Dvorkin on Saturday February 22 2014, @04:30PM
Indeed!
Pipedot [pipedot.org]:Soylent [soylentnews.org]::BSD:Linux
(Score: 0, Insightful) by Anonymous Coward on Saturday February 22 2014, @03:33PM
How do you disable popovers like the one in the main link? They're the new popup annoyance and I want to block them.
(Score: 5, Interesting) by regift_of_the_gods on Saturday February 22 2014, @04:06PM
I gave up serious chess long ago so I've never tried Fischer's Chess960 [wikipedia.org], but it sounds like a neat idea for play between human opponents (per TFA's point #1, it won't necessarily help humans regain the edge against computers). One of the discouraging things about walking into a chess club was realizing that players, particularly youngsters, were spending many hours each week memorizing opening theory. That's not necessarily an unpleasant way to spend time, but we all have many other priorities.
(Score: 5, Interesting) by physicsmajor on Saturday February 22 2014, @05:15PM
Exactly. I enjoyed chess quite a bit until I tried to play competitively at a low level and it was clear everyone was parroting the exact same openers. Instead of the fascinating ebb and flow, gambits, and strategy I liked it became "who memorized the most" and was uninteresting.
Even if computers theoretically could still have an edge, what I like about Chess960 is that it forces HUMANS to focus on actual dynamic strategy instead of rote memorization.
(Score: 4, Interesting) by andrew on Saturday February 22 2014, @05:54PM
960 is interesting. My experience with it is that sometimes you get extremely forcing openings in which there is really only a couple of possible moves, this results in a much less interesting game. It turns out one of the most best parts of chess is that the opening position is so amazingly balanced and able to consistently create such rich positions from a variety of openings.
A lot of people at my chess club play the same solid mainline opening variations, so much so that I adopted some offbeat gambits that a lot of people club level rarely see and I sometimes am able to get some advantage since they are unfamiliar with the opening. I used to dislike opening theory until I approached it that way and studied things like the Danish Gambit, Budapest Gambit, and Sicilian Wing Gambit.
But, in reality, you can be a pretty good club player with minimal opening knowledge. Follow the a few basic principles (like get all your pieces out, try and control the center, and get your king castled and to safety) and study a lot lot lot lot of chess tactics and you'll be winning games, eventually.
(Score: 1) by buswolley on Saturday February 22 2014, @10:18PM
Dutch Defense :)
subicular junctures
(Score: 1) by andrew on Saturday February 22 2014, @10:21PM
The Leningrad Variation [chessgames.com] is a fun one I play from time to time.
(Score: 2) by regift_of_the_gods on Saturday February 22 2014, @10:55PM
I had to google Sicilian Wing Gambit. What happens if Black declines with 2... e6?
(Score: 1) by andrew on Saturday February 22 2014, @11:27PM
a3 is a popular move, then if black decides to take ..cxb4, axb4 Bxb4 and then white can put the bishop on b2 and the a file is open for the rook. If black doesn't take the pawn before moving his black bishop white may have the chance after black plays Be7 to play bxc5 forcing the bishop recapture and thus move twice gaining a tempo.
or just Nf3 which the Wing Gambit Deferred.
But there are lots of playable moves at the club level. The original point I was trying to make is that openings like this get players out of their comfort zone of opening memorization and study.
(Score: 5, Informative) by Koen on Saturday February 22 2014, @04:38PM
If you want a board game that humans can win, just play Go: easy rules (more simple than chess) resulting in games which are far too complex for the brute-force approach as used by chess computers.
If you're interested to learn the game, start here:
The interactive way to Go [playgo.to]
Flash Go Tutorial [allaboutgo.com]
A wealth of information to go further:
Sensei's library [xmp.net]
/. refugees on Usenet: comp.misc [comp.misc]
(Score: 2, Interesting) by tdk on Saturday February 22 2014, @05:17PM
go problems [googleusercontent.com] is also a great site although it's down today for some reason.
Take back your internet [squte.com]
(Score: 2, Informative) by Koen on Sunday February 23 2014, @02:05AM
Tsumego (go problems) are indeed the way to progress. Here is a link that works today [xmp.net].
/. refugees on Usenet: comp.misc [comp.misc]
(Score: 2, Informative) by Aiwendil on Saturday February 22 2014, @06:15PM
(Score: 5, Interesting) by TheRaven on Saturday February 22 2014, @06:27PM
Backgammon was quite difficult for computers to win at, because the random element makes the tree of possible games very broad and difficult to exhaustively search. The first AI to win at a competition level used neural networks and so mimicked the approach that a human takes to learn the game. I wonder how many of the proposed games are intrinsically hard for a computer, and how many are just hard for the specific approach used for chess (pattern matching on similar moves then exhaustive search of the space once you reach the end).
I tend to find purely deterministic games to be quite boring. It's quite difficult to get the balance right: you need just enough nondeterminism that it's possible to play perfectly and lose, but not enough that it's likely. A good player should be able to beat a weaker one about nine times out of ten.
sudo mod me up
(Score: 2, Interesting) by Admiral on Saturday February 22 2014, @10:49PM
I also agree Go is the game to play. Additionally, there are very nice game boards that you can get made out of fantastic natural elements (seashells, amazing hard wood, and slate) that really have a great tactile element to playing the game.
If you think you may like Go there's also an Anime/Manga called Hikaru no Go which is enjoyable if you're in to that sort of thing.
(Score: 2, Informative) by Koen on Sunday February 23 2014, @02:24AM
Hikaru no Go is fun indeed. Here [xs4all.nl] is a list of the games in the Manga, and here [xmp.net] is a list of the problems (tsumego) in it.
/. refugees on Usenet: comp.misc [comp.misc]
(Score: 4, Insightful) by stormwyrm on Saturday February 22 2014, @04:45PM
Go [wikipedia.org] is apparently another strategy board game where the best AI programs so far can't even get close to human players at the professional level, much less the best players in the world.
Numquam ponenda est pluralitas sine necessitate.
(Score: 2, Interesting) by melikamp on Saturday February 22 2014, @06:26PM
(Score: 2, Interesting) by regift_of_the_gods on Saturday February 22 2014, @05:27PM
By now it's pretty obvious that if you allow unlimited hardware, humans will lose.
How about a "Man vs. Small Machine" competition where the machine is restricted to be commodity hardware, e.g. a Playstation 4 or an Xbox One with no custom hardware and no use of the GPU except for video; or an equivalent machine based on the AMD Jaguar chip used by those game consoles.
(Score: 5, Funny) by andrew on Saturday February 22 2014, @05:57PM
I still lose to my goddamn smart phone.
(Score: 1) by drgibbon on Saturday February 22 2014, @06:17PM
At chess, yes. But humans still beat the best Go and Arimaa programs pretty easily.
Certified Soylent Fresh!
(Score: 4, Informative) by wjwlsn on Saturday February 22 2014, @06:38PM
I like the idea of human+computer pairs playing chess against each other. Vernon Vinge talked about this in some of his fiction (e.g., "The Peace War") and Gary Kasparov popularized a form of this as "Advanced Chess":
http://en.wikipedia.org/wiki/Advanced_Chess [wikipedia.org]
I am a traveler of both time and space. Duh.
(Score: 3, Interesting) by jcd on Saturday February 22 2014, @08:32PM
This is exactly why I don't really understand trying to compete against a supercomputer at something that's math/strategy based. What about something more unique? Like poetry? Give it a parameter - "autumn" or something of that nature (no pun intended) - and compete against a human opponent to write meaningful poetry in a limited amount of time.
"What good's an honest soldier if he can be ordered to behave like a terrorist?"
(Score: 1) by dyingtolive on Saturday February 22 2014, @10:25PM
Not only is it computer resistant, it's balance resistant!
Don't blame me, I voted for moose wang!