Chess VS. GO

Sort:
u0110001101101000
HGMuller wrote:

19x19 Go is objectively more difficult than orthodox Chess. This can be measured by the rating difference between a top player and a beginner. For Chess this is something like 2000, for Go it would be more like 10,000.

There has been a lot of effort in computer Go the past decade, now that Chess is considered a solved problem by the AI research community. But the problem is that it is completely unclear how to evaluate Go positions. In Chess there is the simple heuristic of 'counting wood' (Q=9, R=5, etc.) which already gives you >90% of the truth in quiet positions. In Go the number of stones is almost meaningless, as large chains can be doomed in th elong term. The 'poor man's approach' of using mobility (number of moves) when you really have no idea of what to evaluate, which works well in Reversi/Othello, also does not work in Go, as the number of moves you have is simply a function of turn number, no matter what you move.

Most of what you said is wrong. How well do you play both games? Do you know how chess engines work? Do you know how well programs play Go? And 2000 vs 10,000?  This is a silly number. Maybe more like 2500 vs 3000. I'm interested with how you thought up 10,000.

DaMaGor
Carmela_Peru wrote:

Maybe I should not even talk, I play quite poor chess and I never played Go, I don't even know how a chess engine work (someone explained me that they evaluate material and space, mobility, controlled squares and so on, but I don't know precisely how they do).

I just remember that a long time ago any Emmanuel Lasker, it seems that he was not so bad as chess player , when he learned to play Go, then he said that if somewhere in the Universe there are evolved aliens, necessarily they have to know the Go.

He said Go, not chess... I find it very surprising.

Right, of course, the quote every go player saying their game is better than chess uses, though that was never the intent.  Still, Lasker was right, and if you know the rules of go, it's clear why: they're very simple and natural.

Take a square grid of lines and two players.  Have them place stones of different colors on the intersections, taking turns.  If a stone has no connection through stones of the same color to an open point, it is removed.  Positions cannot be directly repeated.  When no more meaningful moves are possible, the winner is the player to control more territory, defined as the sum of occupied and surrounded points.

Everything follows from that, with a few tweaks, like komi and some scoring issues.  It's easy to imagine aliens stumbling on the same construction, almost mathematical in its clarity, and discovering what we call go.  (Well, those of us that don't call it weichi or baduk.)

DaMaGor
0110001101101000 wrote:
HGMuller wrote:

19x19 Go is objectively more difficult than orthodox Chess. This can be measured by the rating difference between a top player and a beginner. For Chess this is something like 2000, for Go it would be more like 10,000.

There has been a lot of effort in computer Go the past decade, now that Chess is considered a solved problem by the AI research community. But the problem is that it is completely unclear how to evaluate Go positions. In Chess there is the simple heuristic of 'counting wood' (Q=9, R=5, etc.) which already gives you >90% of the truth in quiet positions. In Go the number of stones is almost meaningless, as large chains can be doomed in th elong term. The 'poor man's approach' of using mobility (number of moves) when you really have no idea of what to evaluate, which works well in Reversi/Othello, also does not work in Go, as the number of moves you have is simply a function of turn number, no matter what you move.

Most of what you said is wrong. How well do you play both games? Do you know how chess engines work? Do you know how well programs play Go? And 2000 vs 10,000?  This is a silly number. Maybe more like 2500 vs 3000. I'm interested with how you thought up 10,000.

And what's a beginner?  Children just introduced to go might merit a hilarious ranking, like 40k, but children just introduced to chess would probably merit a negative rating if such were allowed, and if they learn anything at all they'll improve from that very quickly.  For average adult beginners, maybe 400 and 20k?  Then the best are ~2900 and ~11d (extrapolating from 7d = 1p and one p ranking = 1/3 of a stone).  As long as I'm basing this on factoids I heard somewhere, a one stone difference in ranking supposedly corresponds to a 67% chance to win an even game, or about a 125 Elo difference.  20k is 30 stones below 11d ("0k" becomes 1d).  So 30 * 125 = 3750, compared to 2500 for chess.  Like I said, based on factoids I heard somewhere.  Take it with a grain of salt.  But 10,000 is crazy talk.

So yes, there are more gradations of ability in go than in chess.  Still, what's often overlooked is that the slightest possible advantage at the end of a game of go means a win for that player -- once the game is over, a win is a win, whether by 0.5 points or 50.5 points -- while significant advantages may not be enough to win a game of chess, even with best play.  So in chess, a weaker player has a chance to get a result, even if their position is arguably worse in some sense, that they don't have in go.  Thus, it's harder to prove one's superiority to the same degree as in go, and fewer gradations of ability are apparent.

krudave

I love both games, and I'm ok at them (expert at chess, 5kyu at go), but I have a different attitude towards each.

Both are beautiful, fun, and allow for creative play, but for me, chess is more interesting to study. I mean, the history, the famous players and their styles, solving problems, studying specific aspects of the game. For me, studying this stuff is more fun than actually playing the game (although I still love to play).

Go is less fun to study, in my opinion, but I feel a greater freedom when playing it to do just as I like, make beautiful structures, get away from joseki (openings) and fight it out my own way. I just enjoy playing it.

The one factor that makes me prefer chess overall though, is the combination. These exist in go, kind of, but they lack the variety, beauty, and the often surprising nature of the chess combination. In go, you could sacrifice a few stones and squeeze from the outside, or even sac a whole group for some advantage elsewhere, but nothing really compares to a queen sac, followed by a series of checks resulting in a pawn mate.

KingofBIunders84

I used to prefer Go to chess when I could still play in a club, as I lived in a different town.

Then my level hit a wall at 5 kyu (go ranking), and I didn't get any better.

These days I just occassionally play Go on the KGS server online, due to the lack of offline opponents. When I loose, I have to go to a special room - the KGS teaching ladder - and wait for someone hopefully to come, and explain to me, what I did wrong. This might be valuable, but only with reservation, as the teachers usually are only slighly better then one self.

I once participated in a chess open. I lost the first game, and my opponent an me analysed it together, although he beat me, his conclusions were wrong, as it later showed at the computer-analysis. In chess I always can run over a game with databases and engines, and learn something.

Long story short: I just feel overwhelmed and lost when it comes to Go, therefore these days I prefer chess.

chessmaster102

The teacher ladder thing should be something chess.com adopts or any chess site.