Hacker News Reader

Comment by dethstarr | Hacker News Reader

I love the game of "Go" -- it's so strategic and it's particularly difficult when you're playing against a human. The mobile app I have for it is okay, but not great, probably due to the issues regarding algorithmic complexity. In fact, I like it a bit more than chess.

I urge anyone to try it.

View on HN

One of my best friends is extremely well ranked at chess (his highest ELO was ~2200, he was in France's top 20 of his age group when we were in high school), and has a marked disdain for go.

Any players well versed in the two games care to expound on any key differences in terms of strategy/balance between the two?

The key difference is the philosophy behind both games. In Chess, the goal is to capture the king, destroying his army in the process.

In Go, the goal is to build a territory, destroying is not necessary (smany games finish with very few catpures).

I also like Chess, but I find this "building" philosophy fascinating.

There have been analyses of western and eastern warfare strategies that hypothesize westerners use chess-like strategies (frontal assault), while easterners typically use go-like strategies (surrounding the enemy).

As a first impression, that seems pretty specious to me. I wouldn't at all describe frontal assault as being a normal chess strategy -- the most common would probably be the idea of a multi-pronged attack where your opponent cannot simultaneously defend everything, which is an equally fundamental strategy in Go, at a deeper level than "surrounding."

It's less a difference of Western and Eastern and more of a difference between Clausewitz and Sun Tzu. Clausewitz was all about maneuvering the opposing army into a decisive battle (capture the king), and although you can use clever tactics, it is ultimately a war of attrition favoring the strong (such as industrialized nation-states). Sun Tzu was about using ambiguity so that the weaker opponent has a chance of prevailing over a stronger opponent (asymmetric warfare, or "cheating"). There have been Western military commanders that have used a distinctively Sun Tzu flavor.

Even if you played chess with a multi-pronged attack, you're still ultimately gunning for the king. Your objective is absolute, and each side knows this. Your objectives within the course of a Go game is more fluid and ambiguous.

There is fluidity to objectives during a game of chess between strong players. For example:

inflicting/preventing weaknesses, seeking/avoiding favorable exchanges, occupying/closing important lines.

Then all such long-term considerations suddenly become irrelevant when the game descends into a tactical melee.

At any rate, I don't see how the objective of chess (checkmating the opposing king) is any more absolute or less ambiguous than the objective of go (surrounding more space than your opponent).

Depends on how you want to play. Some people try to win by a certain amount of points. Some people try to lose by a certain amount of points. Some teachers move in a way to call attention to certain shapes or tactical sequences. Sometimes you just want to stick to beautiful moves. When you throw in handicap stones, komi, and reverse komi, how much you win or lose by is arbitrary. You can win by 0.5 points, but that's largely dependent on komi you agreed to. The losing player still has half of the board.

All fluid objectives during a game of chess between strong players are still subordinate to the ultimate objective: capture the king. There's only one king on each side, so there is not much give. There are 361 empty space to choose from in Go.

I don't understand what you're trying to say.

There's only one king, but both sides have 15 additional pieces. One extra pawn is often enough to win, and in the majority of games between competent players one side resigns long before an actual checkmate is on the horizon.

If you want to exchange off the pieces and make a draw, you can. If you're only interested in playing surprising/paradoxical/beautiful moves, so be it.

In go, regardless of the handicap, or whether you want to win by more or lose by less, the objective is still to surround more space than your opponent.

Like asynchrony, I don't understand this argument at all. You can play however you want in chess. It doesn't change the actual victory condition, nor does it in Go. The sub-objectives during a game of chess appear to me to be equally deep and important as the sub-objectives during a game of Go; namely, they are the whole point. Like in Go, you win by conceding some sub-objectives and pursuing others.

Strategies have evolved massively over time and with technology.

My main analogy when talking to chess players: chess is a battle and go is a war. I played chess, now only go. My main problem was the arbitrary set of rules in chess: go is more like axiomatic. Given two players, black and white pieces and a board, it's almost "natural" to develop it

I'm of moderate strength at both, being a chess expert and a 1d-1k KGS. Some brief comments:

- Go makes me confront my fear of heuristics. My unconscious ability to pattern match the right moves is always ahead of my ability to understand why they are right, although I try to catch it up by thinking really hard. It's a unique experience.

- Both the rules and strategy of Go feel more elegant in the mathematical sense of being a composition of simple ideas, which I like. Chess feels more like a set of arbitrary pieces of knowledge.

- The handicap system in Go is an objectively awesome way to have players of different strengths play competitively. In chess, you can almost never play with someone 400 rating points your inferior and have it be a satisfyingly competitive game -- giving piece or pawn odds changes the game completely. In Go, if you give someone four stones, it feels like you're still playing Go.

- When watching strong players, I like the fact that there aren't draws in Go. It makes the game dramatic until the end.

- I like chess problems better than Go problems, and I personally find a level of beauty and variety in amazing chess brilliancies which surpasses what I perceive in great Go moves. I don't know of a Go equivalent to http://timkr.home.xs4all.nl/chess/chess.html.

- At least at an amateur level, it's harder to make a critical, near-irrecoverable mistake in Go -- there's not as much of a snowball effect making an advantage into a bigger advantage. That makes it feel less stressful for me when playing long games.

I especially agree with your first point. In Go, I often play moves just because I find them beautiful. Even in life-or-death situations, I begin by finding the "most beautiful" moves :)

"it's harder to make a critical, near-irrecoverable mistake in Go"

You must not have seen me play :-p

Go problems feel super compelling to me. Typical problems can be boring "counting" exercises, but I find it hard to beat something like the ear reddening move: http://senseis.xmp.net/?EarReddeningMove

There is a great series on youtube on fantastic moves from professional games with commentary in English: http://www.youtube.com/watch?v=CJ9Oexs59CE&feature=relmf...

They are really beautiful in part because they become tesuji through a confluence of factors that reverberate across the entire board. It can be hard to see for amateurs (including me) but once you give them the right context and insight, they become startlingly brilliant.

Also, it's interesting you find it hard to make a critical mistake in Go - this feels very common to me. For example, a decision like deciding to defend a group instead of sacrificing it (which comes up all the time) often snowballs really quickly.

I do that too, but you have a lot of room to back out before your one mistake becomes a losing mistake. If I neglect a group inappropriately, and then make it heavy, and then try desperately to defend it, and eventually fail to survive with no compensation, that's a lot of mistakes; and I usually could have chosen to stop and cut my losses for quite a while before it came fatal.

In chess, you can have long sequences in the midgame and endgame when the game is on a fine tactical balance, and one not-obviously-awful, ill-considered move can either put you in an immediately resignable state or make you spend the rest of the game fighting on the brink of defeat, trying to draw. And that's not really a style of play you can opt out of if your opponent chooses it.

"Both the rules and strategy of Go feel more elegant in the mathematical sense of being a composition of simple ideas, which I like. Chess feels more like a set of arbitrary pieces of knowledge"

That is exactly the reason that I think go (1) may be solved before chess is solved.

For go, there are some results that give hope that an all encompassing theory exists. For chess, the best we have are results of exhaustive searches of relatively simple situations and a bunch of heuristics. It is true that, together, those have led to spectacular results, but I do not think they will lead to a proof about who wins chess.

(1) to be exact: Mathematical go, as defined in http://www.amazon.com/Mathematical-Go-Chilling-Gets-Point/dp.... ko rules can have variations, and there are variations in how to count points at the end of a game. Both may affect only a tiny fraction of games, but a alpha-beta search may need only one path that is a win under ruleset A and a loss under ruleset B to change the outcome of a game.

>At least at an amateur level, it's harder to make a critical, near-irrecoverable mistake in Go -- there's not as much of a snowball effect making an advantage into a bigger advantage. That makes it feel less stressful for me when playing long games.

I think you mean, at an amateur level, it's hard to see the critical, irrecoverable mistakes that you and your opponent make ;)

I feel in chess it's about concentrating your power with the goal of total elimination of the other side in one battle. With go, there are multiple battles where you make trade-offs - short term vs. long term, territory vs. influence. Focusing on solely one part of the equation is almost always not optimal, yet you can still develop your unique style of play. You would often feel it's not a zero-sum game where both sides are taking what they want(though in the end only one can win). Also with the huge branching factor, go feels more dynamic and fluid.

I'm curious, what is it that your friend doesn't like?

I'm a fairly strong chess player (FIDE ELO 2325) and a much weaker go player (best rank around 5 kyu on internet sites).

Chess is a very tactical game. There are many long term strategic concepts, but victory frequently goes to the alert tactical opportunist (and chess engines are the ultimate alert opportunists).

Also, to win a chess game, you ultimately have to attack and destroy some part of your opponent's position. In go you just need one more point of space than your opponent -- there's no requirement to resort to violence at all.

One thing I definitely prefer about go is the openings. In chess they've been so heavily analyzed that games between professionals frequently go 20+ moves before an original move is played. In go there seems to be an almost limitless number of reasonable approaches to the opening. I guess there just aren't as many good moves available in the chess openings.

I'm not sure why your friend would have disdain for go. I think they're both fascinating games of strategy, certainly more engrossing than any RTS I've ever played.

I am 2300ELO in chess and about 4-5k in go (maybe a bit weaker now). My impression is that chess is almost purely tactical/close fight game while go is much more strategic/feel game. Of course that could be because I am much stronger in chess and go overwhelms me a bit. I like go more in general, it has aesthetic appeal which I like. The only drawback is that even blitz games are quite long while in chess you can fire up few 1 0 (1minute per game no increment) on internet in no-time :)

Reply on news.ycombinator.com