The Game of Go
Go and Chess are generally considered to be the paradigm of strategy games and pure intellectual pursuit. But not many people have heard of Go in the western world, at least not compared to Chess. Anyone that likes chess is doing themselves a grave disservice by not at least trying out Go. I learned how to play Go approximately 6 years ago, but only played sporadically until approximately 2 years ago, when i started to play any chance i could. These days, i try to play at least 4 or 5 games a week, and have seen my strength rise dramatically as a result. Frustrated by a lack of people to play against, this is a small attempt of mine to introduce the game to people, in the hopes of increasing its popularity.
Go is a strategy game from Asia. It is played on a 19 x 19 grid. Two players, one playing black the other white, take turns placing stones on one of the initially empty 361 intersections of the grid. The object of the game is to surround more territory then your opponent. Territory is considered surrounded by a player if there is no sequence of adjacent points in a region which connects to an opponent’s stones. Once placed on the board, stones do not move, and are only removed if all 4 adjacent intersections are occupied by enemy stones. Empty points around a stone are called its liberties, and two adjacent stones of the same color share their liberties. With the addition of a rule, called the Ko rule, which forbids the repetition of a previous board state, these are the only rules to the game.
Playing Go has a quite different feel to it than playing chess. Rather than starting with all pieces at set positions on the board, and then maneuvering them around, the initial board is totally empty, offering black (who goes first) a choice of 361 initial moves. Taking symmetry and transposition into account, he has a total of 90 55 unique moves to choose from. I do not mean to disparage chess; indeed, when I was young I played a lot of chess. But every since I began to seriously play go, I find my taste for chess has waned almost to naught. These days chess feels to me to be confined and lacking in elegance.
A lot of the strategy that happens in go, and one of the things the intrigues me about the game, depends on emergent phenomena that are a direct result of the very simple rules of the game. For example, certain shapes of stones are unconditionally alive. These shapes surround two separate empty positions of one space each. Since either position would be a suicide move for the opponent, the opponent can never take all the liberties of the group. Hence it cannot be killed. Each stone is identical, and so the value of a stone is entirely dependent on the pattern it makes with the rest of the stones on the board.
Conway’s Game of Life was directly inspired by Go. In fact, lacking a computer to run it on, Conway experimented with his “game” using a Go board and its stones. If you are familiar with life, you know about all the strange structures that rise out of its very simple rules. The same thing happens in Go. Certain formations of stones are very strong, others are very weak. Sometimes the difference between a formation being weak or strong depends on who gets to play next. Sometimes you play on one section of the board to a point where one player has an obviously dominating position, and so the play moves elsewhere. But then, as the influence of distant battles propagates across the board, suddenly that dominating position wobbles and falls.
To new comers watching a game, it can seem quite random, with each player plunking stones down seemingly at random, and occasionally taking groups off the board. But to an experienced player, each move has significance. Two moves that might appear practically identical, i.e. the choice between two adjacent points on the board, can and often do have drastically different purposes and significance. For example, consider the first move either player makes. Almost everyone plays the first move in the corner. This is to secure some easy territory, as it is generally the case that a stone behind the 5,5 point can be built into a base. Most people play either on the 3,3 point or the 4,4 point. The 3,3, point is considered an edge facing position, securing the corner quite firmly, but lacking in the ability to project strength towards the center of the board. The 4,4 point, on the other hand, only weakly controls the corner, and is considered a canter facing position. The difference between that two is that from the 4,4 point, your opponent has a 9×9 block of stones in which ti make a position, and it is generally accepted that you cannot stop someone from doing so in the corner unless it is played badly. From the 3,3 point, on the other hand, your opponent only has a 2×2 square in which to try to maneuver, which is simply too tight to do much. The best that can be hoped for is a stalemate in the corner, where each person has a position projecting strength along one side of the board, and one person controls the corner itself, which will generally be the person occupying the 3,3 point.
Often when people see me playing Go with others for the first time, they ask if it is Chinese checkers. While it maybe (very, very) superficially similar, they are very different. The fact is, until you’ve played the game a couple of times, the true complexity of it is not at all apparent. In fact, the first few games people play tend to be confusing. The moves an more experienced opponent makes seem to make no sense, until suddenly they put down a stone and remove a good chunk of yours. Or you find that a position you have been working towards for 10 turns was long ago seen by your opponent, who at that time placed a single stone in a key spot, destroying your position long before it was even formed. It is these kinds of things that set Go apart from Chess, with its more localized character.
In terms of complexity (in the technical, theoretical sense) Go is more complicated then Chess. This is partly because of the drastically larger board space. But it is also caused by the highly abstract nature of the game play. Ever since Deep Blue beat the world chess champion, the AI community has switched its focus to Go. At the current moment, the best Go software players at about an intermediate armature level. The difficulties faced in programing Go are two fold. The first is that the search space is insanely large. The average branching factor of a game tree for Go is 200. As a comparison, for chess it is about 10. The other problem with go is no one has come up with good heuristics to determine what makes one Go position better than another. Without this, standard techniques like Mini-Max cannot be applied. Of course, they cannot be applied anyway when the Mini-Max tree has 200+ children off each branch of the tree.
As is obligatory in any article talking about Chess and Go, a number of obvious differences between the two:
Chess starts with a board crowded with pieces, and removes them one by one. Go starts with an empty board and adds pieces to it one by one.
Chess has complicated rules, go has simple rules.
Chess has several different types of pieces, each of which has different properties. Go has one type of piece, so each piece is identical to every other.
The first move of chess dictates the structure of the whole board. The first move of Go structures only a signal corner.
Chess focuses more on tactical considerations. Go depends in large part of large strategic considerations.
The first moves in chess are a fast paced struggle for domination of the center. The first moves in go are peaceful sketches of future skirmish lines.
If you wish to try out Go, GNU Go is freely available, and you can also play against other people on line for free via the Internet Go Server. In either case, be prepared to have your ass handed to you repeatedly. A wise man once said you should loose your first 50 games of go as fast as possible. The point being that you must bang your brainagainst the go board for a while until certain formations start to glow when you look at them. At that point, you are ready to start learning how to play. You would also be wise to read The Way To Go a free 25 page pdf that explains basic strategy.