Comparing and contrasting to chess - rules complexities


To be more complex in the comparison with chess, I just read this excellent article where it states half way:

“To many go players chess is the ‘c’ word, but the main reason to mention chess in a go article is not to disrespect it, but to provide a reference point. The two certainly seem like diametrical opposites. Chess has complex rules, go simple ones. Chess eliminates, go accumulates. Chess pieces are hierarchical, go pieces are equal. Chess is a battle, go is a war of many battles. Someone famously wrote a PhD thesis on the theme that in the Vietnam war, the Vietcong were using go strategy while the US military were playing chess. Personally I like chess, but I think go is in a different league.”

Images of olden times

It’s also fascinating that go supposedly has simple rules, yet it has rules disputes and even the end of the game is negotiated, wheras in chess, with it’s “complex” rules, there is very little (AFAIK) ambiguity and disputation.


I’ve been playing a lot more chess lately and thinking about similarities and differences. As I just said, this isn’t the right place for me to go into some (quite interesting, to me at least) ideas I’ve had about the nature of the two games, but I thought I’d add some subtlety to GAJ’s post.

Firstly, there can be more complexity to the rules of Go than you think, and the place where that complexity shows itself it at the end of the game, when the scoring and life / death analysis takes place.

For instance, this position shows a situation called a torazu sanmoku:


You’ll notice that if Black captures the White stones at 1-1, White will be able to play 2-1 and capture the capturing stones, which would let him regain some points and leave Black with a profit margin of only two points. This was felt by 19th-century Japanese rulemakers (possibly Honinbo Shuwa) to be unreasonable as “Black clearly has something here” – so a bespoke rule was put in place ensuring that Black would receive three points without being obliged to take the stones off the board. I’ll admit that the ruling changed in 1989, and it’s now considered seki if Black doesn’t capture, but you see my point. There are also a whole set of post-game procedures in the Japanese rules, the implementation of which is being vigorously argued as we speak in another thread. So, Go often has rules that appear simple to the beginner but reveal themselves later to be more baroque.

Secondly, I don’t think it’s fair to say that chess has complex rules, though I appreciate that complexity is relative. There are only five piece types (king, queen, rook, bishop, knight,) plus the pawns, making a total of sixteen men for each player. Compare that to some large variants of shogi – in the biggest, the mammoth taikyoku shogi, EACH player has at their disposal 402 pieces (yes, that’s four-hundred-and-two) of 208 different types. All of a sudden chess is looking very minimal.
– In chess, all pieces have to capture by landing on the victim’s square. Not the case in many shogi variants; in those, there are pieces that can zap their enemies without moving.
– In chess, each piece has the same movement capabilities all its life (well, apart from the two-space starting pawn leap.) That’s not the case in xiangqi or “Chinese chess”: in that game, once pawns cross to the opponent’s side of the board they receive the capabilities to move left and right as well as forwards.
Shogi takes this idea even further, with many pieces getting bespoke promotions once they reach the opponent’s side of the board. In chess on the other hand, all possible piece types are present at the start of the game.
– Another shogi custom is that you can return captured pieces to the board as your own troops. This complicating feature is not a part of chess.
– Finally, all chessmen can visit any part of the board. Not the case in xiangqi, in which the king and his courtiers are imprisoned in the palace and the elephants can’t cross the river that runs down the centre of the board (despite the fact that elephants can swim.)

So you can see why I consider chess to be closer to classical architecture than baroque.

Now to respond to the idea that the ending of a game of chess is very transparent. Yes, you’re that there is no room for discussion when a checkmate’s on the board. But a strong player won’t play to checkmate anyway if he has too bad a position or he’s down too much material. What you’re actually going to see in most decisive games is an endgame in which one player is trying to win with a slight advantage. Sometimes they have an extra pawn, other times it could just be that they have a bishop and the opponent a knight. These endgames can be as difficult as the middlegame itself, and it goes without saying that they’re much harder than an endgame in Go which is just counting and calculating sente / gote.

That’s my take on things, not that GreenAsJade was actually wrong on any point but that she’s looking at only one perspective (that of a Go player) whereas I think there are other ways to look at the comparison.

#4 begs to differ :smiley:

I never played chess that much to be able to discuss these issues more deeply, but I know there have been several disputes over the years, some even in title matches or whatever (can’t find the most famous now, someone with better memory will be sure to suply instead of me :slight_smile: )

I think that is quite inherent of any more “complex” game that given enough samples you will start finding peculiarities and corner cases…

In my opinion Go is said to have simple rules, because the rules themselves ARE simple - not their aplication. All the stones are worth the same and move the same, there is very little limitation on where you can play, and that’s about it… In chess there is special rule for every piece… That’s (I think) why it is said go has “simpler” rules, but the game itself is more complex…


Nice position! I think I’d rule that as mate but you can see her argument.


It’s an interesting position, but I don’t think there is any ambiguity with respect to the rules of chess that it is a checkmate. The author seems to be making that point, and appears to be presenting it as a curiosity rather than a disputable position.


The way to understand torazu sanmoku is to recall that prisoners in black’s territory are removed from the board at the end of the game, before counting. Meaning, black does not have to close up his own territory with extra moves to capture the white stones. It is sufficient to prove that they can be captured in theory.

In this situation, counting black’s territory and prisoners, black has three points locally on the board as it stands. You can verify this by playing (locally) five black moves and also five white moves, starting with black.

It is important that both players play the same number of stones to keep the score balance of the initial position that we want to evaluate. Even if it means that white has to give a prisoner to black for balance (just put it anywhere in black’s territory).

Now, if black actually plays a move while the game is still on, he will end in gote doing so, and, having filled his own territory, his local score drops to two points.

The only difficult part of this is convincing yourself that the local situation is finished, played out as it stands, with black having surrounded three points, just with a slightly unusual border that is hard to visualize. I suggest that you try to play out the capture yourself, or you can look here for a more formal explanation. I find the wiki entry to be a bit too theoretical and hard to understand.

The three-points ruling is more fair and balanced than the 1989 Japanese rules, under which black has to play it out and lose a point. This is due to the quirky definition of seki, which is applied even though the position is not actually a seki of any kind.

Three points is also the score that you get under area scoring.
Edit: I take that back, I have no idea how to score this under area scoring. Probably 0 points.


You’re quite probably right, I just wanted to support my idea that there are usually some special cases to discuss even though the rules seem clear from the first glance (not that there is no solution) :slight_smile:

This one was also fun :smiley: (although I am pretty sure they overruled eventually)

And a fun fact I don’t think is that known:
In official World Chess Federation Rules capturing an opponet’s king is an illegal move :smiley: so what you maybe thought was the goal of the game is not true :smiley:


I’m not a chess player but from my point of view

  1. Chess rules are hierarchical, i.e. you can derive the solution to a dispute rather easily (cf. comment section of AdamR’s linked article)
  2. Go rules are more like loops, i.e. for some disputes, you cannot derive solutions.

Chess is decidable, Go is undecidable. D:

In other words: chess rules aren’t complex, there’s just a lot of them (pieces more differently, etc). As for Go, there aren’t many rules (especially few related to game mechanics), but they generate more ambiguous situations.

As Adam’s last post indicates, it’s not just easier for someone to make an illegal move in chess without someone noticing (than would be the case in a Go game), but it’s also easier to prove that they did.


I think any discussion about the simplicity or elegance of the rules of go vs chess (and its variants) is ultimately subjective to some degree.

One way to consider it is to think of chess, shogi, xiangqi, and other variants as belonging to a family of related games. Each variant has to define several different pieces, their initial position, how they each move and capture, any special restrictions on movement (e.g., the palace and river in xiangqi), any special movements/captures (e.g., castling and en passant in chess), any other special moves (e.g., putting captures back on the board in shogi), etc. Thus, it seems that the family of chess of games is quite complex in having to define and decide upon many things.

The rules of go are another beast to consider, especially since there is no consensus on the definitive rules of go, but rather several closely related (but distinct) rulesets that result in very similar strategy. Some rulesets, such as New Zealand and Tromp-Taylor, are quite simple and elegant (in my opinion), while others, like Japanese and Ing, are very complicated.


Well, chess, shogi, xiangqi etc. are related games. It’s thought that they all have a common ancestor originating somewhere in or around the Middle East which was then dispersed on the Silk Road and other trade routes. I think you’re right in that there are a lot of variables in the premise of a chess game, but to me it seems that some of those choices have defaults. For instance the default choice in terms of board accessibility, in my eyes, is to have the whole board open to all pieces. So does adopting the default state for a variable mean taking on a more complicated premise? I don’t think so.

I was thinking earlier as well that you can think of the king, queen, rook, and bishop as all logical variations on the idea of moving along orthogonally and diagonally. The bishop can move as far as it wants but only on the diagonal. The rook can move as far it wants but only on the files / rows. The queen is the superset of both since it can move as far as needed both orthogonally and diagonally. And the king is the weakest possible variation of the queen type, since it can only move one space orthogonally or diagonally. So I don’t think the movement capabilities of these four pieces are arbitrary; they bond naturally together as a set.

I think in your last paragraph you touch upon an interesting point: that Go has a significant number of different competing rulesets that all try to achieve roughly the same game. They’re like different engineering methods trying to realise a common architectural ambition. There have been different rules for chess, like whether or not to implement en passant capture, but those have been intended to influence gameplay.


I think that it is worthwhile to distinguish between different types of disputes/debates about rules.

One can have a dispute about the how rules should be applied, and if there is ambiguity that prevents resolution, then perhaps there is a flaw with the rules (or at least a lack of clarity in how they have been written).

Another kind of debate is to analyze the implications of the rules and debate whether or not the rules are aesthetically pleasing and in accordance with the spirit of the game. I think this second type of debate is fundamentally subjective, and the position that you shared seems to fall in that category.


Yes, I’m aware that they are related and evolved from a common origin. I think historians generally believe that the game originated in India, and that Chaturanga is the earliest known form of chess. Also, it seems that the game spread into the west via Persia and eventually via the Moorish conquest of Spain (


I’m familiar with that sort of argument by game theory to support the torazu sanmoku and I respect it as being a valid basis for making a judgement on whether it’s a helpful feature of the rules or not. However, you can’t deny that it’s an example of a ruling made bespoke for a single local position rather than extrapolated directly from the underlying rules. Personally I think that we’re better without it: it complicates the game, making it less minimal and more difficult to understand for players and tournament directors. I’m interested to hear what you think of the torazu gomoku (which in the '49 rules granted Black five points without capturing,) since I’ve read people on Sensei’s Library saying that it doesn’t have the same combinatorial support as its little brother.


And if you support the torazu gomoku ruling then what about “points without capturing” of larger sizes? A torazu of ten or twenty stones?


I think this depends on the ruleset. Under most* area scoring rules, disputes are straightforward to resolve by simply playing it out.

I qualify that statement with “most”, since I think Ing rules can still be quite tricky to resolve, due to the very complicated ko rules, which could easily produce a dispute during play out about how to apply them.

I think it is a bit strong to say to suggest that go is undecidable, but maybe you mean something else. Could you please clarify?

Also, what are some example disputes where solutions cannot be derived? I think there are some historical examples, but I’m not aware of any under the modern rulesets.


There’s no shortage of board positions that present problems for different rulesets.


Do any of those examples demonstrate positions that are impossible to resolve under one of the modern rulesets?

It seems that those examples mainly aim to show how different rules will handle these very differently and how it is very challenging to work though them.


Japanese and Ing rules are complex and have loopholes.

Chinese rules are a lot simpler and I haven’t seen any example of dispute around them (except for cheating or accidental move of stones). Same for other Chinese inspired rules like AGA, EGF, or New Zealand. I believe these are the rules we go players are really referring to when we talk about simplicity.


I’d describe Chess as having more explicit rules, and Go as having more implicit rules. In Chess, if you know all of the rules, you know all of how to play. Everything else is strategy and computation. Nothing complicated can arise as a result of the basic rules. (Most contentious arbiter rulings are rulings related more to etiquette than to actual game play.) In Go, there’s a lot of complexity that can potentially arise from the relatively simple rule set, such as bent four in the corner, for an easy example, or even just dead vs. living shapes. Chinese or Tromp-Taylor eliminate some of those implicit rules, but they’re present in most human games, where people don’t want to manually capture everything at the end of the game.

The implicit nature of Go also creates a lot more ambiguity in certain rule sets, such as whole board repetition when there’s no superko rule. Modern rule sets fix some of that ambiguity, but do so by creating more explicit rules, which are often more difficult to interpret, at least for beginners, than chess rules, since understanding the rules requires some understanding of the game.


Very true. Tromp-Taylor rules are even simpler: having read those a few times I’m astounded we don’t all use them :slight_smile:

So my comment that you quoted, and the subsequent observation by ckersch88 about bent 4 etc, is a red-herring: it’s really based on experience with a poor attempt to capture the rules of Go that somehow took hold, and the fallout from that!