Thoughts about the means and role of capture in board games

I started thinking yesterday about how and why pieces are captured in board games.

Let’s start with the how.

  1. Displacement capture. This is when the attacker captures the victim by moving onto their space. This is the only method by which pieces can be captured in backgammon and international chess.

  2. Leapfrog capture. This is when the attacker captures by leaping over the victim. This is how pieces are captured in draughts, as well as how the cannon (pào) captures in xiangqi.

  3. Long-range capture. This is when the attacker captures the victim without having to move, akin to firing a missile. Some pieces in taikyoku shogi can capture this way.

  4. “Gratuitious” capture. This is when the “attacking” player is given the right to remove a victim from the board, without any of his pieces having to exercise an action. An example of this is the “huffing” rule of traditional English draughts, in which a piece is forfeited if it refuses to make an available capture.

  5. Custodial capture. This method requires two attackers to surround the victim, one on each side. This form of capture is used in hasami shogi and the similar Roman game ludus latrunculorum.

  6. Go-like capture. In Go, of course, a stone or group of stones must have all its liberties removed if it is to be captured, and these liberties are common to the whole group.

  7. Environmental capture, where the capture is made by elements of the playing arena, or “traps”, which the attacker forces the victim into. This is the means of capture in Arimaa.

  8. Explosive capture. This is when the victim is caught in an “explosive” radius of the attacker, which detonates in response to a stimulus. In atomic chess, this stimulus is displacement capture of an opposing piece.

Now, why?

Some games can only be won by capturing all of the opponent’s pieces, eg. draughts.

However, most games which allow piece capture do not have capture as their central goal. In chess-type games, one has to checkmate the opponent’s king-piece by simultaneously attacking it and denying it escape. In Go, capturing is only one of the two ways of making points, and most points in almost any Go game derive from territory rather than capture.

In general, we can say that capturing the opponent’s pieces is a way of weakening their whole-board position, so as to hamper them in the achievement of their goal and to lessen their defensive capability. The idea of capture trades is most sophisticated in the chess-type games, in which piece imbalances of nominally equal value but different specialisations emerge; for instance, two rooks against a queen.


I dunno on this last point. The idea of capture trades is a latency in both games and a high degree strategic concept to be mastered.

In the game of Diplomacy, pieces are sometimes removed (at the choice of the owner of the pieces) because the owner of the piece has lost control of some supply centers, and must disband some units to compensate. Would this be a case of environmental capture or something else?


An interesting example to consider here is Lines of action. The goal of the game is to connect all your pieces, which can actually become easier if you have fewer pieces. So before capturing an opponent piece, you really have to consider if you’re helping their position more than your own.

Of course, capturing not always being beneficial is true for most deep games, but it’s an extra prominent part of the strategy in Lines of action.

Edit: Also, Antichess is an obvious example of a game where capturing is usually bad for you! There capture has to be made mandatory to make the game interesting.

A go equivalent is Anti Atari Go which can become quite interesting even though captures are not mandatory.


That seems like it’s own form of capture to me, a self-capture.

We can add suicide in Ing-rules Go to that category.


The specifics of capture mechanics work quite differently across games, so it can be tricky to analogously call many different types of mechanisms “capturing”. Backgammon is an interesting example, since a lot of people do call “hitting” an opponent piece as “capturing” and it does typically produces a negative effect for the opponent by setting them back. Note: I only say typically, since in some strategic situations, it might be best not to hit when that is an option.

However, this “capture” mechanic of hitting in backgammon does not remove the pieces from the board (but rather places them in a special area called the bar), and crucially it makes those pieces the focus of subsequent play, since no other pieces of that player can be moved until all of the pieces on the bar successfully come off of it. A lot of other “capture” mechanics in other games have the effect of removing that piece from the game (at least temporarily, e.g., shogi).

On the other hand, backgammon does have a mechanic of removing pieces permanently from play, that is, bearing off near the end of the game. I don’t think anyone calls that mechanic “capturing”, since it is actually a positive thing with the entire goal being to bear off all of your pieces before your opponent does so. However, since this act permanently removes pieces from the game, it is kind of like (self-)capturing in a way.

Similar to many other “anti” variants, anti-backgammon is also well-known and what you’d expect. In this variant, the bearing off at the end is negative thing, so maybe one should think of that as a “capturing” mechanic, while, on the other hand, getting hit and sent to the bar is like renewing the life of your piece, delaying its eventual capture (by bearing off).

If we view anti-backgammon as the standard, then it seems quite natural to call bearing off “capturing” (or at least a form a self-capture that one is trying to delay as much as possible). Thus, maybe regular backgammon is really like the “anti” variant, where one is trying to self-capture (bear off) their pieces as fast as possible.


What an interesting point!

1 Like

I find this thread really interesting as I like to organize stuff also. My problem is that I find categorizing capture in board games a bit too ambitious task.

Could you provide a more strict definition of what kind of games are you interested in? You surely don’t want to include Catan or Pandemic in your definition, or do you?

Could you provide a more strict definition of capture? Is it capture, when in Stratego a piece moves onto a bomb? Is it capture when in certain Mancala/maya/bantumi variations you finish your round in an empty hole? Is it capture when the pieces taken can be used by you later?


My 2cts, you first should distinguish when you capture a piece or when you capture a place (take control).


I have always wondered about this. If you checkmate the opponents king, assuming that the game continues, you would be able to capture the opponents king next turn. So the goal to checkmate is equivalent to the goal to capture the opponents king, right?

1 Like

Almost, but not quite. If we reformulate the rules as “capture the king” then putting the opponent into what was previously stalemate will now be winning. Personally I find the simpler formulation worth it, but many chess players seem to enjoy the depth added from stalemate traps.

While on the subject of chess rules, let me also nitpick this part from the original post:

Not quite the only method, since there is en passant! A freak combination of displacement capture and “long” range capture, where you both move a piece and remove a piece on another square.


I think it’s impossible to define rigorously what constitutes a capture in a meaningful way. In abstract terms, a game is completely decided by the game tree. Many games have some concept of “pieces”, since this fits well for physical boards and is easier for human minds to think about. But some games can be reformulated differently if we let go of the physical constraints: for instance the rules of capture go can be stated such that the underlying graph contracts when two stones connect, removing the concept of chains by always making sure the board only contains isolated stones (but it’s exactly the same game, just a different perspective).

That being said, let me throw out some other examples of “capture” mechanics for fun:

In self-capture chess you can capture your own pieces. I think most people would agree that this constitutes a capture. So is it only the act of removing a piece from the board that makes a capture? Then a pawn promotion might also be considered a form of self-capture (you capture your pawn, and then place another piece in its place). I think most people would object to this, and want to classify promotions differently.

Is every move in Reversi a capture? You surround some opponent pieces and make them your own color. But you don’t physically take them of the board, so it doesn’t feel like capturing. If we were playing Reversi with go stones, removing and replacing stones every turn, we might feel differently.

Does firing an arrow in Amazons “capture” the square it lands on? You’re not removing an opponent piece, but you could play the game with a set of loose tiles, and physically remove squares rather than mark them removed by placing a stone on them. This would then be a form of long-range capture, but you are capturing a sort of “neutral piece”, which doesn’t belong to either player.


I don’t think the xianqi cannon captures by leapfrogging. Leapfrogging is how it moves when capturing, but it still captures the piece displaced, not the piece leapfrogged, and thus is still capturing by displacement.


Maybe we could evaluate the difficulty of capturing which could put go as the most?

1 Like

The cannon does not use leapfrog capture. It uses leapfrog movement, but displacement capture.


Reversi exhibits Custodial+Displacement capture. The fact that it is simply the same piece flipped is more a practical concern than a theoretical one.

1 Like

What counts as a capture and how to categorize them are subjective questions. Any attempts to rigorously formalize would require some arbitrary decisions and probably still remains ill-posed. Ultimately, it’s as much a linguistics discussion as anything else.

I think that collecting rent in Monopoly should conceptually count as a form of capture.


Which of these count as capturing?

  • Hitting in backgammon
  • Hitting in anti-backgammon
  • Bearing off in backgammon
  • Bearing off in anti-backgammon
  • None of the above

0 voters

Select none along with something else to indicate uncertainty.

My fuzzy knowledge of xiangqi got called out >__<

I shouldn’t have forgotten about en passant, though, that was sloppy.

1 Like

I always thought the goal of chess was capturing the king. As I understand it wouldn’t change the game a bit.
But for most of history even mentioning killing/removing/capturing a king ended up in wars/executions/higher taxes so I understand why it’s traditional to stop one move earlier.

Edit: I just saw le_4TC 's answer. And it does make a difference.