Prisoner Return

In Mathematical Go, Berlekamp and Wolfe introduce the idea of prisoner return. It is an additional rule to be applied to no-pass Go. With prisoner return, instead of making a move on the board, a player can return a prisoner (taken earlier) to their opponent. This is like a pass stone, except that, first, a player can only return prisoners so long as they have prisoners remaining; second, consecutive prisoner returns do not end the game, instead, as in no-pass Go, the first player without a valid move loses.

Correct play will be for players to fill in their own territory, but two eyes must be left in each group and so there is effectively a group tax, as with stone scoring. As such, it is very close to the oldest known form of scoring.

For what ?

Prisoner return removes a peculiarity of no-pass Go:

The structure of territory matters beyond life.

Example:

W’s turn, no prisoners  

No-pass Go gives White the win after three nakade and one fill.
Prisoner return instead lets Black prevail easily, as expected.
(John Rickard’s perfect example from RGG 1997-09-01.)

Prisoner return still has no jigo

luigi87: It would be very interesting to see examples where this gives a different result than stone scoring.

Bill: That’s easy. On an odd board, if there is not an odd number of dame left in seki and Black plays the last dame, the result is one point less for Black than stone scoring (with possible rare exceptions).^[1] No pass go leads to territory scoring, although how to count territory and prisoners depends upon the form of no pass go.

luigi87: Can you explain why this is seen as closer to territory scoring rather than to stone scoring? The behaviour you describe seems different from territory scoring as well.

Bill: Well, both stone scoring and no pass go with prisoner return have a group tax, so no pass go with prisoner return is different from Japanese scoring and Korean scoring. Usually, if the net score by this form of go is 6 points for Black on the 19x19 board, it will be 7 points for Black by stone scoring, because Black will get the last dame. But with an unfillable dame, it will be only 6 points for Black by stone scoring. The reason is that this form of go does not count the stones on the board, only the territory.

Anon: I think this illustrates Bill’s comment.^[2]

Black to play  

Moves 1 to 10  

Black in Zugzwang  

Black is now forced to play at 'a', and so loses. The game would have been a draw using stone counting (or Chinese counting), so it’s equivalent to half a point komi for White using either of those two methods.

Robert Pauli:

[1] Bill seems to be talking about territory scoring versus stone scoring.

[2] Even if there is an odd number of dame in seki here.

Ya Tienna: Hello everyone, I am an amateur on Github. How about Prisoner to Spy?

W’s turn, one prisoner. B has no prisoner.  

1 was played by W  

B tries to make two eyes, but 2 breaks them. So W wins.