RobertJasiek wrote:
Case II is not that common after the endgame.
Isolated for themselves, your alternative local/capturable definitions are possible but now what does that mean for trying to redo Chris's proof for such other definitions? IIRC from memory, conflicting positions are not known. So maybe one could try a new proof for the alternative situation.
OK, Case II is a bit nearer to "common" than to "abnormal", when regarding the examples.
Chris' proof is no problem. Just leave it as it is or follow my suggestion. (2b1) does not do any harm, it may be only superfluous.
The branch for capturable-1 covers all Nakade and Uttegaeshi.
The branch for not-capturable-1 covers the only remaining position, in which one single stone is transformed to one of only two eye points of a two-eye-formation.
-------------------------------------------------
-------------------------------------------------
But let's try another view.
To not interfere with the definitions in J2003, let us connect
"capturable" to "the opponent can force the capture of the string"
and (thereafter)
"pin1" to "the player can force at least one permanent stone in local-1" and
"pin2'" to "the player can force at least one permanent stone in local-2 minus local-1" (local-2 minus local-1 = "local-2'").
You can assign these definitions to my latest version of Chris' proof respectively as "capturable-pin1" for "capturable-1", "capturable-pin2" for "capturable-2" and "local-2'" for "local-2".
What remains open is your fear that a position could have been "forgotten", in which it would be important to use local-2 instead of local-2'.
Your fear is without any good reason. And, in my opinion, your fear is an unavoidable implication of the multiple-step-multiple-aim algorithm you use in J2003.
I think it necessary to add some motivation for "player" and "opponent" to make the application of "force" within your J2003 well-defined:
- The opponent shall try to reach "capture without any successor" >>> "capture with a successor in local-2 (may be local-2')" >>> "capture with a successor in local-1", in this order.
- The player shall try to reach "uncapturable" >>> "captured with a successor in local-1" >>> "captured with a successor in local-2 (may be local-2')", in this order.
Then you will have no mismatch with different aims in different steps of the evaluation, because there will be only one set of hypothetical sequences only.
By the way: One simple way for the opponent to avoid capturable-1 or capturable-2 would be to capture during the explicit search for "uncapturable", but not to capture during both of the following explicit searches for "capturable-n".
-------------------------------------------------
-------------------------------------------------
Let's return to your fear. I hope that the following text will make my view understandable. But it will be not of the sort you would accept as "proof", I suppose.
#4 is not really an example of this type of position, it resembles a snap-shot mid in the evaluation sequence (originally in #4, nothing had been captured). I suppose it will be extremely difficult to find one position of this type. But anyway, let's go on. I've thought a lot, so I found one example.
Let's bring to mind again that the string under evaluation has been captured.
If the player cannot force a permanent stone neither on local-1 nor on local-2', there are two explanations only.
The first one is simple: It is impossible for the player to have a permanent stone on local-2.
The second one is more complicated: It is not the player, who can force a permanent stone either on local-1 or at local-2', but it is the choice of the opponent. Who shouts: "Ha, ha, if you want capturable-pin1, I will let you have capturable-pin2' and vice versa."
- Click Here To Show Diagram Code
[go]$$
$$ -----------------
$$ | X X X . X X X |
$$ | X O O O O O X |
$$ | X O . O . O X |
$$ | X O O O O O X |
$$ | O X X X X X O |
$$ | O X . X . X O |
$$ | O X X X X X O |
$$ | O O O . O O O |
$$ -----------------[/go]
As can be seen with ease, White's 5-stone-chains are not uncapturable.
- Click Here To Show Diagram Code
[go]$$
$$ -----------------
$$ | X X X . X X X |
$$ | X O O O O O X |
$$ | X O . O . O X |
$$ | X O O O O O X |
$$ | O X X X X X O |
$$ | O X . X . X O |
$$ | O X X X X X O |
$$ | O O O 1 O O O |
$$ -----------------[/go]
Black captures with 1.
- Click Here To Show Diagram Code
[go]$$
$$ -----------------
$$ | X X X 2 X X X |
$$ | X O O O O O X |
$$ | X O . O . O X |
$$ | X O O O O O X |
$$ | b X X X X X a |
$$ | . X . X . X . |
$$ | . X X X X X . |
$$ | . . . 1 . . . |
$$ -----------------[/go]
White gives Atari with 2 and it is Black now who can choose:
- If he connects at "a", White will play at "b" and establish a permanent stone in what is local-2' for her right-hand string and local-1 for her left-hand one.
- If he connects at "b", White will play at "a" and establish a permanent stone in what is local-1 for her right-hand string and local-2' for her left-hand one.
If you remember my suggestion above, you will realise that both players will meet at "capture(d) with a successor in local-2 (may be local-2')" for each of White's strings.
White cannot "force" capturable-1 and Black must not abandon the "force" to "force" capturable-2, what would be against the spirit of the rules.
And, what is important also, this can only happen with strings that will not become part of a two-eye-formation, so are outside the focus of Chris' proof.
The opponent has to capture "symmetrical", the player has to answer with a "symmetrical" threat. This is impossible, if any of the strings in action could become part of a two-eye-formation on its own.