Hi,

I've been building a solver for small boards and reached a knot: I don't how to score seki with certainty (I'm using area scoring). I've been looking the Sensei's pages but still feel confused.

Let me give you an example

Click Here To Show Diagram Code

[go]$$c Seki.
$$ ---------------------------------------
$$ | . O O |
$$ | O . X |
$$ | O X . |
$$ ---------------------------------------[/go]

As I understood, even with both sides without unconditional living groups, if they both pass they have points of their own. The points would be:

Click Here To Show Diagram Code

[go]$$c Seki.
$$ ---------------------------------------
$$ | C W W |
$$ | W . Z |
$$ | W Z M |
$$ ---------------------------------------[/go]

White get 5 points and black get 3 points. White wins by 2 points (without komi).

Am I getting it right?

Yes, you are right.

Note that if we were using territory scoring, white would most likely still win by two points, since there are two more white stones on the board, so white must have taken prisoners at some point.

Thank you!

I'm not so sure, because in Area scoring, dame are "shared" inside a seki, so I would score this game 5.5W:3.5B... Still 2 points difference, but Area Scoring has to add up to the total of points on the board, and Komi is 3.75, so that would give White a substantial win: W+5.75 No?

Yes, I was in doubt about mutual eyes being counted or not. But as it doesn't change the difference in points, I think it is ok any way.

And with perfect play black always win by 9 point on the 3x3 board, so the komi would be 9 points, I think. But I am interested in solve small board "tsumegos", so I'm not using komi.

You are not considering the possibility of black passing a few times. The concept of passing is important for the solver.