TsumegoFromGames70/solution

Paths: <= Endgame Problems => · <= Tsumego From Games =>

tderz: problem position in detail view

The full game record is available here: [ext] IgoBerlin4 (W) vs. Han1b4 (B), 16. November 2006, 60min, 15s/5min, W+8.5
It was a game from the 1.Bundesliga (DE), board 4, Han1b4? is [ext] Tsche1 (Tsche Young Il, 2d) (which evolved from [ext] Tsche),
IgoBerlin4? is tderz.

Solution (found by Tamsin): Black lives (just barely!), but White captures 5 stones. Net local result: 7 points for White.

If Black plays first, secures 10 points locally.

Endgame calculations:

The original position has a local count of (10 - 7)/2 = 1.5 points for Black.
A gote play gains 10 - 1.5 = 8.5 points for Black,
or 1.5 - (-7) = 8.5 points for White.

White played in the real game.
White captured 5 stones, but Black made 4 points instead of 3.
So White gained only 7.5 points instead of 8.5. (But that was enough to win the game. :-))

[edit]

Calculation before Bill's comment

tderz: I consider this as the correct solution (best play for both).
Black lives with 2+2 = 4 points (2x + 2x white+circle ),
hence 1 point less than in diagram 30 above (as p won't be a point).
The 5 black+circle are worth 10 points.
The net result is thus
Maximum Territory for Black minus Black's territory loss minus loss by =
White gained = - 10 - (10-4) = - 10 - 6 ,
Black's loss is now 16 points (in Gote for White).
White catches 1 (=2points) more,
loses 1 point at ;
furthermore p is only a dame point (=> 2-1+1 = 2).

This is 2 points better than the 13 points in dia.#30 above.
As both variations end in White's gote, they are directly comparable.

[edit]

After correction by Bill

Above diagram is indeed 1 point inferior than the 17 points in diagrams 30 - 33 as pointed out by Bill.
As both variations end in White's gote, they are directly comparable.

Someone could establish proper diagram links please? I do not succeed.
Bill, could you prepare a deiri and miai counting calculation for the two main diagrams in question? Thanks in advance.

Bill: Is this OK?

[[tderz]: Yes, thank you.

Though I still have problems understanding the local count above.
What if there are several variations - all not bad - say in total 4 - each one in gote, resp. sente for both Black and White?
How to calculate this?

Is this addressed in a page explaining local count?

In a quiet moment I have to do a 'Master edit' of this page
(at present, there might remain some debris of results 13, 15, 16, 17 mixed up)
and compare diagrams 30 - 100 by 3-4 different(!) countings methods:

miai
deiri
local count
my own.

Afterwards I will have understood it for the rest of my life, I hope.

Bill: When we are talking about the size of plays, we are not necessarily talking about good and bad. It is usually right to make the largest play, but not always. In comparing lines of play with, say, Black playing first, we look for one which yields the greatest gain for him if it is a gote or reverse sente, or, if it is a sente, the best local count for him. (The count after the sente is the same as before.)

The count is an estimate of territory (or area), but it is more than that. For instance:

Assuming, per convention, that all stones are alive, each Black group has a local count of 1/2 point of territory. That's more than just an estimate, because, unless there is a ko fight, they add up to 1 point exactly. They are miai.

Miai values tell us how much a gote or reverse sente gains. It does so by comparing the local count before and after the play or sequence of play.

Deiri values compare the local count after each player plays first. They do not tell us how much a play gains, except for reverse sente. And to compare different types of plays you have to make more calculations (essentially converting them to miai values).

Paths: <= Endgame Problems => · <= Tsumego From Games =>

TsumegoFromGames70/solution last edited by tderz on November 25, 2006 - 17:01

RecentChanges · StartingPoints · About

Edit page ·Search · Related · Page info · Latest diff