Numsgil wrote:
Like is there ever a time where responding to a "sente" play with a threat larger than the current temperature is suboptimal?
Yes. You may be interested in Berlekamp's "Economist's view" chapter in
Games of No Chance, in which he explains Sentestrat and a lot of other things.
Quote:
It just gives us a hard and fast rule for which plays are actually sente.
It is a specialized meaning of
sente. Go players in general will not understand it.
Moi wrote:
Do you mean score or count?
Numsgil wrote:
Bleh, I can never remember which is which, so for the record I have and probably will butcher these terms, and even use them interchangeably.
You meant
score. To call it
gain implies that you are starting from zero. Since I use
gain starting from any position, we could have a language problem if you mean something else.
Quote:
All are correct, but for C, it should be:
Code:
C
/ \
C1 0
/ \
C2 4
/ \
C3 7
/ \
Big 13
Please note that C1 is a 3 point sente, not gote. And that the count at C1 is 7, the same as the count at B1.
With the correction to C, there are two orthodox lines of play:
1) Black plays to C1 in C, White plays to 0 in B, then Black plays the sente to 7 in C, and then to 4 in A, as though it were sente. Result: 11.
2) Black plays to B1 in B, White plays to 0 in C, then Black plays the sente to 7 in B, and then to 4 in A, as though it were sente. Result: 11.
Quote:
Code:
A
/ \
A1 0 -3
B
/ \
B1 0 -4
C
/ \
C1 2 0
/ \ (sente for white)
C2 0 -3
Note that C is gote.
Quote:
But are there endgame sequences where other issues than tedomari prevent the biggest-move-on-the-board-first from reaching the optimal solution (still not counting kos)?
Yes. I used to compose problems where the key was what plays were left at the end, before I realized how stupid that was.
Still, tedomari helps humans recognize such sequences.
Quote:
Is the loss in points always no more than 1?
No, but I like problems where the incorrect play loses only 1 point.
Login
Register
FAQ
Search
Numsgil
