Value of 3-3 as a ko threat [#2911]
18.104.22.168: Value of 3-3 as a ko threat
(2013-07-22 00:51) [#9787]
Early in a six-stone game and there's a one-sided ko fight; losing it will be expensive for black (maybe 20 points) but losing it will cost white almost nothing. Neither side has much by way of ko threats yet, though. Trying to invent a ko threat, W plays 3-3 in an almost empty corner -- nothing in that quadrant of the board except the two handicap stones on hoshi points.
My question is how to evaluate the cost of ignoring that.
22.214.171.124: Re: Value of 3-3 as a ko threat
(2013-07-22 16:38) [#9790]
One way to think about it: Komi is 6.5, which is supposed to represent half the value of having the first move. Since the first move is usually a corner move, we can say that a corner move is worth about 13 points. If you approach/come into a corner and B ignores, then the corner becomes equivalent to one where you had the first move and B approached. So to a first approximation (ignoring the appropriateness of a 3-3 move or a 4-4 approach on that board) we might guess it's a 13 pt ko threat.
126.96.36.199: Re: Value of 3-3 as a ko threat
(2013-07-22 21:12) [#9791]
As you describe it the situation is what is termed a "flower-viewing" ko for White. Because White loses very little if she loses the ko, ko "threats" for White are probably abundant, but threats for Black are few. The thinking is that an effective Ko threat has to be big enough to gain more than would be lost if the opponent ignores the threat and connects the ko. Thus White is most likely to be able to find moves worth more than losing the ko would lose. Black has to find moves with big follow-ups for ko threats. As for your example with the 3-3 point, I think ignoring a 3-3 invasion will cost Black a fair amount, though how much depends on the board position. Usually White will get more territory in the corner than she would have if Black had responded to the 3-3 invasion or won't be sealed into the corner.
188.8.131.52: Re: Value of 3-3 as a ko threat
(2013-07-23 17:24) [#9795]
If the difference between winning and losing the ko is on the order of 20 pts., that means that each ko move gains around 7 pts. In the opening, even of a 6 stone game, that is too small to worry about. Answer the 3-3 and let White win the ko.
184.108.40.206: Re: Value of 3-3 as a ko threat
(2013-07-23 23:37) [#9796]
The ko is worth 20 points, but white winning vs black winning is 3 moves difference. So it is worth 20/3 = 6.67 points per move.
Big gote opening moves are worth 13 points, but the difference between white vs black getting a particular big gote point is 2 moves is 13/2 = 6.5 points per move.
So the ko is slightly larger. Ignore white's 3-3 and win the ko.
220.127.116.11: Re: Value of 3-3 as a ko threat
(2013-07-24 20:19) [#9798]
If you instead attach at the 3-4, the threat value is larger, black can subdue it less quickly, and also with more difficulty as it's a handicap game.
: Actual game
(2013-07-28 21:17) [#9809]
Thanks. The posts above are very helpful.
I was Black, chose to connect the ko, and ended up getting beaten by something like 50 points (we neither played out the end game nor counted). I would guess the levels involved at 4 dan vs 12 kyu (see Rank in China for why I have to guess), so I considered that a good result.