When to Tenuki

Recently, the phrase "When in doubt, tenuki!" has been referred to so many times here on SL, that it probably has acquired the status of a Go proverb. Why this is reasonable advice has been summarized by Bill Spight in comments to the teaching game 53:

" [...] in general, it seems to me that kyu players often miss good plays because they didn't even see them. They did not think about throwing stones away, they did not look for dual-purpose plays, they did not consider the whole board.

As for tenuki, I say, when in doubt, tenuki. Sure, you will make mistakes, but when in doubt, you are going to make mistakes anyway. The reason for my advice is psychological. Most people, myself included, tend to get overly involved in the local situation. To the extent that I have been able to find errors by pros, failing to tenuki is a common one. Even they are not immune. Besides, if tenuki is wrong, the refutation is likely to be impressive, whereas, if it is not, then the refutation just looks like the next big play, and the error may go unnoticed."

I fully agree with this advice! I'd say in a typical nine-stone handicap game, I may well catch up around 20 points in the endgame just because my opponent answers my moves too submissively. Instead he/she should sometimes take sente to play elsewhere. And I certainly include myself in the list of people that may get overly involved in a local fight.

But then, I also disagree, and that's why I created this page. (Probably I disagree not so much with the advice itself, but with the way it sometimes gets applied... I would believe -- maybe I should say, I hope -- that our most-respected sensei around here would actually agree with the revisions of his advice. I would be glad about comments!)

If your opponent has just invaded your territory -- please, please, attack him! (Not to kill, but to benefit from the attack, of course.) If your opponent just played an attachment -- please, please, react! In any situation where an additional move makes a huge difference shapewise for both side -- please, take the good shape! If you have a weakish group, and your opponent has one nearby, too -- go and turn the balance of power to your favor immediately!

So how do you know when to tenuki? I'd suggest to rephrase the statement a little:

If you are in doubt, and if it is just about points -- then rush to tenuki!

So if it's about getting superior shape, or about two running groups, chances are you shouldn't tenuki.

Probably I should add examples here, but instead I will begin by discussing this on various game pages where this issue has shown up. See the links listed on the left below "Referenced by" for such discussions.

--dnerra

Bill: Nice page, Dnerra! :-)

I think we are largely in agreement. :-)

On Tenuki is always an option I give some concrete indicators of when to consider playing tenuki.

Tamsin: As an interesting exercise, play through the first 50 or 60 moves of some games from www.gobase.org and count the number of times tenuki is played. Note also under what conditions tenuki is not played.

I have taken the liberty of summarising the above information in the following table:

TENUKI

Profit moves

Plays where you have little potential

Take big point after kikashi

Punish opponent for his tenuki

Already sealed in, life assured

Don't save heavy unimportant stones.

DON'T TENUKI

Shape points

Attachments to important stones

Invasions

Running fights

Weak groups when enemy weak groups nearby

Avoid being sealed in

Save stones that cut apart weak groups.

All the same caveats apply: there will be times when you can tenuki in situations listed under "Don't Tenuki" and times when you cannot tenuki situations found under "Tenuki".

Dieter: I did the exercise. I added the issues about being sealed in.

Imagist: I added the stuff about saving unimportant stones. I have taken the liberty of saying "stones that cut apart weak groups" instead of important stones, because it seems to me that more often than not that's the kind of stone that is important. Stronger players can correct this.