Amashi/Discussion

Discussion

Adamzero: Are you certain about the definition? From what I can see in the 'Amashi' section of the Game Gallery at Jan van der Steen's GoBase, Amashi is a strategy in which one attempts to destroy any attempt your opponent makes to build territory or a framework, instead of building one of your own. To quote: "Amashi: You usually win a Go game by making more territory than you opponent. Sometimes a pro tries to win by preventing that the opponent makes any territory at all."

NickGeorge: "You usually win a Go game by making more territory than your opponent." That's brilliant.

Warfreak2: You can win by getting less territory but enough captures, so the statement is sensible enough. Also you can win by resign without having made very much territory at all.

On the other hand, looking it up in the InterGo Dictionary at GoBase gives this definition: "a strategy for White in no-komi games in which he lets the opponent take good points but as compensation takes territory, aiming to 'outlast' the opponent."

Can anybody clarify for me? --adamzero

Charles Matthews: Amashi is a type of weak group strategy. Any weak group strategy is Greedy Go: take more territory early on, and hope to defend well enough to win in the end. Territory is taken in exchange for (own lack of?) influence, normally; and influence is what you need to attack.

The real point is that amashi strategy is the sophisticated, pro version. Players of all levels can play big points and hope their groups don't die. But you have to be really strong to 'dangle' weak groups successfully: leave them just weak enough that your opponent can attack them, and really must attack them, but has no attack that can successfully be pressed home. The amashi concept is related to amarigatachi, the creation by the attacking player of overstretched shape in the vain hope of killing a group.

John F.: Much of this page is badly wrong. Charles is almost there. I suspect his last sentence is simply badly expressed, but as it stands it is the reverse of the truth. The attacking player does not create overstretched shape and then hope to kill a group. He tries to kill a group and, being unsuccessful, ends up with no relative profit or a relative loss. His final shape may be overstretched, but often isn't - that's beside the point. The point is simply that his manoeuvre led to no gain or even to a loss. Bob Terry put it pretty well with the phrase muscle-bound shape, though that doesn't cover every case.

Here is an example.

Example 1

W starts 'amashite utsu'  

B ends up with 'amarigatachi'  

B has gained relatively nothing - he has lost the corner and still has a weakness at 'a'.

One Japanese definition of 'amashite utsu' (a better formulation than 'amashi') is: "A high-level technique: while appearing to follow the opponent's orders and doing what is asked, to play in fact on the basis of precise calculation of relative positional gains so that the game will become favourable."

That in itself is not a sufficient definition, but it is followed immediately by 'amarigatachi' which is defined as: "A shape in which you have not made the slightest gain after you appeared to be attacking in good form but the opponent has settled himself. You are in a state where you have suffered from the opponent's 'amashite utsu'.

Those definitions are backed up by the above diagrams, but a Japanese reader would also have the benefit of knowing the core words. The core meaning is ama = "too much" and the core grammar is -su = a causative verb (i.e. force the opponent to expend too many moves) and -ru = aru = to end up in a state (of having overdone things).

Bill: I have seen amashi used in a different way, closer to the first meaning given here, in a fuseki book by Ishida Yoshio years ago. It indicated a strategy principally of invading or reducing moyo or negating influence. Not really taking territory every chance you get, but taking away potential territory or making the opponent's stones inefficient by shinogi.

John F. I can't see what difference you are pointing up, Bill. in the "W starts amashite utsu" diagram seems to fit the example you quote: invading a moyo to take away potential territory and making the opponent's stones inefficient by shinogi. If you are saying that the emphasis is on invading or reducing and that that's the meaning of amashi, well that's plain wrong IMO and doesn't equate with any ordinary meaning of amasu or any definition of the techncial term I've seen.

PS Maybe it's worth adding that Go Seigen claimed that Huang Longshi invented amashi. Even Dosaku hadn't got round to it.

Bill: The difference I had in mind mainly had to do with strategy. The examples that I vaguely recall from Ishida were on a much larger scale, and had nothing to do with "appearing to follow the opponent's orders", which strike me as having a more tactical focus.

John F. Well, it seems to me that a kitten is still a cat. It's true that amashite utsu is a large-scale notion. But large-scale doesn't have to mean moyos or the like. In this case it usually refers to impact on the game as a whole. The Japanese definition does not say anything about tactics, but it very clearly implies whole-board strategy because of the phrase 'precise calculation of relative positional gains' (keisei handan). Invading and so on, like making one weak group, does also surely imply surrendering apparent control to the opponent?

Example 2

Shuwa - Nakagawa  

Bill: This example comes from a game between Shuwa (White) and Nakagawa Junsetsu, in Kenrui Shuwa (Shuwa, the Fortress) by Fukui Masaaki, p. 79. The single kikashi of , followed by the oba of , and then the tenuki to play another oba at are, says Fukui, "a model of amashi uchi".

This is the sort of thing I had in mind, which I did not describe very well, but which also seems different from the definition quoted by John.

John F. This is a very good example, and exactly what I would expect from the definition. From your earlier remark, I'm wondering whether you are reading 'appearing to follow the opponent's orders and doing what is asked' differently from me. As I said, I did not regard the Japanese definition as complete. What I'm supplying (mentally) to fill the gap must differ from what you are supplying. In the example you give, the area where White is going to have to do what he is told is in defending the group that includes 1, 5, 7. He has calculated that any gain Black makes in doing that is outweighed by the "free" moves 3 and 9.

Incidentally, my objection to the very first definition on this page is the notion that it starts with grabbing territory.

Bill: Moi aussi. :-)

That is a possible way of playing amashi but is not the only way. The amashi player may take big points, defend another group, make thickness, etc. The important point is for the amashi player to make a calculation of profit and loss - all kinds of profit and loss, not just territory. You'll see this even better in the games of Huang Longshi. Because of the starting stones and the group tax, it was very difficult in Qing China to play a territorial or moyo game - but it was still possible to execute amashi. BTW I'm not really qualified to say this, but I have a feeling that the Chinese did not develop the concept much further (I've never seen a special term for it - the authority for saying they discovered it is Go Seigen's, not mine), but the Japanese did, and the game above is a splendid example of how they did. It would take us too far astray to explain fully, but (in my view) one direction they developed in was kurai - a term that became almost redundant once hoshi stones became normal in fuseki, and is rarely seen now, but was common up to Meiji times. I'm referring to kurai used without an attributive, as in the modern shogi sense - it means to get "standing" or a strong presence in the centre or other dominant points; not just influence but access to the centre, or control. When the fuseki was mainly on the third line, much effort was expended to achieve kurai. But once they realised that hoshi stones supplied kurai for free, that effort became redundant and so did the term. (I stress, this is my own theory - I've never seen it discussed anywhere else.)

In this Shuwa game, I would argue, White has important kurai with the stone on the lower hoshi point, but the real value of 3 and 9 is not that they (apparently) make some territory at the top, but that they deny Black any centre kurai. If you follow the game through, you will see that Black's thickness on the right gets rather little there, and next to nothing in the centre, but White 3 and 9 also make next to no territory - they end up as a 3-point group. Black's amarigatachi is evident in the final position (he lost by 7), as most of his stones are long lines surrounding nothing.

Bill: While arguably makes territory, is just an oba.

Bob McGuigan: While this is a clear example of amashi it seems to me that the hard thing is not in the recognition but in the execution. This is where very delicate positional judgement comes into play and this is what makes it such a high level technique. I am quite capable of leaving a vulnerable group and taking big points or territory elsewhere but I often lose more than I gained when my group gets attacked. The positional judgement must be exercised fairly early in the game, when there is still a lot of scope for future developments. By the way, Cho Chikun was mentioned earlier as a skillful amashi player and I thought it might not be well known that he admires Shuwa among the classical players, another amashi specialist. Maybe the difficult positional judgement required is in part responsible for Cho's famous long thought in the opening.

John F. That's an interesting last thought, because Go Seigen has been very severely critical of how long Cho spends on the opening (Go was famous as a very fast player, of course, but he and amashi are rarely associated).

Calvin The glossary at the end of Appreciating Famous Games has this definition: "a high-level strategy in which over the board as a whole one permits the opponent to attack one and thus seemingly take the initiative, while planning to catch him unawares when he over-extends himself." I find this translation easier to understand than the phrasing given on this page, especially in the context of the Shuwa vs. Shusaku game given in the book. John, since you translated it some time ago, do you still like that definition?

John F. Yes, although it could go further. I certainly don't agree with the Go Base definitions.

Warp: By the way, if the first definition of amashi given in this page is wrong (as seems to be the consensus), shouldn't it be changed to a better definition (something from pro literature or whatever)? After all, when someone is looking at the meaning of the term here, he will probably think that the first given definition is the correct one. (And another btw: Shouldn't all this discussion go to a separate /discussion page?)

Bill: Here is the definition in the Go Players Almanac:

amashi: Letting the other player do what he wants, typically by letting one's weak stones be attacked, but coming out ahead, anyway. Literally, "having something left over", i. e., having enough territory left over to win the game.

Now, that's quite different, isn't it? (And questionable, I think.) Defining amashi is not easy.

Somebody: Does anyone have Japanese definitions (untranslated) to give here? It would clear up things.

What is being left over? When something is left over, it usually means that you don't know what to do with it. Is it extra moves in the area (might be, as "muscle-bound" sounds like a weird way of saying "overconcentrated"?), weaknesses (which can't be fixed efficiently), thickness that's OK as such but is difficult to put to use in that board position, or something else entirely?

As such, what this page seems to say is that amashi is a strategy where you let your opponent attack a weak group of yours in such a way that the thickness he gets as compensation is inefficient. This may happen by invading an area where the resulting thickness is devalued by the positions of other stones or leftover weaknesses in the shape (which you could not avoid because of tactical considerations). You may also play elsewhere during the opponent's attack, which reduces the value of the thickness indirectly, as the opponent has spent more moves to gain it. When playing elsewhere, moves that also serve to devalue the (expected) thickness become dual-purpose moves and thus more valuable. In short, you only expect to live with your group and that your opponent gets some stones around it as compensation, and then play in a way to reduce their effectiveness.

If the previous paragraph is wrong, it would be nice to have some counterexamples from game commentaries.