Attack while dozens of points ahead
This page is an extended discussion of how the stronger – or at least winning – player should behave when they have established a near-certain win but their opponent does not resign. It started as a counterpart to the page Continue while dozens of points behind, but the scope of the discussion has broadened greatly, so that a better title would now be something like “How to play when you opponent will not resign”.
The following approaches for player in the lead are suggested and/or discussed:
- Play as well as you can:
- Try to maximise your winning margin, i.e. to find perfect moves.
- Attack strongly (as in the page title).
- Consolidate your position and play safe.
- Practise “the difficult art of winning a won game”.
- Try to make the game more interesting / more fun
- Try to win by a specific margin, e.g. ½ a point.
- May not making it more interesting offer unjustified encouragement to the trailing player?
- Try to get your opponent to resign:
- Ignore their ineffectual threats.
- Fill dame.
- Pass a few times.
- Suggest resignation.
- Get out of the game at any cost:
- Resign yourself!
The following types of game are considered:
- A teaching game
- A tournament game
- A game for fun
- A game to estimate the relative strength of players (a Westerner in a Chinese go club)
- On line as opposed to club games
- A game on a go server where you have not yet established your rank
- A game where a complete beginner tries to get a feeling for go
The following motivations for the weaker player continuing are offered or postulated:
- Thinking they may be able to win:
- Being unable to estimate if they are so far behind as to make resignation the best option
- Thinking their opponent might make a mistake
- Misguided tact:
- Thinking (or claiming to think) it would be ungrateful to resign(!)
- For their own benefit:
- Practising, particularly the endgame
- Aesthetic: wanting to see the game completed and/or positions played out
- (They may assume that the opponent feels the same way.)
Other issues raised are:
- What do we play for? Or perhaps, how do we benefit from playing go?
- To win? For the fun of playing? To understand go? For practice?
- Do some players get satisfaction from finding small yose moves to minimise their losing margin?
- Can playing out a clearly settled game still be interesting?
- How far should one treat a win by 1 point as equivalent to a win by 100?
- Etiquette and culture
- Is it rude of the losing player not to resign?
- Nakayama Noriyuki is quoted as suggesting that if they do not they are insinuating that they think they can still turn the tables.
- May/should the stronger player suggest resignation?
- (The page Suggest resignation considers it acceptable in a teaching game.)
- Is it not much worse to complain when you have lost than to attack the losing player?
- Should the losing player should ask his opponent if he minds continuing?
- (A number of players said this was their practice)
- Perhaps cultural issues such as loss of face sometimes prevent a much stronger player resigning when that would be a way out of a tedious game.
- Is it rude of the losing player not to resign?
- How should the losing player behave:
- Should they try desperate measures to achieve an upset? (And perhaps resign if that fails?)
- How far behind (how many points) should they be to resign?
- Can they be proud of a narrow loss (after playing it out rather than resigning)?
- If they do not resign, they may deprive themselves of the lessons of the analysis or other players of the chance of a game (teaching or otherwise).
- How strongly should a teacher play?
- Playing trick moves or impossible invasions appears to be disparaged.
- Is it not bad to overwhelm the pupil?
- Some pupils object to them steering the game to a narrow White victory, preferring to be attacked powerfully.
- How many mistakes should they leave unpunished?
- Isn’t it better that they play moves appropriate to the pupil’s level all the way through rather than grabbing a lead and then easing off?
As a lower ranked player, I occasionally see a game where the stronger player attacks for the maximum point difference win, rather than letting a weakness or two pass. I’m not suggesting that the stronger player should deliberately lose; rather I’m saying that it would be more polite to stop playing and count the score when it's obvious that the stronger player is ahead,
DougRidgway They may not be toying with you — perhaps they are just practicing the difficult art of winning a won game. It’s not over until there are two passes or someone resigns.
jfc: Is this an attempt at humor? If so, your act still needs a bit of work.
Harleqin: jfc, I am not sure whether You refer to the page or to Doug’s comment, but I don’t think this page is intended as humour. If you are ahead, you should strive to simplify the game by securing your positions and covering your weaknesses. Then you can safely bring home your victory. Just attacking like it doesn’t matter who is ahead often gives the opponent a chance to stage an upset which obviously you don’t want, especially having been ahead before. Knowing how it stands, and playing accordingly, is an important part of go strength. This is not even a question of politeness.
jfc: I was referring to ggleblanc’s original comment. I think it is a troll (in which case it is working as we are responding) or an attempt at humor.
While following the proverb simplify when ahead can improve your winning percentage, there are other factors to consider.
In ggleblanc’s scenario the players are of clearly different skill levels with the stronger player far in the lead. ggleblanc’s suggestion that the stronger player should “stop playing and count the score” suggests a teaching game or, at the very least, a friendly game. In this case the stronger player continuing to attack when she already has a clear win may be a strategy to combat boredom until her opponent resigns.
Scryer: “Simplify while ahead” makes sense only in the context of a contest where the outcome is in doubt: it’s a heuristic that helps you try to win a close game when you’re ahead. In the case hypothesized, the player who’s ahead will win; the only questions are when and by how much. What should happen is that the losing player resigns when their loss becomes obvious. The stronger player has several options available, including: (1) Kill another dragon and make the loss more obvious to the loser, in hopes of terminating a game whose aji has been keshied; (2) Fill dame and ignore losing invasions, hoping that this will encourage the loser to pass; (3) Amuse yourself by trying for a specific winning margin: 1/2 point, for example. It’s going in the ‘W’ column anyway, and if you’re as much better as you think you are, you may as well play with your food while you’re waiting.
Tas: I dont really see why go has to be so much about either winning or losing. If the opponent made a mistake and lost a big group in one corner, getting hopelessly behind, why not continuing to play equally on the rest of the board? Why does the one point between loss and win matter, while the rest do not?
You might not have time for another game, but time to find the outcome of another interesting fight, and practice some endgame, though it won't matter to the overall outcome.
Most of us play to play, not to win, right?
If I was the one behind, I would expect my opponent to continue to do his best attempt to play perfectly, making the win as large as possible; while I’d try to minimize that difference. But of course I’d ask first if he minds that I do not resign.
Dieter: I play to understand Go better. The purpose in the game itself is to win. If it doesn’t matter who wins, then what’s a good move? If there are no good moves, what’s there to understand? I do not play just to fill the board.
I agree that maximizing/minimizing the win may be a good exercise, but then you must set these goals beforehand. Silent convention is that B+1 = B+100.
Tas: It does matter who wins. I just don’t see why the rest shouldn’t.
I guess what I am asking is why B+1 = B+100? Of course it’s the easiest in tournaments, but many people do not play tournaments. The game is what the players make it, right?
Anyway, perfect play would be the same, regardless.
Bob McGuigan: Of course you can continue playing, trying to optimize every move. In my games, though, when one person loses a big group, say, it makes the rest of the board uninteresting. It is too easy to play in other parts of the board for the person winning the group.
An interesting related point, though, is what happens in handicap games. When one player loses a big group in an even game it gives the other player a huge lead in the game. When you receive 5 handicap stones you are starting out with a huge lead (60 points?). So continuing to play after losing a big group is simply accepting a big handicap from your opponent. Of course in a handicap game one player is a lot weaker than the other and the handicap is supposed to equalize the game. If two equal strength players are playing and one loses a big group he has accepted a big handicap from someone of about equal strength.
Malweth: I disagree that B+1 == B+100. Certainly there is no difference in result, but if all you care about is results go will be a very hard game for you. The difference lies mainly with White. In the case of B+1, White has played a good game and lost with small margin. In the case of B+100, White has continued to play a hopeless game, failed to understand resignation, and should be embarrassed.
There is also a difference for Black - B+1 is a closely won game and was likely a good fight and equal match. B+100 is either a teaching game against a weak kyu student or a boring game for Black. I’m of the opinion that, unless teaching weak kyu, a kyu player should be within 25 stones of winning. Anyone who’s learned to count (weak to mid Dan players) should be within 15 stones of winning. Of course, one should only resign if there is no possible strategy for comeback :)
Dieter: Hmm. I didn’t mean that the nature of games ending in +1 or +100 is the same. But there is no sense in taking pride in a 5 point loss. It happens you know, playing White in a high handicap game, playing your very best to squeeze the most out of it, count and recount and confidently win by five points and then the opponent says “Oh, five points, I did well against you!” I’m speechless when that happens. The reasoning behind such statements is very far from Go. You can also lose by resignation and still having done well, because you gave it a tough fight. There is no shame in that, on the contrary. There is shame in playing cowardly because you don’t mind losing if it is only by a small margin.
Malweth: True... I guess since I don’t play like that I didn’t look at it that way. Relative strengths, however, always need to be examined separately. There’s nothing wrong with the Black player in a handicap game being pleased with a close loss — especially if the handicap favored the white player. I remember a time when I would have been very pleased with a W+5 win versus my teacher with 9 stones. Now, when that handicap seems numerically appropriate, I don’t feel the same way. With players of equal strength, a close loss usually indicates a good game. When the losing move is relatively small, a good game is indicated. A loss due to a misread in fighting is a bigger failure, IMHO.
Bill: Well, if the opponent keeps missing the opportunity to resign, maybe a strong attack will open his eyes, or maybe it will inject some interest into the game.
ilan: It is obvious to me that the complaint comes from people who were losing by dozens of points and were still attacked by their opponent. In other words, it is a complaint by the loser against the winner, which is much worse etiquette than what they complain about. Accept your losses graciously, and if you believe that someone has no need to attack in a clearly won position, ask yourself instead why his opponent is not resigning.
Malweth: I certainly have to agree that attacking all out is not a problem — it is a style of play. It’s fine to defend territory when ahead — but if the player ahead does not need to defend the territory (because of a very great point advantage or already secure territories), attacking is the only option.
If it is a teaching game, it is certainly bad etiquette to overwhelm your student.
EdwardHammerbeck: Maybe it is a personal style thing. I am frequently, ahem, in the losing end of the B+100 scenario. Often I want to finish rather than resign in order to practice endgame, close up groups for a sense of aesthetic completion, or to just keep playing an enjoyable, if doomed, game. But when White starts going for my jugular vein, that’s when I resign and think a little less of the other person. I regard it as “kicking a guy when he's down”. In the interest of full disclosure, in the rare occasions where I am in the winning end of a lopsided game, I do practice what I preach. I calmly close my groups and end the game. I have no sadistic need to utterly humiliate and destroy my opponent. Bad karma.
ilan: That is exactly what I was talking about. I suggest you read the book “The Hustler”, by Walter Tevis. The subject is exactly how winning makes you the “bad guy” and losing the “good guy” and the various excuses people find for losing. Such psychological factors are intrinsic to pool playing, where self confidence is a key issue, but I think it also applies to Go.
EdwardHammerbeck: I see your point. One thing I failed to mention in my two cents above is that I normally, when I want to play despite being so far behind, will say to the other guy “hey, I realize I am over 100 points behind ... would you mind if we played on? I’d like to finish (for whatever reason).” In that way, I, rightly in my opinion, defer to his/her superior play and give him/her the choice. And then the second part of that is that I respect the choice.
George Caplan I suggest those who fight on read Nakayama sensei’s chapter on resigning in The Treasure Chest Enigma There is no point in continuing if a game is lost. Endgame practice? Play when you are 100 points behind is completely distorted — the only thing you will learn is how to play when the game is already over. As a player, who though weak, is often the strongest player in my club, my goal is to teach. My goal is not to play all night with someone who refuses to resign a 100 point game. While this person is busy not resigning, I could be going over the game with him, or playing the next eager student. It is not sadism that leads me to go further ahead, it is proper play, and irritation.
Alex: Allow me to add my two cents to the rising chorus of voices saying it’s fine, etiquette-wise, to attack viciously when ahead. I do think that playing safe is a better strategy for converting your lead into a win. Many is the game I’ve lost after getting ahead by 30, growing frustrated with an opponent who wouldn’t resign, aggressively trying to kill yet another group in the hopes that he would then resign instead of forcing me to play through a tedious endgame, overstretching myself and, in the end, accidentally letting him bring a dead group back to life or kill one of mine. Not only is it very frustrating to lose this way, it’s probably not a good strategy for teaching the person some manners — injecting further excitement into the game probably only encourages him. Next time I encounter such a person, I’ll try to make the game as boring as possible for him. Bore-you-to-death as an antidote for Annoy-you-to-death.
ilan: To me, this is the best answer. If you think that a certain behaviour is wrong, then the right thing to do is find a way to take advantage of it over the board.
George Caplan Of course it is better to play simply when ahead. I was reacting to the post that tried to defend playing on when you are 100 points behind — and thinking less of a player who still plays since you are still playing.
Here is what Nakayama says on the subject:
On occasion, when playing a teaching game with someone for the first time, I’ve had the opponent carefully make a two -point gote hane plus connection even though he was losing by 100 points. After the game the player rationalizes it by saying: “I thought it would be rude to resign when you’ve been so kind as to play me”. That implies that he plays differently with his friends, but if you watch you find him fighting a half-point ko when 100 points behind. He fills in all the dame, lays out the captured stones then exclaims “Well, what a surprise! I’m quite a bit behind,” He then pretends to resign. I don’t recognize this as a genuine resignation.
‘The most painful chore for a go teacher is playing with this kind of pupil. Another name for go is “shudan”, meaning “talking with the hands”. One may not utter a word, but each and every stone played relates the feelings of the players. If you translate “shudan” of a player who won’t resign a 100-point loss into plain English, he’s saying “You're such an imbecile that I can easily catch up 100 points” It’s hard to imagine greater rudeness in a go player’
sjd123: So do you suggest you should suddenly play more defensively if you are ahead?
davos: I am not very inclined to play soft when I am 100 points ahead against a weaker player. Some posters criticize the stronger player for continuing attacking when ahead by a wide margin, but usually the weaker player is also “continuing attacking” although it is usually just adding dead stones in gote, but in their mind it is an attack. What should I do then? Answer a move that adds a dead stone in gote? Pass? Passing often means that the weaker player still continues to add dead stones (at least in online playing). I have played lots of games where I was forced to play on until my opponent was out of legal moves.
kevinwm: Personally, I most often play for fun. If I’m ahead or behind, I try to make the best of the situation. If I’m way ahead or behind, I might try new things that I wouldn’t try in a close game. Clearly if a game is important, the proper strategy is to avoid risks to win — I'm not talking about that. I’m talking about when it doesn’t make a difference, play to have fun. And recognize that your opponent may be doing the same.
Spiritweaver: I understand both sides of this. I’m not good at estimating score, so I’ll keep on playing until I think there’s nothing else I can really do. On the other hand, I’ve been the butt of some pretty brutal attacks. I had a game where, in the endgame, I failed to make proper eyes in a group that was quite large. My opponent moved and killed the group. I figured there was nothing else I could do, so I passed. The opponent moved there again. I said, “It’s already dead,” but they kept placing more stones, while I kept passing. Then, after they captured that group, they went on to attack another group that I had mistaken for secure. They had the advantage, obviously, and after a couple moves I saw that the group was dead, and passed. They moved again, and I resigned.
--- My counting is often off by +- 30 so attacking while dozens of points ahead has won a game that would have otherwise been lost before. -JH
I think playing aggressive or defensive when points ahead is a strategic choice; if you’re sure you’re winning you can just have fun widening your margin or trying weird attacks. When I begun playing I regarded a 30 points margin as almost even; now I can tell if I’m winning by 10 points or so so I call it a victory even with small margin, but when playing even, with weaker players, I’m hardly satisfied with a 20 points victory. Actually I like playing until the end when ahead, I just try to win by more, I think it’s instructive for the weaker player too. Of course the kind of player that just waits for your mistakes playing aimlessly is irritating, but I think the bad part is that he believes go IS like that and when he loses he thinks it’s bad luck.
— Anyway, I believe the main point is that go is about “making more points than the opponent”; how you do it is another matter, but if you are behind you have to attack; if you are ahead by a few points you can slow down a bit (though it’s your choice). The person behind 100 points continuing playing (especially playing 2 point gote moves) destroys this balance because he doesn’t try to catch up, in other words his game makes no sense. When I play 50 points behind I try to come back (that is I play for a 50 points difference gain); if I can’t I resign but that’s personal. To illustrate what I’m saying, the famous Shusaku opening, so successful before the introduction of komi, became obsolete with komi, because it assumes Black is trying to win by one point or so (with komi black needs 6 points more). This to say a move may be good or bad if it gives you a higher chance of winning; if you are 100 points behind there may be no move that gives you a chance of winning.
This as a normal game; of course if you are a beginner and just want to get a feeling for go then you may play on to see what happens in the end (but keep in mind most opponents will get annoyed, so maybe its better if you ask first). The thing with internet games is you hardly find an opponent willing to teach and much stronger than you — that’s why I may want to play on when 30 points behind...
As for the teacher thing, I suppose many teachers win by a small margin even when they could safely get more, in order to leave the pupil some self confidence. I don’t like that; if the teacher tries to crush my game I try to crush his, no big deal, if the teacher goes for a small victory I have much more trouble finding the direction I should played in (I mean in mid-game counting points is not possible but better players have better estimation) but that’s also learning.
Polama Living requires enough territory inside your groups. When a player is up many dozens of points, it’s often the case that their groups are safe, and the opponent, having been forced into small shapes, is in peril of losing everything. If your groups are secure, what else is there to do but attack? Where’s the fun in a game where one player is left searching for small gote moves to reduce their margin? Similarly if you get ahead by eating a large group, you’re usually left with a living group radiating thickness in multiple directions. From that setup it’s hard not to pull further ahead. Simplifying is still a good strategy, but sometimes the best simplification is to read out the death of an opponent’s group and just remove it from the board.
Even in teaching games I’m reluctant to let too many late game mistakes go unacknowledged. I might let a mistake go once, but if the player is repeating bad moves, or failing to ensure eyespace in their territory, I think it’s better they get upset at my “running up the score” than letting bad moves form into habits.
Even when teaching, I think the only useful way to draw from a large advantage is to attack strongly to force the other player to keep their groups alive. Keeping games close in a teaching game is very valuable, but should be accomplished by playing moves at your pupil’s level through the entire game, not grabbing half the points on the board as fast as possible and then filling dame. If you’re dozens of points ahead you overplayed or your pupil underperformed and there should be enough for a useful review on the board already.
What I would agree is bad form when ahead is a speculative invasion, playing inside an opponent’s territory just to see if maybe you can get away with something. I’m not a big fan of that in general, but certainly if you've already won it seems unsporting. They left the 3-3 invasion open? Surrounded an area but without making provisions for 2 eyes? Sure, remind them they should have thought about that.
tapir: This might be a strange complaint from the loser of the game “who wanted to practice endgame”, but often players taking a significant lead try to “wrap it up” too hastily, expose themselves to a counter-attack and lose the game. Happens all the time. I guess this should count as a bad habit.
Pashley, about 10k: I think it depends very much on the situation. I once played an exploratory even game “to assess my level” as the first foreigner to wander into a Chinese club. I resigned after 20 moves or so before any fighting started; it was quite clear he had a substantial win just from the fuseki. In the second game, he gave me six stones and beat me easily but it was a much more interesting game. In general, though, I won’t resign until the end of the middle game unless I am hugely (~50 points) down. Against a stronger player, I will quit if I’m over 10 points down and all that is left is endgame play. I expect the better player to gain in the endgame, so there is little point in playing it out; the result is already clear. On the other hand, I won’t suggest to a weaker opponent that they should resign at that point unless I am up 30 or so. I almost never suggest it earlier.
C (SDK): I saw Michael Redmond 9p on a NHK cup commentary game state “if you have the advantage, you need to keep pressing your opponent in order to win the game”. I have also seen commentary about Alpha Go saying that the program’s approach to (end)games when it is ahead is to play the move with the greatest chance of victory — not to go for victory by the biggest margin. This is very much in line with Dieter’s B+1 = B+100.
That said, I don’t think Michael Redmond 9p would be throwing in speculative attacks in his opponent’s corner when there was a 4 point sente on the board unless he was way behind!
I think this mostly comes down to a cultural issue that we don’t have in the West, though, where it would be perceived as humiliating for a 1 dan to resign a game against a 12 kyu. Personally, if I was ‘trapped’ in a game where I had won by 100 points and I didn’t want to play any more, I would pass consecutively, then if the opponent didn’t take the hint, I would just resign and walk off into the sunset / to the bar.
If you’re in a tournament game, then suck it up and let the opponent have their moment in the sun — there’s nothing in the rules that says they must resign when behind, so you should be able to accept anything within the rules with good grace — even if that’s playing to win on time from 100 points behind.