With Black to move, the top left corner is a sekifiable shape.
While most if not all of the introductory go books pay attention to killable shapes, I never encountered a book that payed significant attention to sekifiable shapes.
I find this lack of knowledge and theory about seki’s rather surprising. Okay, a seki is certainly not as big as a dead group, it nevertheless is a nice reduction of the opponent’s territory. And there are many games that are decided by a seki.
Having said this, I have two questions:
are there books or articles that deal with seki’s?
do you know other sekifiable shapes?
Thanks for reading this and hopefully sharing your knowledge with us.
It’s been fascinating to encounter some very interesting seki, or seki/ko shapes in the corner.
Here are a few…
A classic when there’s an open exterior liberty in this shape (without the exterior liberty, it’s a thousand-year ko, but there is a cool seki there) :
I’ll post more i I can re-find more of them, or find more.
There’s also very interesting nuance as to whether a seki is sente or gote, and thus when or whether they are worth playing at a particular point in the game. ^^
I think common shapes show up in Tsumego books, certain one in corners or on the side.
But in more general life and death books, it’s more the case that some heuristic is given like the most common killable shapes with nakade fit in a 3x3 square, and when the eyespace is 7 points or bigger, it ends up being a seki or alive with points.
There’s lots of exceptions though, where weaknesses make the eye space smaller than it looks, and I feel like you’d want to be a very advanced player to confidently survey all of the possible seki shapes and special killable shapes.
As in
if you were going to write a whole book or dedicated chapters to all kinds of seki.
I can flick through some of the books I have later and see if there’s any ones that stand out in this regard.
The classical ‘Life and Death’ from James Davies covers a few basic cases. Not as a separate category/chapter though, just while covering corner patterns etc.
Haha, that is definitely a nice idea, but first let’s finish me the Dutch translation of @JethOrensin 's Multi Lingual Go Project and second since I am a SDK I probably would never be able to finish that, so best would be to undertake such an enterprise in some form of group writing.
I think that sekifiable shapes would simply be a sub-category of unsettled shapes. I guess that they are not common enough that an entire book would focus on them, and for the sake of study, one would also consider these alongside killable shapes, in the context of unsettled shapes and corner patterns as a whole.
Well, I have only written half a page for seki, so yeah, quite embarrasing on my end.
There is a reason though. I don’t think I’ve ever managed to actively make a seki and in the rare couple of times that I might have had, I blundered it. I had to do a lot of digging in my games to even get something close to it and, if memory serves (surprisingly, it has been 7+ years since I wrote chapter 1. Time flies!), the only image/example I could find was from a real game where a group died and I heavily modified that to make an artificial seki (it definitely didn’t happen in that game).
It took me hours to even locate that example and all those years later I still do not think that I can point out to any of my games and say “There! I went for a seki and I made it!” My best guess is that I could take days to re-view all my games and find like, five seki or seki candidates? I’d be surprised if they are more than five.
On the other hand this might be because I am so totally bad and ignorant about seki.
Just like pros in chess can seemingly tie more than win or lose, and I cannot even think about how to aim for a tie in a game, this utter lack of sekis in my game could be totally self-inflicted.
Have you never made a seki in the endgame? Like in the late endgame when liberties get reduced so you may get an opportunity to steal some points from a living corner group by turning it into seki?
Not really.
In the games I am in, the endgame/corner mistakes are either catastrophic and something dies or the players involved cannot properly kill whatever group is in danger and it survives.
As far as I can tell, if you have the skill as an SDK/DDK to force a seki, you probably have killed whatever group it is that you are targeting, way before the seki option came into play. And if you don’t have such skills, then it will probably cleanly survive whatever attack you can come up with.
I could be wrong, but that’s just what I’ve seen happening in my games.
In my mind, the puzzle I showed is not about attacking and killing, but more like a little endgame reading puzzle. And the shape is fairly common in real games I think (though perhaps more so in handicap games). @Atorrante’s OP puzzle is also a pretty common corner endgame shape I think.
You are OGS 2k right? In my mind that’s a significant step up from SDK/DDK level. I didn’t really expect that even at your level, catastrophes happening left and right throughout the game is still the normal game experience.
I mean, it can definitely happen in my evenly matched games that a group unexpectedly dies, but it’s not normal I think (outside of blitz games). And we are not that far apart.
I’ve probably had a lot of games where groups unexpectedly live or die.
The reason why we’re not dan players is inconsistency in some cases
You make a misread here or there and your corner dies or some other group now has a ko for no reason etc.
I think in a lot of players play, there can still be some missing fundamentals for example, because we don’t have to have passed a test to say our endgame has to be at least X level and Tsumego needs to be at least Y level.
Probably pushing beyond 2d, 3d and so on that might be the case though.
Anyway in some of these cases the sekis only show up when the dame are all filled like
and it’s just a mistake of one player not to realise the aji.
Other players when behind might have resigned rather than fill in the dame, at least if they’re not actively spotting the aji. So that can lead to maybe an under representation of certain types of seki maybe.
I think this corner shape you’ve posted and other sekis in endgame or midgame happened to me a fair bit at around KGS 2-4k or even around 5-6k more than a decade ago when the rating system was maybe similar to current OGS.
Especially when trying to kill or live groups, capturing races… – it’s still fairly common for me now(or more so now ?) to make all kinds of seki, especially in long thinking time games, like correspondence, and can happen when fighting, or making life in the centre or sabaki too.
Maybe it depends also on how reading- and fighting-heavy the style is-- a couple of French players I played with often who are around 2k-1d OGS have this in our games and fights a lot too, and they both have fairly good time use+spend time reading the intricacies available in a shape (one also really likes to fight ). And my style was very fighting-heavy too even at 5-6k KGS, a lot of capturing races and life and death+reading aji, which can easily turn into something where seki is possible in yose or midgame.
Technically, this belongs to endgame technique, not life and death or tsumego. Making a seki means taking around 7 points of territory away from the opponent, similar to doing a monkey jump. Although I agree that learning how to do this (or defend against it) feels more like learning how to read out life and death problems, so it would make sense if tsumego books had a “seki” chapter.
I’ve not seen it happen more than once, I think. On my end I usually cover the flaw earlier and most of my opponents do too. Sure you lose some time and sente, but who said that we are efficient?
They are not often, but when they do happen, you notice (because someone resigns, usually me). In my case it just happens that I rarely play against people of my rank. They are either ranked and I challenge two-three ranks above me or they are friendly “teaching” games so they are three or more ranks below me, so maybe I have a bit of a skewed perpective on the issue.
Good find! As you can see, noone noticed that this is a potential seki. I also just noticed that we even forgot to mark the dead stones on the snapback (thankfully they do not change the outcome)
Do note that unlike chess where the board has a definitive “north-south” axis, pattern recognition in Go is much harder because all directions are equal.
Same image, same shape, different orientation. Good luck memory cells!
Also, thank you for providing the link, this way I could play it out and see how it works. I could see the danger of the shape, but I was not exactly certain what the actual moves would be. This seems to be the solution (ignore moves 1 and 2, they are there to take the snapback out of my sight):
In which case I would have lost the game. Good to know
My best guess is that I could take days to re-view all my games and find like, five seki or seki candidates? I’d be surprised if they are more than five.
). And my style was very fighting-heavy too even at 5-6k KGS, a lot of capturing races and life and death+reading aji, which can easily turn into something where seki is possible in yose or midgame. 
As you can see, noone noticed that this is a potential seki. I also just noticed that we even forgot to mark the dead stones on the snapback (thankfully they do not change the outcome)
