The Global Problems Of The Tesuji Encyclopedia
261 views
October 25, 2026
Context
On a previous post, I shared the workbook for the Nihon Kiin's Tesuji Encyclopedia. However, still, most people are probably unaware that problem collection features two parts: a first for the expected, local tesuji problems; but also a second, for global problems.
At a first glance, those might seem like just some extra problems to contextualize what the reader has learned through the first, core section. That's not the case. The problems in that latter section grow on me every time I take a look at them, and I believe solving them is perhaps an even truer mark of expert skill than solving the more standard problems.
That's not to say they don't have issues though. Many of those problems are outdated with respect to AI, of course, and some even show wrong ideas for solutions. Nonetheless, going through those challenges and identifying imperfections within the problems are great training experiences.
The Tesuji Encyclopedia is Book #3 on 101weiqi.
The cover for the Chinese edition of the Nihon Kiin's Tesuji Encyclopedia.
Problems
Problem 2479
This first problem is a brilliant illustration of flow and interaction effects.
Problem 2479, White to play.
We have to link the forcing moves on both sides in order to have them collide:
Black needs to save one side, saving both isn't possible.
If Black tries to save it all...
Black fails in a ladder.
Problem 2478
Some problems in the book, though, are definitely quite... problematic. Not for their setup, but for their solution.
Problem 2478, White to play.
As a not-so-strong player, I cannot approve the proposed solution here, and neither can AI, which estimates the loss from move 1 on Dia. ? to be around 2.3 points.
The book's solution to problem 2478.
If there's a vital point on the top, it's 1 on Dia. ?, not 1 on Dia. ?.
White strikes at the vital point to stay minimally safe, before taking the crucial point of 11.
Black could try to resist, and might even kill White, but at what cost:
Black should have even bailed out from this fight with 3 at White 4. Crawling on the second line is so short on liberties it's still difficult to kill White.
According to AI, the top is still very hard to kill, so, surprisingly, it would choose to play 11 on Dia. ? first. If Black, tries to kill the top, only then does the book's tesuji shine:
The book's tesuji shines even brighter when we tenuki first. Of course, this demands outworldly skill on White's part.
Problem 2482
Black has a potentially troublesome capture on the right side. How can White use it?
Problem 2482, White to play.
White uses a forcing move on the right with 1 to then drive the A stones, on Dia. ?. Due to the shortage of liberties, disconnecting White's group in the B direction isn't possible. And Black 8 is necessary, given the C weakness.
White pierces through the right side.
Problem 2534
Here's another problem with great techniques from both sides, especially when considering the AI refutations.
Problem 2534, Black to play.
The book recommends trying to squeeze White from below, reinforcing the A group:
White is squeezed, while Black reinforces the A group.
AI thinks White shouldn't follow Black's plans. If you're gonna get squeezed at the beginning of the game, and all you have to show for it is second line territory, things did not go so well.
Also, on another topic, AI dislikes the joseki on the top-left quite a bit; among many other reasons, Black should at least turn at B instead of 19.
AI recommends resisting with an interesting technique which pops up in this joseki — and which appeared in my game against Uyama Hissao during my 2nd place run in the 2026 Brazilian Go Congress —, with White 19:
A joseki illustrating the key move to resist, White 19.
Translating that technique to problem 2534, we end up with this sequence, according to AI:
White's AI resistance, after Black 1.
However, that's not the optimal sequence for both sides. This is what AI recommends as one of the optimal solutions:
AI's optimal solution for problem 2534.
If Black insists on trying to kill, White has a tesuji up its sleeve, which hinges largely on counting the ko threats:
White creates a ko with 8 as a sacrifice.
Problem 2516
This one is more direct than most of the other global problems:
Problem 2516, Black to play.
Through a clever sacrifice, Black manages to find a way out, potentially killing White's A group:
As Black escapes, White's A group is going to suffer tremendously, if it doesn't outright just die.
Outro
What do you think about this sample of the global problems? Leave the ones you found best in the comments, perhaps in the future I might add them to this article.
Comments
You need to sign in to comment.
No comments yet.