Improve your SDK fighting abilities 🐽

So this is a focused training at a shorter rythm as i used to propose till now. One by one, one problem for a quick fun (My hope ofc)

From the feedback of other threads, it seems there is some interest around tesuji, middle game and a SDK public.

As always discussion and feedback are welcome. Post your screenshots!
Stronger player may come in but then please use the spoiler tool. ​

01- The broken snake.

White to play.



Considering the answer,.how black should have played?

demo for your screenshots


I prefer these kinds of problems because even if you can’t read it out entirely, the first move is just so good for you that no matter what happens the end result is great.

If I was white...

I wouldn’t allow black to atari my stone at 5-7.
So I 'd play at 4-6 to make a tiger’s mouth which should also cut black someway.

Tiger’s mouth with a wedge?

Coordinates please next time. :grin:

A screenshot corresponding to the previous post, with coordinates.



Good! So what about the cookie?

02 Take the most

Black to play


demo for screenshots

1 Like

If i understood well the cookie is not well cooked.


For the cookie of problem 1 I would have played like this.



cookie spoiler


Problem 2

I would try and cut white with R7.
I don’t know if it would work.
Or maybe S6 as an alternative. I think S6 could connect white, so I’d remove it.

But I’m quite scared by something possibly happening around P2. Are those black stones in the corner safe?

Probably in a real game I’d play P2 as black, letting white connect at S6


Well you can agree that if white is dead already if surrounded, better not let him connect. Quite a big fish here. So checking P2 is a good idea but nothing happen there, just a distraction as black can give up T2, that will not help white to live if he gets surrounded
Now how you suggest to close the escape makes me think that this topic is very well fitted to you. Try a bit more reading as i am sure you can.

1 Like

Yes, I think so.
The only issue is my poor reading skills. :grin:

1 Like

Ok you already have interesting ideas of possible moves. Now add a bit of checking mostly on liberties. Building a complete image with all the stones around where you played at step1 2 3… helps to check these libs.
These “not so deep but still a bit tricky” problems should be like an invitation to sharp your teeth because, you guess, the reward is there and great.

More: the goal is to reach the state where you are convinced you have the solution. So you can check if another move is possible which is a good way to get convinced or to find a failure.
Problems are made in a way that there is only one starting answer.

Follow-up question to first cookie:
What is the problem with this black move?


chocolate cookie!

The wedge seems working pretty well

wedge with variations

1 Like

What if Black responds to the wedge with R6?


Yes i forgot that branch. And i have difficulty to find any other move satisfying for white. I guess then white can at least enclose black if It’s in his interest (still costly, letting b live in big ) So as a variation of the problem, unlike the hane, the jump is relatively connected with the forcing move (still a cutting aji for an attack on the center stones later)
But maybe @le_4TC has something else in mind?

hint (spoiling a bit)

After R7 my guess is you expect a classic hane+connection but here, well, check if there is not an alternative for white

1 Like

I didn’t consider that! I thought the wedge sequence from @Groin was cleanly working.

So the jump does not get easily cut, but the R6-S6-O7 sequence seems like a cleaner way to get out (and when black plays R6 immediately, it is a bit more sente than after the wedge, I think).

However, I tried to ask KataGo about it (setting up some arbitrary stones in empty corners) and as usual it seems like us poor humans have no idea what we’re talking about…


It doesn’t mind the immediate hane, because if white goes for the cutting sequence (which is kind of bad shape for white) there is aji with S4 that let’s black live in the corner. And if R6-S6 is not strictly necessary, it’s a sligthly bad exchange.

The tesuji for white to cut is still very useful to understand of course - sometimes it will be good in a real game, sometimes not.


That’s a nice toping by the AI. Thanks for this check.