I'm really having a hard time understanding probes properly.
Going through the Invincible book (games of Shusaku), I constantly read that people are/were very impressed with his perfect timing playing probes, even at such a young age.

Many times, the 'probe' stone(s) end up as dead stones. I know what a probe is supposed to be, but I feel like I don't understand it properly. Why is there a perfect timing for a probe? If the probe is played a few moves earlier or later, what is the big difference?

Could someone direct me to a good site/book/... where I can learn more about probes, or maybe someone can give a good explanation here? The page on sensei makes some sense, but is short.

So while Shusaku was a genius, playing these perfectly timed probes at age 14, I am still trying to understand what a probe actually is at 25

Curious to know if the bots agree with the human assessments for these cases.

Part of the value of most probes is that you can throw them away. If you look at today's top bots, a lot of their early moves act as probes. That is, having made a few moves in a area and gotten the opponent's responses, they often leave a few stones where they can be sacrificed later, if need be, and play elsewhere. OC, they don't always do this, but they always play the whole board.

Some of the skills involved with probes are 1) learning to throw stones away, 2) learning how to throw stones away, and 3) playing the whole board. IMO, bot vs. bot games show how.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

I'd love to input a game into Leela.11 and see, but these games are without komi so I'm not sure the robot will asses the position correctly.

A key idea of probes is you are asking your opponent how to respond to your move, where there are several reasonable choices but it's not clear yet (so timing is important) which one is best. Depending on his answer, you then play differently to try to turn his answer into the wrong answer (typically via inefficiency). It's not on probes per-se, but the book Beyond Forcing Moves has lots of nice worked examples of skillful timing and technique of move order. Here's a post I made on it: forum/viewtopic.php?f=17&t=5150&p=86025&hilit=beyond+forcing+moves#p86025. Also see forum/viewtopic.php?f=12&t=7185.

I think that it would be instructive to try and see. Especially if you play through Leela's main line to see the eventual fate of the probes. I wouldn't worry too much about whether Leela.11 can assess the position correctly. After all, Shusaku was better then than Leela.11 is now.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

Okay.
In Shusaku's game (the last one I replayed) his probe stones died, I'll try to go back and look what he gained by them. Maybe I'll input them into Leela, too.

I have recently ordered (and it arrived half an hour ago) Invisible, the book with AlphaGo's games against Lee Sedol, the 60 Master games and the games it played against Fan Hui.
After going through Invincible, I plan on going through Invisible. I'll watch out for the early probe moves then!

The only book I know about timing (and associated probes) is a book written in German by a national champion, J. Mattern from my feable memory. Quite difficult high concept, the right timing means if you do it too early it may be some bigger moves, too late no more the time to get it answered and usually involves some yosumiru (fix a choice) aspect in my short view.

You know, this is a good question. I'd be curious if there's any detailed (perhaps mini-book-length, e.g. lecture-turned-into-book) analyzing probes and exploring some of the implications.

I went through some of the games of Jowa recently and it's very startling how much of the early-mid game in classical go involved probes all over the place. Cynically you could say it's a product of overthinking each and every little move in an era without time limits, but it undoubtedly changed the character of the middle game in those games. There's a lot of aji left in corner and side fights and a lot of resurrection or surprising kills (or at least, surprising kos). Which strikes me as a good indication that there's a lot of skill and finesse but I do wonder how deep the reading really was vs. it being stylistically trendy to probe all over (and top players doing so--who would have created and exploited aji regardless of style).

I am delighted by pro moves and commentaries on these,but I am 1000 km of them, still struggling to get right direction and efficient shapes, basement of any superior concepts like probes and timing...

Strong AI (AlphaGo Master and AlphaGo Zero even more so) are exceptionally good at probing (sometimes I even feel that every AI opening move is a probe). Check out Micheal Redmond's reviews of the AlphaGo games and you'll see some exquisite probes that have mind boggling long term effectiveness. This supports the notion that probes are a good indicator of skill.

In some ways even the early 3-3 invasion can be thought of as a probe: you are asking the opponent to pick which side to block now when it's not clear which is best (though if there is a 'wrong' one it's rarely by much) and then later you can choose your direction of play to try to make it the wrong one.

As has been noted above, probes were a major feature of old go. Perhaps longer time limits encouraged this style of go. Perhaps nowadays players have become more used to blunt force and perhaps AI has served as a brake on the modern style of play.

But however you look at it is all a process, and even in Edo times probes were still being, er, probed.



Everyone takes this probe for granted now but it was not tried until 1857 and it had to be championed by Honinbo Shuho for quite a few years after that before it became routine. It can be seen as the Edo equivalent of the early 3-3 invasion, with other pros not quite sure what was going on. It was Honinbo Shuei who was the first other player to get the hang of it, and he of course was effectively a kid brother to Shuho. He probably got it all from the horse's mouth but other players took an awful long time to see the light.

Even now, we are still in the middle of the learning process.

But I would suggest that learning about probes can be hampered by just calling them probes. In exactly the same way that westerners have painted themselves into a corner by using terms such as thickness, aji and forcing moves in a blunt way, overlooking all the nuances of the Japanese terms (e.g. aji has aya, fukumi, etc, thickness has atsumi, atsusa, seiryoku, etc), I see here the same blunt tool being used to probe probes.

If you look at Japanese commentaries on old games what you notice about probes is first, as already observed. that they are very frequent. But, second, they are very varied and they are often discussed as part of a nexus which includes aji and forcing moves and - perhaps surprisingly reminiscent of the new AI style - erasure. In other words, there is often discussion as to whether
a move is a reduction or a probe (or both). Perhaps this strong element of erasure and focus on the centre (very AI like, be it noted) is what prevented them from seeing earlier the kind of second-line probe illustrated above. Even in the 1950s and 1960s pros were still being shocked by Sakata's jugular-slicing use of the second line (which is why he got the nickname "Razor").

Ian, I have struggled just as much with this fluffy concept, which has been conveyed to me as "a move that invites the opponent to make a choice, so that next you can take that choice to your advantage in subsequent moves".

In a way, this should apply to any move.

But I do see the difference with a forcing move, which is forcing the opponent to confirm a prior choice (usually defending some territory claimed) and reducing his options (like expanding the territory), while your stone still has an added value (usually some sort of influence). Then again, the opponent can always try and resist the forcing move, which is then - after the facts - converted into a probe, since the opponent apparently thought he still had a choice.

"Probe" is a term applicable in situations where choices still need to be made. "Forcing moves" are applicable where choices have already been made.

Both reduce uncertainty, which is trivial, because every move technically reduces uncertainty by bringing more information, except perhaps ko.

Bots are essentially "reducing uncertainty" all the time. That's how they operate, by measuring the likelihood of a move to lead to a victory and choosing the move with maximum likelihood. But it's a whole different way of reducing uncertainty. The human concepts of probe and forcing move yield more information about the nature of a local situation and its implication on the global position, not on the likelihood of winning.

I don't think bots duplicate any kind of human concept and I'd be very interested to know what matter the AI conists of.

Specific to the question of timing, I found this video very helpful: https://www.youtube.com/watch?v=TjhGaSHXPz0&t=2134s