My observation is that many times AI don’t bother that much on group security, ready to sacrifice/tenuki while we did like the proverb says (assuring a base for ex)
1 Like
By the way @hektik_jekyll a search with “proverbs” in this forum will lead you to more similar topics.
Happy reading!
1 Like
Kyu players aren’t AI. Most kyu players—certainly DDKs—will do much better, I think, by following the old proverb. When kyu players are nearly shodan, then maybe they will have developed the judgment to know when to sacrifice urgent for something bigger.
2 Likes
Well there is a saying that what is most difficult is to get rid of bad habits…
Motoki Noguchi, 7d EGF, has written a book recently, “l’usage de la force”. He says that AI didn’t change fundamental principles, and one of those is “urgent before big”. (Motoki does acknowledge that AI has changed our way of thinking. However what AI has changed is not the proverb, but our way of evaluating strength.) Many examples are given, for instance Yasui Senchi (white) vs. Honinbô Jôwa (1820):
Black’s next move is approved by AI. What is it?
3 Likes
It’s a nice way to say it – it’s true that some groups seem much stronger with analysis and can be tenukied from !
Although, on the other hand, how one can play also still depends on the skill of the player – there are situations AI can make life easily in that humans find hard, even pros, and may be 50/50 or worse to make life in (easy to die or make mistakes) even for higher level players.
And apart from that, many key concepts about key points & shapes (especially sente/anti-sente moves) still seem to apply the same ways in AI analysis.
(the AI seem to also like good shapes & key sente/anti-sente points in them, in many ways – maybe even valuing them more than human amateur players in many situations, in my experience)
Tesuji (spoiler) : f3 looks like a vital move for affecting the life and death (base and eyeshape) for both groups.
And white playing there is also big territorially, in sente, along with un-settling the black group and making the white shape very alive
2 Likes
F3? (a move that affects the bases of 2 groups is urgent)
I think AI are just very flexible and their positional judgement is so good that they meet the conditions to sacrifice/trade skillfully. AI can accurately tell how much compensation they need for a fair trade and if that compensation is available. So they can walk very thin lines with confidence, much more than humans (even pros) can.
Still, AI are superhuman, but not supernatural. AI are very good at keeping stones light and avoiding overcommitting to anything for longer than humans can, but eventually groups can become too big to fail and they will settle it, preferrably leaving as few forcing moves and ko threats as possible. I think AI hate to have a heavy weak group bullied and chased around the board.
Edit: another proverb, one that Rob van Zeijst 7d EGF likes to quote: “a weak group is worth -30 points”.
1 Like
I think this is the case for AI and many shapes, somehow – when reviewing my games they play and emphasise counter-sente key points in shapes, where squeezes etc. can happen to squish/force against the shape, a bit over taking big points, earlier than I notice/am sure of taking a point there, at OGS 5-6d
And a pro good in fighting who could consistently win very strong players with supposedly “bad shapes” (common with fighting styles & in my experience as a fighting player), said that they began to learn good shape properly & improve it with the advent of AI.
My experience is that it points out key urgent shape points too, or ways to prevent future situations where forcing moves can be played against a shape in general, causing problems later in the game or fighting, so in fact many things which may be “honte”, and urgent from that perspective.
(often the moves/analysis I notice from my games of players around the same level (+/- a bit), involve both players missing playing such points (“slow”-looking but strong shape moves) as early as the AI suggests – and they do look very good to me to avoid leaving the potential for future complications/forcing moves in that situation)
1 Like
Yes. I think AI are very good at spotting and getting the timing right to make the transition from a light shape to a thick shape, while avoiding heavyness and overconcentration. That can be really difficult for a human, because the distinction can be quite subtle. Though sometimes it’s not actually that difficult to spot, but you just lack the confidence to play it because you worry that it is too slow.
1 Like
Yes I’m often surprised when looking at AI analysis that I missed a reverse-sente move during early midgame. So for now I’m using the homemade proverb: “reverse-sente moves are bigger than they appear”.
3 Likes
I prefer the version
“Trade plums for peaches”
Mainly because it sounds risky to put your plums on the line when you’re not sure you’ll get the 
1 Like
Counter-point: aren’t we all playing the same game of Go? There is no “Go for idiots” separate from “Go for adults” (paraphrasing).
If a move is good, why not play it? If we do not understand a good move, why not seek to explain it? If we cannot explain it, why not believe it on the stronger authority until we can? Let’s use the tools that we have instead of rushing to put on a straitjacket.
Anecdote time. 
I was watching a dan-level Go streamer instead of being productive (as usual). Their opponent played the shape: 3-4, low approach, kosumi, kosumi. Without diagram; use your imagination. Upon which this streamer, who had apparently never seen this, commented that it cannot be good and then went on a whole tangent about people playing AI moves without understanding them and how bad it must be for them not to have the AI insight to find the follow-up and viewers are advised not to do this.
This irks me because this shape has been popular for centuries; you can see it in games of Honinbo Dosaku, for example. Even if it was an “AI move”, what is really so bad about that? Nothing that dan players can articulate. When I hear some monologue where every third word is “AI”, I tune out.
I tell my students that the opening is very free, there are few bad moves and you can play what you like.
4 Likes
I don’t think I understand your point, so let me try to simplify mine. The choice between urgent and big requires good judgment/experience. Most kyus—especially the weaker ones—lack the requisite judgment. Therefore, following the proverb is the better course of action for those players.
It sounds like the go streamer was uninformed, so he/she isn’t a good example to illustrate a principle.
I agree about AI, but I’m not the one that dragged AI into this.
2 Likes
Yes, that advice is something that we learned from AI. In the early opening, there aren’t many urgent moves. Though perhaps we knew that already (proverb: there are no ko-threats in the opening)
But I’d consider the proverb about urgent before big to apply more generally than the early opening.
5 Likes
Thanks for the heads up @Groin
Thank you for clarifying. I agree about the proverbs.
All I wanted to say is this: let’s not discourage learning players from any moves that are basically fine, especially for such reasons as: you are not strong enough to play this, it is too complicated for you, and of course because AI. 
4 Likes
Yes i get this. When i watch AI games, the limit between flexibility and weakness (with the abundant use of tenuki) is what made me wonder how the proverb still hold but yes. I have to reevaluate where and when is the switch. 
2 Likes
It’s not a proverb as far as I know, but I think it could be one:
It’s more valuable to read 4 moves deep correctly in 5 different directions, than reading 15 or 20 moves deep in only one direction. (At least at DDK and SDK level)
1 Like