Why are our instincts so bad?

[Bill Spight]

If repeating 'C' from board position 'P' is not great since 'C' is still fresh in the mind, what about a small modification?

1. Take an ordered set of board positions (P0, P1, ... Pn) where you played a bad move (B0, B1, ... Bn), and know the correct move (C0, C1, ... Cn). That is to say, on position Pi, you played bad move Bi, but the correct move was Ci for all 0 <= i <= n.
2. Select some index j at random, and practice playing Cj on position Pj.
3. Repeat step 2 several times.

This way, you still get practice strengthening the connection between the position and the correct move, but since you keep iterating to a fresh position for each iteration, you give a chance for your bad moves to show up again.

Would that work better?

I think so. The newly presented problems interfere with the short term memory of the original problem, effectively erasing the working memory. So that when the original problem is presented again, it is without that baggage.

So far, my experience is that mixing up the problems is nice, but I still get the feeling that I am recalling board position as soon as I see the problem.

Maybe I just need more problems to fix this.

Yes, there are problems with simply remembering the answer to a problem instead of working it out afresh. When I was learning go, I tried to avoid that by waiting at least one month before reviewing a problem. I now think that that was not such a good idea. I did not have many problems available, and so by not reviewing problems and not trying problem above my level I ended up not doing enough problems. I think I would have done better to forge ahead anyway. Not having enough problems at the appropriate level is not the case for Western go players these days.

One problem with simply remembering answers is the same one of the stimulus situation. In order to recognize a problem position, only relatively few features of the problem are necessary. Thus, the desired connection between the relevant features of the problem and the answer is not reinforced if all of those features are not noticed.

Two things can help. One is to take the time to look at the problem being reviewed and take in the relevant features. Focusing on those features is even better, IMO. Another, which I like, is to make similar problems by changing a feature so that the correct answer is different. A good example is the 2x3 eye in the corner, where the right answer depends upon the number of liberties of the eye. To make such a tsumego problem all you have to do is to move a stone one space or remove a stone or add a stone. Whole board problems may require more alteration.

[Bill Spight]

But rapid repetition of the right play still alters the stimulus situation.

Hi Bill,
( Off-topic? )

Interesting.

One wonders if there are similarities to this scenario:

P: Meeting someone. ( Interactions with people, in general ).
C: Greetings ( good feeling ).
B: Some nasty, or at least unpleasant remark. ( Bad feeling ).

Suppose a person has a tendency to express or otherwise project some unpleasant feelings toward others.
Now, rapid repetition of C -- say, "How are you" or at least holding the tongue long enough to think before speaking --
could alter the behavior of this person. But maybe that's only the symptoms, and not the root of the problem.
Which could be some bad experience in childhood, insecurity, anger issues, etc.
Unless and until we dig deep enough to uncover the buried issues, later on, especially under stress, B will likely re-surface.

But maybe that's only the symptoms, and not the root of the problem.
Which could be some bad experience in childhood, insecurity, anger issues, etc.
Unless and until we dig deep enough to uncover the buried issues, later on, especially under stress, B will likely re-surface.

This may be controversial, but I think that we have good evidence that bringing old traumas to light is not sufficient to release their effects. In fact, doing so can retraumatize the person. And, as evidenced by so-called recovered memories, a person can be traumatized by imagined events.

That is not to say that realizations such as, "My mother used to do that to me, but you are not my mother and I am no longer a child," are not helpful. But notice that that involves inhibition of the childlike response. It is possible to inhibit that response without conscious memory of the past.

[John Fairbairn]

Kirby wrote:

If repeating 'C' from board position 'P' is not great since 'C' is still fresh in the mind, what about a small modification?

1. Take an ordered set of board positions (P0, P1, ... Pn) where you played a bad move (B0, B1, ... Bn), and know the correct move (C0, C1, ... Cn). That is to say, on position Pi, you played bad move Bi, but the correct move was Ci for all 0 <= i <= n.
2. Select some index j at random, and practice playing Cj on position Pj.
3. Repeat step 2 several times.

This way, you still get practice strengthening the connection between the position and the correct move, but since you keep iterating to a fresh position for each iteration, you give a chance for your bad moves to show up again.

Would that work better?

Bill wrote:
I think so. The newly presented problems interfere with the short term memory of the original problem, effectively erasing the working memory. So that when the original problem is presented again, it is without that baggage.

If I understand this correctly, it appears essentially to be the method used by Meijin Inseki in Today We Have a Splendid Feast. At any rate, he is essentially getting you to do the same problem many times, each time presented slightly differently. If it worked for the guy who wrote Igo Hatsuyoron....

[EdLee]

Hi Bill,

In fact, doing so can retraumatize the person. And, as evidenced by so-called recovered memories, a person can be traumatized by imagined events.

There is a drama teacher -- he used to be one of Julianne Moore's teachers -- he really dislikes the idea of re-visiting personal experiences to prepare for a role.
Especially unpleasant ones. One reason he said is why would you want to go there again. (re-traumatize)

That's one end of the spectrum. As we move to more mundane situations -- say, just bad parenting, or copying parents with bad behaviors -- there are likely more ways to find out the root causes, without re-traumatizations.

[Darrell]

By coincidence, I am reading a book that touches on this subject. The author talks about how psychologist Daniel Kahneman separates decision making into two systems - a fast, intuitive one and a slow, analytical one. While the analytical system may be the one used to choose the final decision, the initial candidates are generated by the intuitive system.

The part I thought was interesting was that in order to become an expert at a skill, one must use deliberate practice to train the fast, intuitive system. I can see this watching Haylee's videos; she casually discusses her upcoming streaming schedule while playing her openings against 9Ds and 1Ps. After playing a move, she often immediately points out her opponent's next move with great accuracy. While Robert may thoroughly analyze every potential move and counter-move, Haylee clearly doesn't.

I think a high-functioning intuitive system also explains differences in reading ability. Using brute force, I have to consider 5 potential moves, then 5 potential counter-moves to each of those moves and I still often miss the best play. With good intuition, a player can a) identify the best candidates and b) prune the tree to just the best candidates and thus have the capacity to read further.

[RobertJasiek]

the initial candidates are generated by the intuitive system. [...] in order to become an expert at a skill, one must use deliberate practice to train the fast, intuitive system. [...] Robert may thoroughly analyze every potential move and counter-move

I do not generate my initial candidates by "intuition" (which, IMO, does not exist anyway) and have not become an expert in that manner. What has made me an "expert" (i.e., helped me to move from 4d to 5d and then to stronger 5d) has been the consideration of all moves and then applying choice filters to them, such as identifying the one relevant part of the board (if there is exactly one) and then using Local Move Selection to identify the best move in that part. Application of the first few filters is very fast. So fast that subconscious thinking is not needed. E.g., one of the filters is "Identify the intersections that the opponent's last move affects tactically / strategically / initially unclearly and significantly.".

I think a high-functioning intuitive system also explains differences in reading ability.

For me, not at all.

With good intuition, a player can a) identify the best candidates and b) prune the tree to just the best candidates and thus have the capacity to read further.

With application of esp. generally applicable reading theory, candidates are chosen, the tree is pruned and reading becomes manageable. Subconscious thinking has never had any noticable effect on my reading. Reading due to specialised theory (such as status knowledge of particular shapes) without also using any / enough generally applicable reading theory has been successful for me only in those cases where generally applicable reading theory can explain the seemingly accidental success (such as verifying the status of the particular shape).

Replacing reading theory by "intuition" would immediately make me weaker by 24 ranks if only I were able to forget and not apply the theory. (I can't.) Relying on good "intuition" instead of good reading theory is the worst possible advice. What improved my reading at any rank has always been more, clearer or better knowledge acquisition about reading theory.

[Bantari]

Replacing reading theory by "intuition" would immediately make me weaker by 24 ranks if only I were able to forget and not apply the theory.

Since you don't believe in intuition, you would be replacing something with nothing - and so your huge drop in rank. If you actually believed in intuition, and made the rational realization that nobody asks you to *replace* theory with intuition, but only to *augment* it - you might actually get stronger. In any case, you would not get any weaker - in the worst case, it will have no effect since adding nothing to something gives back the same something you started up with.

Personally, my intuition at present adds around 5.74 stones to my playing strength, although I might be an extreme case.

I wonder what others think about that, and what their numbers are.

[Tami]

Replacing reading theory by "intuition" would immediately make me weaker by 24 ranks if only I were able to forget and not apply the theory.

Since you don't believe in intuition, you would be replacing something with nothing - and so your huge drop in rank. If you actually believed in intuition, and made the rational realization that nobody asks you to *replace* theory with intuition, but only to *augment* it - you might actually get stronger. In any case, you would not get any weaker - in the worst case, it will have no effect since adding nothing to something gives back the same something you started up with.

Personally, my intuition at present adds around 5.74 stones to my playing strength, although I might be an extreme case.

I wonder what others think about that, and what their numbers are.

There can only be a few possibilities:

1) Robert really does somehow experience an explicit and fully conscious simultaneous awareness of everything he knows while making a move. That is to say that every principle and every example he has learned actually flashes through his consciousness on each and every turn.

2) That there is disagreement over what words such as "intuition" and "instinct" mean.

With respect to No. 2, I think there are two kinds of "intuition" or "instinct", etc.

The first kind

What comes to mind without the benefit of "learning". The method by which animals know what to do in the wild. The way a complete beginner at anything might go about performing that task.

The second kind

"System One" thinking, which is informed by learning. The way in which most people would automatically spot a snapback or choose not to touch a hot iron; the system by which one knows how to drive home without giving it any thought. The thing that causes certain moves to come to mind in certain kinds of positions.

I am quite prepared to entertain that No. 1 above is true for Robert. I would find it extraordinary, to be sure, but if that's how he thinks, then that's how he thinks. However, I hope that he would be equally prepared to entertain the possibility that not everyone's mind works like that.

As for improving at go, I think for many of us our instincts, whether of the first or the second kind, are as the OP said, are not good. We don't study enough, we don't review our games sincerely, we don't receive sufficient guidance, whatever, and we end up with a faulty knowledge and understanding of the game, that causes sub-optimal moves to present themselves in our minds as we play. The answer is, in my opinion, to go "back to school" and question what we know and improve our knowledge and comprehension of go theory; and indeed to rely less on our instinctive responses (whether native or learned) as we play, but as much as time allows on careful thought. That way requires effort; but isn't that the real magic pill for progress?

[John Fairbairn]

I am quite prepared to entertain that No. 1 above is true for Robert. I would find it extraordinary, to be sure, but if that's how he thinks, then that's how he thinks. However, I hope that he would be equally prepared to entertain the possibility that not everyone's mind works like that.

I likewise would find it extraordinary, especially as RJ is known as a rather good blitz player. But I'm not sure that either of the above possibilities apply here. I don't thinks he explains it very well, but I infer the following:

He believes in the effects of subconscious thinking, and relies on it, but he also believes that in his case his subconscious thoughts have all been trained logically and consistently by himself, and so should not be classed as intuition.

So, if he learns elements A, B, C, D and E in that order (he's a list man) he expects that, with due training, these elements can be called on instantly and in that order to reliably affect his judgement of a move.

But this is where the rest of us part company with him. We (?all) accept that when we learn A, B, C, D, and E, at the same time we may also learn bit of F, G, H and so on. These extra elements may apparently have little to do with A to E, but there may be accidental connections, and our subconscious brain ALWAYS logs these connections. As time goes on, F, G. H experiences may be repeated, and new experiences I, J, K... may (unintentionally even) be added to the mix, and through constant feedback some effect of all of these may impinge on A to E. In addition, it may turn out that we don't use A to E as often as we initially expected, or we use some of those elements more than others, so that in time some of them come to dominate and so affect the planned ordering.

All of these extra unplanned additions to A to E have unexpected effects and these are what I think most of us class as intuition or feeling. RJ seems to have a horror of these unplanned effects, but it seems that they can often, maybe much more often than not, be beneficial. They can, for example, provide tweaks that enable us not to make the mistake of using raw A to E in a non-standard position - the linkages with the extra elements F to K, which we may not be aware of consciously, can alert us to danger or make us more effective.

Furthermore, I glean from RJ's actual words that he considers someone (like himself) who has learnt A to E thoroughly and in an orderly fashion to be an "expert." Again I think this is somewhere the rest of us part company again. We would see someone like that as merely well trained (e.g. a doctor who has just graduated), and someone who has added the elements F to K, unknowingly but through rich experience, as more like a real expert (e.g. a doctor who has worked in a hospital for ten years).

If I'm wrong about this interpretation of your remarks, Robert, I confess to being baffled. And I have to say that if you intend to continue producing books for other people, as I hope you do, you should seriously consider making sure you and your intended audience are on the same wavelength. One starting point is that English speakers know what intuition is (even if intuitively) and believe it exists. The only question mark is how reliable it is. The indications are that in intellectually complex activities like go it is necessary and mostly reliable. Which is why we can get away with passable play in blitz games.

[RobertJasiek]

I do not use a single list for storing all my current knowledge of go theory but I use various hierarchical and interdependent forms of storage. Therefore, I do not always need to access all the items of knowledge (such as principles) but I access those relavant for a particular application. (This is similar to abstract data structures used for efficient access, such as a tree with the fast search time log(2) n instead of slow search time n.)

Apart from the already perfect theory, theory is subject to improvement over time. This also applies to interdependencies of knowledge.

I know that there is subconscious thinking but I do not care much for how it works. It has no noticable impact on my go thinking. With the following exception: When a decision is too complex (e.g., a choice between two seemingly equally valuable opening moves), I draw sort of a mental probabilistic lot instead of thinking until the thinking time is out.

There are dynamic principles etc. with dynamic input, such as the result of reading or a mental lot.

Blitz is just faster application of go theory with more errors (blunders or insufficient reading etc.) and lots.

Go theory described by me and learnt by others can be applied directly or, if used for training, indirectly. The problem has essentially never been the contents of the theory but its insufficient learning and training leading to application blunders or omissions. (And then there are missing concentration and other psychological factors, and of course those players never wishing to learn anything about my theory at all.)

[John Fairbairn]

I know that there is subconscious thinking but I do not care much for how it works. It has no noticable impact on my go thinking. With the following exception: When a decision is too complex (e.g., a choice between two seemingly equally valuable opening moves), I draw sort of a mental probabilistic lot instead of thinking until the thinking time is out.

Thank you for trying to explain, but in all honesty I think you've just explained that you accept intuition exists and you use it. So maybe we are back to Tami's point that you may not have the same feel for us native-speakers as to what native speakers understand by intuition. It is not a single concept incidentally: e.g. for me, intuition in go is not quite the same sort of thing as a woman's intuition, and a hunch has nuances of its own... Intuit as a verb can imply an element of conscious thought, and in go intuition can manifest itself differently in slow play and fast play. They way we unconsciously pick up on body language or tone of voice can be classed as intuition, but with elements of instinct. The important point, though, is that we all seem to find common ground - never an exact match maybe, but close enough - in this range of meanings, and that very fact alone tells us that intuition is important, because we never normally analyse the word the way I am trying to do now. We just respond instantly.

[MinjaeKim]

Wow this interesting thread made me register to post.

Here are my thoughts.

I often feel that the westerners (okay I'm an easterner in this context) often think the learning process of this game the opposite way from what I experienced myself directly and indirectly. Compare the kyu level games in KGS and Tygem. In KGS I see that the kyu players do have surprising ability to play an opening that makes sense, but when more stones are added, it becomes quite obvious that they played those moves without really understanding what they are worth. In Tygem, the a pair of kyu players will play a non-sense opening, but at least they do seem to know why they put their stones there and those stones do show some value when the game gets complicated. In other words, you can copy the pro moves and look neat, but it's more important to understand why such moves came out among the alternatives. Without this understanding you're playing some philosophy rather than go. The only way to gain this understanding is by experience and strong reading.

Some people here said that it's important not to develop the 'bad habits'. In my opinion, you really don't have to care. When you get strong enough to win most of your opponents with similar rank, you'll play with stronger ones and see whether your winning stratagy still holds. It won't, so you have to change something. In the kyu level, really minor tweaks can bring huge improvements, and the mistakes you make get obvious enough when your reading improves and more of your groups get killed. Just play your games and in my opinion that's enough to bring you to dan. Knowing some basic opening shapes will help but you really don't need all those fantastic books, yet.

After you stay a while in the dan level, you may feel some unfillable void in your baduk by your own. Here is where the great books help you greatly. If you've come up here 'the hard way', the loads of pretty shapes in those books will be mind opening. Some people here seem to be unhappy for the lack of 'explanation' for the bunch of diagrams in typical go books, but if you really think the explanation is lacking, then probably you're not prepared yet to read those books. Just gain more experience until you naturally feel why those 'shapes' suggested by pros are indeed beautiful.

Back to the main question of this topic. Why are our instincts so bad? It is because you haven't had enough experience. I believe this quite generally can be applied to anything else.

[Kirby]

In other words, you can copy the pro moves and look neat, but it's more important to understand why such moves came out among the alternatives. Without this understanding you're playing some philosophy rather than go.

IMO, learning can happen from the "top-down", the "bottom-up", or some combination of both. In baduk, starting with nonsense theory and learning it on your own through experience is one way. But I don't think it means that you can't learn by imitation, and fill in the gaps with experience along the way.

Compare it to learning a foreign language. You can try to learn all of the vocabulary and grammar yourself by repeated study. But you can also learn by imitation sometimes, even if you don't understand every phrase that you're using 100%.

The goal is to learn everything, so I think there's value both in learning from the top-down and from the bottom-up.

I agree with many of your other points, though.

[Bill Spight]

Compare it to learning a foreign language. You can try to learn all of the vocabulary and grammar yourself by repeated study. But you can also learn by imitation sometimes, even if you don't understand every phrase that you're using 100%.

How do children learn their native languages?

[RobertJasiek]

you've just explained that you accept intuition exists and you use it.

As long as you do not define what you (or, acccording to you, native speakers) mean with the word intuition or as long as you bend its scope of meanings to include everything, it is meaningless to claim that I would be using it.