Quotations references:
viewtopic.php?p=174771#p174771
viewtopic.php?p=174772#p174772

[A discussion arose about whether teaching the principle "Defend your weak stones." or "Defend your weak important stones." is better, e.g., in the book The Basic Principles of the Opening and the Middle Game.]



Here we agree.

(However, the book The Basic Principles of the Opening and the Middle Game teaches neither as or in principles.)



Here we disagree. While it is a possible alternative to teach two separate principles, one must then clarify that the principles apply simultaneously and their application must be joined.

Although there could be principles for
- developing important stones
- preventing development of the opponent's important stones
- defending important stones
- attacking the opponent's important stones
etc.

I disagree that having many principles about closely related topics creates an advantage.



So, after having read the books, what is your opinion about First Fundamentals and Easy Learning: Joseki? I have been told by readers that the contents is extraordinarily clear, but maybe your opinion differs?



While I agree that the list of timing principles in Joseki 2 Strategy is too dense, your other extreme misses the necessity for combined application of related principles.



I get your point, but most important of all is to teach contents at all instead of forgetting to write about it entirely and so hiding it from the readers.

How explicit do you want to get? If I am being a mindless principle-following robot I might interpret "Defend your weak stones" as meaning I should defend my stones when it is my opponent's move which is illegal and would cause me to forfeit the game. "Defend your weak stones when it is your move" is thus more correct but more verbose. Btw, "Defend your important weak stones" sounds better.

Accuracy of better-than-guideline-principles need not be the accuracy of maths. Rather such principles should be correct in the 90% to 99% range.

"Defend your weak stones" is right in 50% to 70% of all cases, i.e., not good enough. The by far most frequent exception is the bad defense of non-essential stones.

"Defend your weak important stones" is right in the 90% to 99% range, i.e., good enough to be taught as a principle.

Now, the most frequent exceptions are, I think,
- the strategic exception to leave behind a weak group on purpose, gain some value elsewhere and then get away with a smaller loss when the opponent profits from attacking that weak group,
- the intentional sacrifice of a previously important group for the sake of an exchange,
- as above, but intentionally leaving behind several weak groups.

You can't embody all Go knowledge/advice/strategy about defending weak groups in one pithy principle, unless you cop-out with "Defend your weak groups, except when you shouldn't". To add further exceptions to your list there is:
- when the opponent cannot profit well from attacking it (which may be due to your tenuki as alluded to in your first item, or could already be a consequence of the situation)
- when it means you would lose the game through lack of points "X can't afford to defend his group, but has to take a big point".
- making some beneficial exchange first: kikashi before defending, or some other beneficial exchange or probe that may make defending unnecessary based on the opponent's answer.
- counter-attacks
- similar to above "take a friend running with you": before running away with your weak group, make your opponent have one too. For example:

Click Here To Show Diagram Code

[go]$$B
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . 3 . . . . . . . . . . . . . |
$$ | . . O . . . . . . 2 . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . X . . . O . . 1 . . , . . . |
$$ | . . . . O X . a . . . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]

If we just think about "defend your weak important stones" we would want to defend the two stones on the left, maybe with an extension around a (but this is too cramped) or a jump around 3. But then white could extend on the lower side. So it's better to counter-pincer to make white have a weak group too and then he runs, which induces us to run out too.

Trying to put all these ideas into one principle will just make a train wreck of language. Seems better to me to explain them all (with examples and exceptions) so the reader can assimilate them, learn, and develop their understanding and judgement.

Of course.

This, however, does not invalidate the purpose of principles. One must not only learn the exceptions and special strategies, but one must first know the ordinary case.



If something is complex, one need not put all in one principle. Instead one can declare that every principle, unless a declared truth, can have exceptions. Such exceptions can be studied separately, but with a reference to which principle(s) they are exceptions.

EDITs

That's what bridge players say.

Bridge, like go, has a lot of "rules" that have many and varied exceptions. So bridge players often say, do X unless you shouldn't, to express such a rule. Go proverbs, OTOH, tend to be simple injunctions; e. g., hane at the head of two stones. In fact, humans perform reasoning with exceptions quite well.

IMO, a better rule would be, Don't defend your stones unless you have to.

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

Visualize whirled peas.

Everything with love. Stay safe.

I am very interested to learn how do you come up with these numbers.
Off of the top of my head, I would say that case #1 is 80%-85% while case #2 is 83%-88%. The exceptions account for at least 12% in my calculations. In either case, the ranges should overlap, I think.

However - the more important question is: don't you need to define "weak" and "important" in your principles?
These principles are meaningless unless the reader knows the exact meaning of those two terms. On the other hand, it seems to me that if the reader knows the meaning of those terms, then the principle itself is trivial, if not redundant. Regardless of which of the two versions of the principle you pick.

I would say that one approach is that of progressive refinement.
You tell the student "Defend your weak stones" and let him practice, and after some time you refine the principle to "Defend your weak AND important stones", and so on... I think we all learned like that, and it pretty much applies to all disciplines, not just Go.

_________________
- Bantari
______________________________________________
WARNING: This post might contain Opinions!!

Many proverbs are weak knowledge in the 50% to 70% range. It is better to replace them by 90+% principles.

And one more thing...

If you speak of accuracy, then the principle "Defend your weak important stones" is also not accurate enough.
Much better is "Defend your weak stones unless there is a better move". This covers many more cases, for example:

• when stones are not that important, but still defending them is the largest move on the board (for example in yose you often defend some unimportant stone which stood in atari for half the game)
• when the stones are important but there are more important/urgent plays
• when the stunes are important but there are some big sente plays you should play first
• etc.

It might be that there are even more accurate ways to present this principle, but the form you gave is definitely NOT 99%. I would say it cannot be more than 88%, as I said before. The new principle I presented here is 85%-90%.

Of course, the most exact version of this principle would be "Make the best move in the current board position." And THIS might be close to 99%-100%.
As you see (also from my previous post), the more exact the principle, the more general it sounds, and the more knowledge and experience is required from the user to fill in the blanks and make the principle actually usable.

_________________
- Bantari
______________________________________________
WARNING: This post might contain Opinions!!

Bantari, the numbers are my educated guesses.

In my principles, I am sometimes more accurate than using "weak"; instead I use "unsettled", "unstable" or "unsettled or unstable" (terms defined elsewhere in a same text). For "(un)stable", I have provided a careful definition:
http://home.snafu.de/jasiek/Joseki_2_Sample.pdf
Needless to say, the terms used in that definition are also defined elsewhere. "Unsettled" is generally known, and my definition is accordingly. Depending on the nature of the text, I define or do not define "important". When using a more easy going approach, "weak" occurs more frequently, such as in "A player's _weak important_ group can be attacked by the opponent easily and must not be sacrificed." in Easy Learning: Joseki, p. 154.

It is an exaggeration that these principles were meaningless unless the reader knew the exact meaning of their used terms. However, I say that accurate knowledge of the meaning of used terms greatly increases the value of such knowledge because then one can (much) more frequently apply the principles correctly.

It is nice to have seemingly trivial principles, because they can be recalled the most easily. However, do not think that every such principle can be (re)discovered easily. For some such innocently trivially looking principles, I needed many years before I could formulate them. OTOH, in the meantime I have educated my go theory understanding enough to (re)discover more principles quickly whenever I start developing some. What helps me greatly for this is having a) the reference pool of already (re)discovered principles in my earlier books and b) the experience how to study game databases or other sources to empirically verify new candidate principles. Discovering the obvious can be the hardest, because it is obvious only once it has already been discovered or expressed clearly.

Yeah, Defend your strong stones, and Defend your unimportant stones, are unlikely principles.

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

Visualize whirled peas.

Everything with love. Stay safe.

I'm waiting for Attack your undetermined if strong or weak or important or trash stones but maybe they can have a ko so decide later.

Like "Notice cuts.". This is a 100% principle, but to notice all cuts, quite some reading can be necessary.

And this is the problem with high-accuracy principles. To cover more cases more accurately, and still be confined to the size of one sentence or so - they need by necessity to assume a lot of knowledge, experience, reading and whatnot from the user. This is why we usually stop at the low-accuracy principles for teaching. Once the reader has the necessary skills to understand the issues involved in high-accuracy principles, he might not need the principles anymore.

Your insistent quest for more accuracy in every word, as admirable as it might be, misses the point slightly - for exactly the reason I give above.

When you look from top down, as we do here, talking about principles we understand but would like to teach to weak(er) players, more accuracy makes sense. But when you look from bottom up, such added accuracy often only confuses the issue and renders the principles useless.

For example, a low-accuracy principle like "defend your stones" might make sense for a beginner. On the other hand, "defend your weak stones" can be meaningless unless the beginner knows how to tell weak stones from strong stones - which requires background knowledge and experience which is not that easy and should not be assumed. When you add the word "important" to that, the problems for beginner get multiplied and the principle even more meaningless. And so on...

And this is why most "proverbs" are low-accuracy.
And contrary to what you say, they might actually be better so, or at least - there is a good and valid reason for them to be so. Not sure if trying to change it is all that good.
Accuracy can have its disadvantages as well, so I am not sure you should generalize as you do when you say "more accurate is always better than less accurate."

_________________
- Bantari
______________________________________________
WARNING: This post might contain Opinions!!

You let this sound negative, as if exploration of the nature of strategy and tactics should not dare to reveal systematic insights.



No. The reason is that previously too few have studied generalised knowledge carefully. Without careful study, only very weak forms of principles or proverbs can be created.

I do not stop there but go far beyond it.



Without a replacement for another, equally successful and efficient approach, he would not abandon their use.



How so, for which sample principle? I prefer to develop bottom up principles. I do not find the time to do so for each principle explicitly, but there is close relation between lowest (rules) level via basic level (such as connection status) to higher level (such as typical, somewhat advanced strategic concepts, e.g., influence).



The lowest levels can be encapsulated to avoid confusion. E.g., strategy depends on left-parts of move-sequences, but this low level can be hidden.



A distinction of weak from strong stones should be assumed, but not every player can make this distinction easily.



No. A beginner can use an easy approximation for "important": "many stones" or "visually big".



I am sure it is very good. Nothing has helped my understanding of go theory more than my principles. Before them, everything was confusing and unclear. With them, everything is clear (as far as the principles reach) up to the point, where principles have dynamic input (such as "[the status] determined and verified by reading"). Without principles, reading (or other dynamic input) often is clueless - with principles, reading focuses on relevant aspects.



Only if you lack the imagination to ignore more accurate details where, by principle, they may be ignored.

Well, I will let most of the below pass, since you are continuing on your usual path of insisting you are correct and everybody else is wrong. We have been there before, and it led nowhere - you learned nothing. So for now just a few remarks about accuracy, since I don't think we have discussed it before:

In theory, perfect levels of accuracy are good.
But then - in theory, theory and practice are the same. In practice, however... they often diverge.

In real life - the level accuracy has to be selected according to the need, not expected to be arbitrarily high or even perfect. Most of the time there is no need to be perfectly accurate (as a matter of fact, it is not even possible) and it is not desirable. Pretty much everything around you runs based on some approximation of accuracy, and the trick is not to make it more accurate but to pick the *appropriate* level of accuracy.

Same with teaching Go.
You simply cannot regurgitate all the details and intricacies of everything you know with 99% accuracy to a beginner or you will drawn him. Knowledge needs to be presented step-by-step, and accuracy increased over time.

For example:
You yourself say that "A beginner can use an easy approximation for 'important': 'many stones' or 'visually big'." - but this in itself is just an approximation with low accuracy, might as well say "defend all visually big groups" since this is what this translates to for a beginner. You seem to admit yourself that such loss of accuracy is needed, and apparently have no problem with that. And this is exactly what I am saying - such loss of accuracy is not only needed when teaching, but unavoidable, and with progress this accuracy gets refined.

The question is - what is the *appropriate* level of accuracy?
Personally, I like "Defend weak groups" better than "Defend weak and visually big groups" as you suggest. Both statements are not very accurate, although you might say that the latter is slightly more accurate. The difference is - the former statement needs to get eventually *refined* while the latter, more accurate one, needs to be *unlearned*. I hope you can appreciate the difference.

This is why "proverbs", which are nothing more but stepping stones to refined accuracy on the path of very low-level learning, are left not very accurate. And I think this is on purpose, backed by centuries of trial and error.

But go ahead, and prove me wrong.

_________________
- Bantari
______________________________________________
WARNING: This post might contain Opinions!!

Wrong.



Wrong. The problem is not that I would learn nothing, but the problem is that I learn unfortunately little on average from others' go theory, which is too weak on average.



No. In go theory (and go teaching), there can be a variety of degrees of accuracy up to "always correct", but some aspects of go theory are easy enough to allow learning and application of truths.



No doubt about that. However, I do not take this as a necessity to present beginners with incorrect advice to be unlearned later. Incomplete or approximative in details, yes - incorrect, no.



No. Even a beginner (old enough to be able to count at all) can, e.g., count numbers of physical liberties. There is no need for him to be imprecise about that.



It depends on purposes, contexts and ability for dynamic factors (such as reading).



This is not what I suggest. I suggest "...important...", and if a beginner (with apparent problems of progress) wants to know what "important" is, I explain to him "many stones" or "[visually] big" (or, in other contexts, there are other aspects of important, such as for important cutting stones). (A more interested beginner I also tell that an intersection with a captured stone is 2 points.)



Since I suggest "...important...", the principle remains the same. Also the means of approximation need not be unlearned, because they are still useful for other applications. The assignment of these approximations need to be replaced by more accurate approximations and later by accurate calculations. This is not the same as "unlearning", because from the beginning it should be clarified that the initially used methods of approximations for a principle are temporary until the player is ready for methods of greater accuracy. The idea always is to be as accurate as one already can and wants to be.

Really? Every sentence I wrote was incorrect?
My my... next time how about you split it by words, maybe some of those were right.
You don't want to break my spirit like that each time. Give me some points at least... A right word somewhere? Heh.

But seriously... I think your worldview is much too simplistic. Still, whatever rocks your boat.
I can point you towards wider horizons, but the journey you need to make yourself.
Until then, "rigid" and "inflexible" is as valid a dogma as any, I guess...

_________________
- Bantari
______________________________________________
WARNING: This post might contain Opinions!!

Best quote ever. This concludes the discussion. Thank you for participating everyone.

Wrong!

Actually this is a pretty confusing thread. Someone should make a few distinctions. While "axioms" for playing go would be valuable, we more often get "heuristics". Those are supposed to allow to find something (your next play, a direction or plan, for example). The one about defending weak groups is actually dependent for explanation on an actual English-language proverb ("a stitch in time saves nine"); because much of the point is that giving up a tempo now to help settle a weak group can save several tempi later when it is attacked strongly but still can be saved. It is certainly not just about life-and-death, which can often if not always be done on a just-in-time basis (''shinogi'').

And so on. The meta-proverb is that explaining a heuristic drags in further concepts. Obviously axioms are supposed to be free-standing.