Is Japanese or Western literature more brilliant?

[Joaz Banbeck]

[admn]

*** Bit Shortage Wrning ***

L19 is runing out of bts. Pls do not mak unncssry psts.
Fr the durtion of the emrgncy, pls abbrvt all wrds lik xlnt + brllnt.
We expct mor bts nxt wk.

Thx,
JB

[/admn]

[tundra]

i just noticed that you dont even understand the rules of go.

under current japanese rule:
in the case of triple ko where one side having one eye and other side having no eye, it is an automatic win for side that have one eye.

did you know that??

Well, this is news to me. The 1989 Japanese rules are here, as translated by James Davies. (As far as I know, these are still the official rules of the Nihon Kiin.)

Article 12. No result
When the same whole-board position is repeated during a game, if the players agree, the game ends without result

I see no mention of one eye vs. no eye making a difference in triple ko. Am I overlooking something?

[Redundant]

Just looked over the paper
your definition is harder than any math class i took.
and you try to sell that to people?

There's no possible way that Robert's definitions are harder than those reached in math classes. For an example, look at the definition of a scheme in algebraic geometry. To parse that, you need at least a semester in ring theory and a semester in topology, and to do anything with it you need a semester in category theory. Yet somehow people learn algebraic geometry and use it to advance the field of pure mathematics. Hard definitions are a necessity for advanced study.

[RobertJasiek]

JB, I try. Bt int. messages deserve detailed rpls.

oren, which strong plyrs/pros do have time for free reviews?

Magicwand, ko def: yes, the def is hard. The paper is not for strength improvement, but research for its own sake. But... the examples can make a reader a bit stronger, like a book with ko examples. Your rules Q: bit warning:)

Kirby, criticising quite a few books of pro players is a criticism of their works, not a personal attack on them. If you think that their works do not derserve the crit., pls show evidence (such as a joseki dict with stone difference stated for every var).

hyperpape, yes, my theory allows new joseki eval if the "joseki" does not belong to the pure tactical reading deeper and yet deeper variety entering the middle game in a pure tactical fight immediately. My theory can correct accepted joseki as unequal, but I have not found such cases yet. What I have found is that different josekis can hv different values. They are all called "josekis" (by pros), but some have a value tendency to the territory end, others to the influence end of the value ratio's range. E.g., I hv found that josekis w big corner territory but groups close to corner can have a high ratio. This makes sense: The territory advantage is compensated by the relatively greater isolation of the corner group, which is a disadvantage. So rather than continuing to speak of "josekis" indifferently, IMO, one should speak more of also joseki types and positional contexts. (I omit / delay more details bec of bt warning. Until then, see J3D.) No, not solely pedagogical. I think, I hv nt gvn these ansrs here bfr, x f asking.

Mef, (3), yes, subject to additional evaluation of relevant strategic concepts for every joseki candidate.

tundra et al, rules pls in rules forum, after bt warning interruption.

Redundant, ko def hardness: it is almost starting from axioms. It is hard because several layers are piled upon each other. Those finding the bottom layers to hard can forgo them and read the ko-intersection defs immediately while trusting the lower levels until they want to study them afresh. OC, one still should apply the DRR.

[lorill]

I AM STRONG AS YOU BUT I KEEP MY MOUTH SHUT BECAUSE I KNOW BETTER!!!

offtopic, but I'd like to see you two play.

[topazg]

where's the evidence that demonstrates these numbers?

In the related book Joseki 3 Dictionary, of course. See "influence stone difference" in the sample http://home.snafu.de/jasiek/Joseki_3_Sample.pdf

Obviously, I don't mean evidence from your own work, that's just self-fulfilling prophecy. I mean confirmation of your work by others, either by confirming your number theory or agreeing with the conclusions your model makes.

2) The chosen influence model works extraordinarily well in the joseki evaluation method. (399 of 400 correct joseki versus non-joseki characterisations. The 1 failure is by nature outside the system: a huge ko exchange during the middle game.)

How do you know it works well? I have been informed by both very strong amateurs and professionals that the conception "joseki is a locally even result" is not true, and that the application of its dynamics impact the rest of the board in a way that some josekis become considerably weaker or stronger depending on surrounding stones. I'm obviously preaching to the choir here as even at our levels we both know this obviously.

However, you no doubt designed your joseki evaluation method, and your model's success in that method doesn't extend to being an accurate metric of evaluating a result accurately. I could rebrand "influence stones" as "Robert stones", and evaluate my corner result as +1RS or -3RS accordingly, but there's no guarantee my evaluation is in any way meaningful or accurate with respect to the quality of my result.

[RobertJasiek]

How do you know it works well?

By studying (and having studied) the theory itself, its assumptions and its application.

"joseki is a locally even result" is not true, and that the application of its dynamics impact the rest of the board in a way that some josekis become considerably weaker or stronger depending on surrounding stones.

The theory allows consideration of the positional context. E.g., "if the player having either [local value] disadvantage has some other advantage of equal importance". Elsewhere in the book series, I stress the important of strategic concepts etc. and state that also they must be taken into account. Furthermore, the functional joseki classication chapter explains usage of josekis in the global context.

The value theory assesses local values but also allows global considerations.

(Also see my earlier messages.)

[Kirby]

Let's stop the personal attacks already.

I don't feel it's much of a personal attack since he started the thread. Is not insulting the works of professional players a personal attack? I personally think he likes the attention.

There's a difference between saying "your book is rubbish" and "you are arrogant".

What would that be? Besides, I didn't say "you are arrogant", I said his statement was bold, if not arrogant.

[Kirby]

I AM STRONG AS YOU BUT I KEEP MY MOUTH SHUT BECAUSE I KNOW BETTER!!!

offtopic, but I'd like to see you two play.

Yes! This would be great. I suspect that at least one of the players won't be up to it, but it would be a great game to watch if it ever happened.

[topazg]

How do you know it works well?

By studying (and having studied) the theory itself, its assumptions and its application.

On what basis has your study evaluated the theory and assumptions? What other strong players have concurred with your theory and assumptions and their relative superiority?

"joseki is a locally even result" is not true, and that the application of its dynamics impact the rest of the board in a way that some josekis become considerably weaker or stronger depending on surrounding stones.

The theory allows consideration of the positional context. E.g., "if the player having either [local value] disadvantage has some other advantage of equal importance". Elsewhere in the book series, I stress the important of strategic concepts etc. and state that also they must be taken into account. Furthermore, the functional joseki classication chapter explains usage of josekis in the global context.

The value theory assesses local values but also allows global considerations.

(Also see my earlier messages.)

Effectively, this comes down to a few separate issues:

1) You claim that because you say the number of influential stones, your method provided more useful information than traditional methods.

This is unsubstantiated. Not only have you failed to explain why these additional numbers provide enough use to be of meaningful value, but you've also conceded that the precise nature of whether a stone is influential or not requires judgemnt - something which as a 4.5d you will find considerably easier than your target audience. You've also conceded that it doesn't discuss the amount of influence each stone offers, either in itself or due to its relationship with others. This is an honest and fair limitation of your model, but you haven't demonstrated that the limitation is free from significant impact on the overall evaluation. As a result, you have demonstrated beyond doubt that your method provided more information, but not necessarily that it provides better information.

2) You claim that the method has passed the test (although it's not clear what this test is from this thread) on 399 out of 400 josekis.

This sounds like self-congratulatory fluff, even though I'm sure that you went through considerable pains to get the results in the first place. In what way did it pass these tests? How as these tests good benchmarks of quality?

3) You claim that theory allows for an application of wider strategic knowledge.

This sounds like saying "well, I have this great method, but there are factors outside of its control, so if my method doesn't seem to be working, it's those other factors at fault, not my method". I'm aware of the complexity of integrating all these different aspects into one model, and I don't envy the task, but I think you're approaching your proofs in a non compelling and not particularly valuable manner.

If I may, I'd like to suggest an application of your theory that would be of interest to me and most of the community following this thread I suspect:

1) Can your theory and method evaluate the quality of the result in josekis played professionally in a whole board context?

2) Can your theory point out the incorrectness of josekis considered acceptable by current professionals?

3) Can your theory point out improvements to existing josekis where controversies remain, and give reasoning why one line is superior to another?

4) Can your theory make groundbreaking advances by proposing improvements to existing josekis with substantive reasoning that professionals can use?

5) Can your theory offer contributions to which joseki to pick in whole board situations - which pincer is appropriate, and how to best handle opposition deviations?

These are all questions that, if the answer is yes, can provide genuinely great advances to people's application of understanding and knowledge in real game contexts. So far, I've not seen anything that can answer them, but that doesn't mean your methods, theories and models can't do this.

[daal]

However, you no doubt designed your joseki evaluation method, and your model's success in that method doesn't extend to being an accurate metric of evaluating a result accurately. I could rebrand "influence stones" as "Robert stones", and evaluate my corner result as +1RS or -3RS accordingly, but there's no guarantee my evaluation is in any way meaningful or accurate with respect to the quality of my result.

It might be interesting to start with the assumption that Robert's method in fact produces an evaluation that is equal than that of a professional who presumably relies on things such as his or her professional sense, experience and reading. Why not? He says it does. In other words, whenever the question arises in any game: "how good is the result of joseki A?" both Robert and a professional would come to the same conclusion.


Obviously, this wouldn't be the end of the story. Clearly the professional would be in a better position to put an abstract result into a concrete use than you, I or Robert, but it would nonetheless be a fine thing. How would it compare for the average amateur with the current methods that amateurs rely on? Well, how do amateurs evaluate josekis? In one of two ways: They either do it like the professionals do, only worse, or they rely on the judgement of professionals. Relying on the judgement of professionals is questionable at best. When we read their joseki dictionaries, typically all we see is a corner and indeed, that is hardly enough to make an accurate assessment. Relying on one's own ability to determine what's best has the advantage of taking into account the whole board, but the disadvantage that there is too much information for most of us to process. Wouldn't it be great if there was a better way?

Again, assuming that Robert's method provided professional quality results, the question would be for whom would this method be preferable?

First of all, it would be for someone capable of applying the method. This rules out people such as myself, with limited analytical ability. Although I would love to be able to use a "Robert stone", I cannot bear the thought of using anything resembling an equation during a game. What is the level of analytical ability necessary? We can assume that Robert's is good enough. Are there any other people who can use it? I don't know, but it would be interesting to find out, would it not?

Robert - do you know of any of your readers or students who can and do apply your method? If not, perhaps we could find a volunteer to put your theories into practice. What would that volunteer need to know? Would he need to purchase all of your books and read and digest them - or would one suffice? Would he also require training in the method? How much time would it take for a sufficiently analytical player to learn to apply your method?

[quantumf]

2) You claim that the method has passed the test (although it's not clear what this test is from this thread) on 399 out of 400 josekis.

This sounds like self-congratulatory fluff, even though I'm sure that you went through considerable pains to get the results in the first place. In what way did it pass these tests? How as these tests good benchmarks of quality?

There's also the usual risk of bias - i.e. consciously or unconsciously including items that work well and rejecting those that don't

[Magicwand]

If. Robert's claim is true we can have computer program strong as professional soon.
I for one not going to hold my breath.

[topazg]

As I respond, I'm going to start by saying what a lovely post that was, nice to see some really good debate going, particularly with someone who owns some of Robert's material (I appreciated your feedback comment on the book providing improvement by the way, such things are always valuable):

It might be interesting to start with the assumption that Robert's method in fact produces an evaluation that is equal than that of a professional who presumably relies on things such as his or her professional sense, experience and reading. Why not? He says it does. In other words, whenever the question arises in any game: "how good is the result of joseki A?" both Robert and a professional would come to the same conclusion.

I agree. If we can apply a system that allows a beginner to use Robert's method to choose a collection of moves, joseki or pseudu-joseki, in a given board position, and ask for a professional judgement on it, this would be interesting. The problem is this isn't necessary an objective assessment either - as magicwand has said, even professionals can completely disagree on whether a result is good or not, and the New in Go article I linked to shows how professional opinion can swing backwards and forwards over time.

However, I would still trust a professionals assessment on the quality of a result more than I would Robert's, simply because if they played 100 games against each other, Robert wouldn't win very many. I think it's relatively reasonable to assume that the stronger the player, the better the judgement in how good a result was. If you asked a beginner, boidhre, you, me, magicwand, Hwang In-seong, or Lee Changho the relative merits of a position, I would expect the average player's opinion of each viewpoint to be in ascending order of credibility across those players, and I think that's as good an assessment as we can make.

How would it compare for the average amateur with the current methods that amateurs rely on? Well, how do amateurs evaluate josekis? In one of two ways: They either do it like the professionals do, only worse, or they rely on the judgement of professionals. Relying on the judgement of professionals is questionable at best. When we read their joseki dictionaries, typically all we see is a corner and indeed, that is hardly enough to make an accurate assessment.

How interesting. I'm probably one of the few people on here that doesn't own a single joseki dictionary, but all of the professional judgement I've seen of josekis has always been with regards to a whole board context. I don't remember seeing any commentary that referred to the quality of the local result in isolation. Maybe I spend too much time reading professional game commentaries.

I agree that relying on the judgement of professionals is questionable, but I see it as less questionable than reliance on anything else, as we don't really have any empirical data to go on.

Relying on one's own ability to determine what's best has the advantage of taking into account the whole board, but the disadvantage that there is too much information for most of us to process. Wouldn't it be great if there was a better way?

Yes. It would also be great if I never missed clever opponent moves and didn't repeatedly blunder away my own groups. If is a very big two letter word

First of all, it would be for someone capable of applying the method. This rules out people such as myself, with limited analytical ability. Although I would love to be able to use a "Robert stone", I cannot bear the thought of using anything resembling an equation during a game. What is the level of analytical ability necessary? We can assume that Robert's is good enough. Are there any other people who can use it? I don't know, but it would be interesting to find out, would it not?

This is part of the crux of an earlier point I made. This particular conversation was about Robert saying his method is better as instructional material because it provides the reader with more powerful tools than traditional professional teachings, allowing them to make better judgements. If you can't use them effectively, and so far I'm not convinced I would be able to, who indeed is supposed to benefit from the material? My understanding is that you are right in the perfect skill level to make use of his material - maybe this part of his material is not yet up to the usefulness of some of the rest of it?

I don't think we can assume Robert's is good enough. We can assume that he thinks it is, but we can't therefore assume that his use of it constitutes valid strategic decision making based on this method. Again, if Robert said "my method says this result is good for Black", and 3 out of 3 professionals said "no, it isn't", I would consider Robert's method to be at fault unless it could be demonstrated otherwise. Professionals repeatedly have to demonstrate their understanding of positions by the games they play, Robert's method hasn't passed such tests yet as far as I am aware.

Robert - do you know of any of your readers or students who can and do apply your method? If not, perhaps we could find a volunteer to put your theories into practice. What would that volunteer need to know? Would he need to purchase all of your books and read and digest them - or would one suffice? Would he also require training in the method? How much time would it take for a sufficiently analytical player to learn to apply your method?

All good questions. What I would like to see is an intelligent and analytically minded 12k take this method, apply it to a few joseki decision points in a whole game, and ask a professional player to judge how good the end results were. To me, that would provide interest insights into the applicability of the method in question.

[daal]

Probably my comment about professionals or their Joseki dictionaries not presenting enough whole board context is faulty - I don't own a joseki dictionary either. My main sources of joseki however such as Kogo's or a book by Naiwei Rui don't say much about whole board context and usually seem to end with something like: "this result is considered even," or if I'm lucky, "this is good if white has support around 'a'."

As to Robert not being as good as a professional despite his methods; I think this doesn't discredit his method at all. Even if his assessment of a local result were perfect, he would still have to make the best use of that result in the ensuing conflicts, and stronger players are probably simply better at this than he is.

While I would also assume that a professional's opinion regarding the result of a joseki were more reliable than that of another player, I'm not 100% convinced that it must necessarily be better than some sort of careful mathematically-based analysis. If I understand Robert correctly, the business of "399 of 400 correct joseki versus non-joseki characterisations" means that his method typically comes to the correct conclusion - however that's defined - of whether a sequence is or isn't joseki.