I agree, but generally a professional will be able to recognise when the result was good but Robert's subsequent play was insufficient. It would be interesting to have two roughly equal professionals play with one of them basing their choices exclusively on Robert's method, but sadly I just don't see it happening.



How can anything be considered correct without a reference point? We can't say that a mathematical approach is better than professional judgement and use test success as a criteria to demonstrate it, when the test analysis is simply a comparison with how closely the method matches professional judgement

That's exactly why I asked what 399 out of 400 actually means. How do we know that some of those josekis aren't actually inferior, and eventually they'll be superceded by improvements later on? I'd actually be more impressed with a score of 387 out of 400 with an output from the method that says how those 13 josekis could be improved on, which could then by opened up for discussion, ideally with professionals input.

An impressive performance at matching the status quo doesn't predispose me towards the idea we're viewing a groundbreaking approach here

Of course, that's irrelevant if the point is that the method allows weaker players to reach joseki quality outcomes by applying the principles in the method. However, you've already said you don't think you could use the additional information that Robert alludes to, to make improved decisions yourself. I don't think I could either. If the purpose is valuable additional instructional techniques, who is the target that benefits from it?

My original supposition was that the additional information doesn't actually help with the application of knowledge, and I'm not seeing contrary evidence yet.

topazg, I understand your curiosity and do not mind your devil advocate's play. At the same time, I want this bit cencorship of my user name to end as quickly as possible. So, in every message, I need to guess (it is the only relevant source of decision I have) what admins might think of it. Currently I am still hoping that admins do not mind factual discussion. If I find out that they do (because good factual discussion can require many bits easily), then I will need to shift my place of replies to rec.games.go, where there is no bit censorship for factual discussion and no confusion of it with flooding. BTW, for those joining reading of this thread late, I am Robert Jasiek, usual L19 username RobertJasiek.



The basis includes
- the ca. 400 josekis, uncounted number of major variations and non-josekis occurring in J3D,
- the pretty representative nature of that joseki selection,
- on the educated assumption of the correct komi being 7 (compare J2S ch. 4.4.1),
- the assumption that stone difference, territory, influence, turn and stability, strategic concepts (such as development direction and global positional context) etc. can all be relevant,
- assumptions for forming the ration of territory and influence,
- assumptions for transforming values for stone difference 1 or 2 to stone difference 0,
- the assumption that dominated aspects can be ignored.



With my joseki evaluation theory as a whole: none so far.

With the assumptions: It depends on which assumptions. From many (e.g., that komi 7 is the best educated guess) to none so far (e.g., that the stone difference 1 transformation "imagining a local continuation of, e.g., 1 extra play" is possible to enable a stone difference 0 evaluation.

Their relative superiority: see before.

Let me comment: I am not aware of stronger players buying my books. Maybe they exist, but if so I do not know who they are. Anyway, there would be only a few, especially for the still new book JD3. The next question would be: Is anyone of those reading L19 and wishing to express his opinion? I mean those having read the book - not just the sample.



No. I refer to the DIFFERENCE of numbers of SIGNIFICANT OUTSIDE influence stones. Which traditional methods of assessing influence? AFAIK, they do no even exist! (Symbolic number guesswork models for specific applications exist, but there appear to be only occasional proponents, so it would be a very great exaggeration to call that "traditional methods".) AFAIK, it is me who has invented methods of assessing influence and defined influence in a measurable manner. In particular, numbers of influence stones is one such model I have invented.

So what I am saying is: This my influence stones method is very good for josekis (and for determining middle game center dominance). Having this method is much better than (as it was the case previously) not having any reasonable method for assessing influence.



Of course, and it is not what I claim, see before.



Sorry, but I have explained it earlier.

1) Joseki: they agree to professionals' judgement about is or is not joseki (or equivalent to joseki), as far as I have studied it for the represenative selection.

2) Particular josekis: the ratio values retrieved by the theory make good sense for every particular joseki.

3) Middle game: I have used this in my games' planning and it has accelerated decision making drastically. (Similarly: in my teaching of pupils.)



To save bits, I ignore the Turing Test contained in your statement.

Yes. Therefore JD3 has examples demonstrating the theory. Besides, the 400+ variations with values stated below the diagrams allow the reader to exercise his value determination skill.



Why. See my earlier messages with 399/400. If you don't trust my numbers, recalculate them! If again you are referring to other aspects such as global context, read the related information in my book series on about two thirds of the pages. You can call that insufficient, but then I suggest that you call every other book on the topic countless times as insufficient.

Not only have I "not demonstrated that the limitation is free from significant impact on the overall evaluation", but rather I (also in the book series) stress or imply frequently the importance of positional context and strategic concepts. E.g., notice the JD3 chapters 2.1.3 to 2.8.3!



The all professionals' implicit agreement on which results are joseki and my evaluation method's judgement pretty much agree. In this respect, I provide the same information.

Indeed I provide more information (except concerning sheer number of joseki or failure variations and the like).

I provide a better method in the sense of being more efficient: You do not any longer need to refer to or ask all professionals, but now each player can (if he wants to) reliably evaluate josekis or non-josekis by himself.

I provide also better information because I show structure how quite a few of the various bits of information can be integrated with each other! Besides it becomes easier to recognise essential vs. immaterial information.



Read my book, judge whether there are indeed ca. 400 josekis, judge whether they are a reasonably respresentative selection of josekis and recalculate the values. If this does not convince you, then take out your 50,000 further josekis, calculate the values for them and report.



1) I calculated every value about 5 times to be as sure as possible.

2) Then the value types explained in the book are applied to every tested variation's values.



That is the evil perspective:) The positive perspective is: There are more aspects for a joseki than stone difference, territory count and influence count. These are integrated in the value part of the theory, which does not contradict but can be enhanced by the relevant non-value parts.

E.g., if Black has an advantageous ratio while White has an advantageous development direction, then it is fair. E.g. if Black has both an advantageous ratio and an advantageous development direction, then the result is "favourable for Black".



I have not discovered the theory of everything yet. Having found a very good joseki evaluation method does not determine perfect play until the game end yet. Sorry for that.

If you apply your skepticism to previous joseki judgement ways, then you must be more sceptical, mustn't you?



Partly. The joseki evaluation method itself assesses local values. Strategic concepts etc. can also be considered to get a nearby enviroment judgement (up to, say, adjacent corners and center of the board). To get a whole board positional judgement, one must furthermore apply methods of whole board positional judgement and analysis, see J2S ch. 4, 6, 8 etc.



Answered earlier.



Yes, but tewari etc. also remains useful. In particular, when two different variations are almost equal (e.g. in my theory's judgement), then tewari might be the better tool. OTOH, tewari might not be able to judge whether either variation is joseki or equivalent to joseki while my theory can do this.



My theory is not a tactical move finder. Once a sequence of moves is halted (e.g., by manual interruption), my theory can be applied to judge or contribute to a judgement of the (temporary) result.



Contributions, yes. But... J2S ch. 6, J3D ch. 2 + 4 are more relevant for that purpose.



J1F and J2S are more relevant for these questions.



Indeed, therefore my J123 books make great efforts to provide all the necessary tools or principles.



Have you read my books?

From what I understand, you checked 400 actual joseki and your theory confirms that they are joseki. If this is all you've done, this is not a sufficient test. To make it complete, you need to add 400 sequences that look like joseki to kyu or low dan players, but actually aren't. Unless your theory correctly identifies them as not joseki, it's worse than useless.

E.g. my competing theory is as follows: every corner sequence is joseki. This worked for 400 out of 400 corner sequences in professional games.

Concerning the joseki evaluation theory, you must be able to count stones, to distinguish influence stones from rather not influence stones (or stones on the inside), to count territory (at least roughly, this is learnt from 5k or stronger), to devide two numbers, to add / subtract 1 to / from a number or to add / subtract 7 to / from a number, to notice if something else is particularly important (such as a big gap in a wall). IMO, all tasks are reasonably easy. The most difficult part for kyu players may be to identify a necessity for a strategic concept's application. This is learnt while becoming low dan.



Can: see above. Do: I have motivated a few students but will have to see if they continue to apply it. Do, influence stones in middle game: yes (some have learnt it, others agree it is useful but still need more practice).



Each reader can if he wants. The concepts are meant for application.



No. Each volume can be read pretty much independently. But there are things useful also for a more profound understanding in other volumes.



Most methods / knowledge need training. (There are exceptions. To save bit width, I keep them secret:) )



Between no time and weeks. Players are so very differently skilled!

Thanks for the in depth reply Robert



I'm pretty sure professionals are on the whole very happy with their methods, so prior methods obviously exist. The fact they may not be mathematically written out in stone with finite numerical outputs doesn't necessarily mean that that's not what's going on subconsciously. Of course, written down methods for instructional material is beneficial, provided it serves the purpose.



Most other books I've read are making rather less grandiose claims I trust your numbers, I don't yet trust the process reached to calculate those numbers (thus making the purpose of doing further calculations rather pointless). The value of an outside influence based stone varies depending typically on what's going on on the facing side. It's hard to say that the model accurately confirms joseki theory unless it explicitly takes that into account. If it's a more generic confirmation that doesn't take into account the additional factors, then what makes it a valuable tool for people to apply, when those additional factors _will_ be present when it comes to apply them?

I'd be very keen on Yilun Yang's impression of your method, particularly as he has written a lot of material on what joseki to apply to what situation and why.



Can they? Unless I misunderstood, daal says he finds it difficult to feel confident that he can apply the methods with confidence. I would be interested to know how many players choose to disregard joseki from memory in favour of the application of principle (even if the result is the same). I would also contest that, even if your method is fast to apply, simple memorisation is faster - obviously that doesn't help with respect to deviations from joseki, and in which case your method could prove superior. Is there evidence of disciples of your method consistently creating good results by applying your methods when josekis are deviated from?



Read my response earlier for why I consider this irrelevant.



That doesn't mean that the calculated values are useful or a demonstration of quality. Just that they have been calculated. Against what benchmark are they compared?



Even this is a simplicity. It avoids the more complicated issue of how ratios that go in different directions but by non-equal amounts equates to a not quite equal result. Can we confirm situations where a professional can agree that, in this circumstance, when a slightly biased outcome favouring one player happens, that professionals agree this to be the case. There are a few josekis that are locally favourable for Black but globally favourable for White. How does a player know when to apply your method and how to incorporate other factors to avoid applying it correctly and ending with a bad result? In these situations, what additional value did they have from using your method in the first place as opposed to using more vague general principles?



Indeed, I happily throw skepticism about quite liberally. You receive more than your theories deserve primarily because you make such strong statements as to their brilliance. In general, bold claims require strong evidence, outlandish claims require extraordinary evidence. When you say that your method is the best yet, the burden of proof lies with you to demonstrate it, and I haven't yet seen it demonstrated



I don't think you can have a sound method without an incorporation of these in one system, particularly not when being used as instructional material. The way I read this response of yours is "yes and no, to be honest the value you get out of my method is only useful in a wider context", which to me is no better than I already get when reading other professional comments about joseki evaluation. Maybe I'm stronger than the target audience?



Is it, I don't remember that? Which josekis does your theory consider incorrect, and why?



Great, can you please post some josekis in which your method has provided the answer to a controversial joseki area.



What value does your system have in evaluating a result, prior to the application of other information?



No, and I'm not going to pay money to see if your arguments are sound. If there are excerpts that demonstrate your point, I'm sure you can post them accordingly (or someone who owns them could).

Similarly, I could assess every joseki based on the number of empty triangles it creates. Less is better. Any sequence that doesn't involve an empty triangle is a good sequence, regardless of whether or not is joseki. That would have some interesting results

Magicwand, current programs are strong because of calculation power (MC & Co). To apply knowledge, expert system programs are needed, and I would not dare to say "soon" for them. They do particular things very well (see GoTools for particular types of Life and Death problems). Integrating all things is the real problem.

topazg, the joseki evaluation method is not meant for beginners! (Well, you can test if it is nevertheless, but I would be surprised.)

Trusting a pro more because he has played 100 games? This is less than the 70,000 GoGoD games (we do have this empirical source, and my method relies on that evidence!), on which I ran Kombilo to check every move to be declared as joseki in my dictionary. Where necessary, I also checked changes over time.

You doubt my ability of strategic decision making. Please read J3D and point out where you have doubts about my global context advice and strategic choices statements!

You are suggesting that my joseki evaluation method would be the best tool for a 12k tester to make strategic decisions. It is not. J3D chapters 2 and 4 are more suitable for making strategic choices and a 12k make make the latter easier with the former (if he has the book in front of him).

You think my method's success it too good and would wish it to identify pro mistakes? My method is so good because, fortunately, pros' judgement has been so good! This is now confirmed. Nevertheless, a careful study of values generated by my method points out where some "josekis" are more or less profound. (So far, I have done that only mentally.)

daal, stronger players tend to better at decisions but there are their occasional mistakes, which I would not do in the positions then.

EDIT:

palapiku, let me repeat: I have also checked that every non-joseki variation I tested is identified as such by the theory.

Perhaps lost in these walls of text. Did an admin change your name or did you do this as a joke? I couldn't find a place to change my username, and its not cool if others are changing your name.

Temporarily, yes. Apparently, as I guess, as a "bit width warning" or something like that.



It is pretty much the most uncool thing I have seen for years.

I don't understand this at all. Is this some kind of joke? What's going on?

As a mod, I can say I have no idea at all. Joaz?

I assume that this involves examining a corner sequence, doing some counting, some dividing, some subtracting and coming up with a result close to 0. In which book is this process described? I couldn't find it in book 2.



This makes sense. In order to evaluate a joseki choice, one must be able to produce the variations from which to choose. If we were to devise a test to substantiate your evaluation method, how strong would the player have to be to make professional quality joseki choices using your method? Are you convinced that a month or so studying your book would enable him to do so?

_________________
Patience, grasshopper.

If you're going to pull the admin card out, then it least be serious about it.

this is stupid.

_________________
Occupy Babel!

Hey! his thread isn't about trashing mods! Can't you people read the title?!


Uh... By the way, I recently read the opinion that there is "an incomparable depth and richness to" Japanese go material.

_________________
Patience, grasshopper.

+1

This looks like an abuse of mod powers to me.

Edit: "This" refers to the changing of Robert Jasiek's username.

topazg, for letting influence assess what is going on elsewhere, there is my precise influence definition, which, however, is not compatible with my joseki evaluation method (you cannot devide a number by a 5-tuple).

You can ask Yilun Yang:)

daal has referred to some of my methods, I do not know if is already referring to the joseki evaluation method. Which is daal's real world rank?

Unlike you seem to suggest, I do not advocate to forget about joseki variations (just because of the joseki evaluation method); in fact J3D offers also them for learning.

Benchmark: My method could serve as one for some other approach. But is there any benchmark to assess my method (other than what I have described and you do not accept as benchmark)?



We can: apply my method (or read the values created) and compare it with pro statements for such local positions. I have seen such but not written down the evidence.



I would not call it joseki but standard variation for special circumstances.



Always when he thinks he can gain information from it.



How: consider the other factors. Judge for whom they are advantegous.

Avoid incorrect application: be careful with your judgement.

Judging end result: judge. Use my method and positional analysis methods. Use strategic planning. (Details: see J123.)



Precision. (At the moment, I lack time to find more. How does Bill say it? Gotta run!)



See previous posts. - There is not remotely a second similar generally applicable joseki evaluation theory at all. So already trivially mine is the best.



By this argument, every other go theory book is worse. I make at least attempts of integrating the various aspects such as strategic concepts and analysis methods in the overall strategic planning. Where else do you see any such attempt?



Decide. Before you complained that the wider context was missing.



I will repeat when my user name is correct again. Otherwise I better save bit width.



No. J3D teaches frequently useful josekis - not the latest controversies. Will be interesting future study.

(Gotta run. More later.)

Oh come on, this is hilarious.

I would probably email any who speak English and teach regularly on KGS ( http://senseis.xmp.net/?KGSTeachers ). Who knows, maybe some would be interested or maybe not. You would have nothing to lose. I think it's worth the effort for you to learn more from stronger players and be able to present even better material in the future.

Has Western literature ever delivered any innovations, or is it purely derivative?

_________________
North Lecale