Robert Jasiek's Go Theory Research

[Kirby]

I thought this is a thread about theory rather than facts

Perhaps, but this isn't the first time a call has been made for "factual discussion", so if we're going to say that, it seems natural to be "factual" about our definition of "factual".

The fact that this argument is kind of circular should indicate that I don't feel that "fact" is universally agreed upon.

That being said, I will exit this thread.

[quantumf]

Perhaps some examples:

Any number multiplied by 0 equals 0 - fact
Salt is bad for you - not a fact

Playing on the vital point of a bulky five shape reduces it to one eye - fact
Chinese pros are stronger than Japanese pros - not a fact

[HermanHiddema]

As has been discussed in another thread, Robert uses the word factual as having the same meaning as the German "sachlich", which is not really correct. Perhaps a better term would be "objective".

[RobertJasiek]

In viewtopic.php?p=127304#p127304 John Fairbairn compares me to the chess book author Liu Wenzhe. Although I find it much more relevant to compare go authors or researchers with each other, every such comparison should not just count numbers of described aspects but should also take into account the effort to reveal findings.

John suggests that having a full-time research job allowed creation of more research. While there is a spark of truth in it, quite a few amateur researchers (such as in mathematics or natural sciences) did spent extraordinary amounts of time on their research. E.g., Fermat or the early Einstein (working as a patent officer third class) were such examples. At times, during my research period since about mid 1996, my time spent for research was small - at other times, it was great. E.g., creation of the Japanese 2003 Rules (and their inherent discoveries of go theory) kept me busy for 11 months every day from awakening to falling asleep (and occasionally I was doing research even in my dreams, as noticed when awakening from a dream with the solution for a problem of the previous evening).

John wonders how much go theory can still be expected for go by the kind of establishing facts in the style of Liu (if it were for go) or me. About 1/2 or 1/3 of the top level go terms are defined reasonably or very well. On the order of magnitude of 1/10 top level classifications have been given. However, both just scratch the surface because lower levels can study many more details and, much more importantly, decision making is extraordinarily deeper. Thus, research so far is nothing compared to what it will be when, one might hope, go will be solved completely one day in the far future. John, you did not overlook that decision making is much more complex than (traditional) top level terminology and classification, didn't you?

[tj86430]

As has been discussed in another thread, Robert uses the word factual as having the same meaning as the German "sachlich", which is not really correct. Perhaps a better term would be "objective".

My (internet) dictionary gives the following translations:

sachlich -> objective, factual, functional

factual -> faktisch, auf Tatsachen beruhend, tatsächlich, sachlich

While I'm not a native speaker in either English or German, and I only understand some German (I'd like to think I understand English quite well), the way I understand "sachlich" is that it means mostly what we Finns call "asiallinen" and Swedes call "saklig" (which is, as anyone can see, of the same origin). This indeed is not really "factual" as in "based on facts" but more like "appropriate" or perhaps "correct", "objective", "to the point", "matter-of-fact".

[RobertJasiek]

In viewtopic.php?p=128288#p128288 Knotwilg mentions a conflict between belief in the amateur expert (stating me as an example of somebody claiming to be an expert) and knowing that it can't be true. I have written down knowledge on a couple of topics. If it "can't be true", then please demonstrate any factual mistakes in it and why professionals would be the better experts on the same topics!

[uPWarrior]

While not wrong one may question whether your facts/knowledge are relevant to play this game any better.

I refrained from commenting on your research after asking you where I could find your papers, but I found the 4 papers about KOs that you provided to be not much more than a list of positions. I do not think that this "research", though time consuming, in any way provides a better understanding of the game or even about KOs. You just beat the professionals by exhaustion.

[RobertJasiek]

uPWarrior, the Ko and Dame Endgames paper is relevant for playing better under Area Scoring. Maybe "only" on average 1 point for the final score per game. But, think about it: If there would be 100 such papers, you would be 100 points better per game... - Under Territory Scoring, the paper is almost useless.

The 3 papers about ko definition and ko types are almost useless for players and their playing strength. These papers are important almost only for researchers. In fact, the papers are research papers rather than go books.

Most of my go theory that is specifically useful for players you do not find in research papers, webpages or online messages, but can find them in the commercial literature. (There are a few exceptions: a few things are already available, e.g., at Sensei's.)

In general, my go theory falls into two major types: either 1) written for go players' strength improvement or 2) research for its own sake.

I have listed the, in my opinion, most important inventions for players. uPWarrior, you seem to be doubting even their relevance for strength improvement? Consider, e.g., the 4th most important, the definition of stability. I see most players around (European) 5 kyu still making too many mistakes WRT to creating stability versus playing elsewhere. Everybody wishing to become a dan player is required to develop an understanding of stability expressed by or very similar to the definition. How can you call such essential knowledge "irrelevant for players"?!

Knotwilg, in viewtopic.php?p=128299#p128299 you miss deeper understanding in my writings, when compared to Go Seigen's writings. To give a useful answer, can you please clarify which of my writings you have read? What do you mean by deeper understanding, as conveyed in writings?

(You don't just mean that a writing was specifically for your 2d level, do you? That could explain a lot because so far, in the texts for strength improvement, I have written also for kyu players and not exclusively for dans only. I could write specifically for dans to stress and encourage a deeper understanding, and for the better show that I have some:) However, some of my writings do also contain knowledge meant for dans. So I am really curious what kind of deeper understanding it is that you miss.)

EDIT:

You doubt that my selection of aspects, or my way to deconstruct and reconstruct knowledge, is the path to follow.

I do not select particular aspects to disregard all other aspects. Instead I have selected particular aspects for research or writings because I can, to simplify a bit, only explore one new aspect at a time. Basically, all topics of go theory could be important one way or another.

There are many paths to knowledge, and (apart from mathematically proven truths) certainly my way of structuring knowledge is not the only way. Apparently unlike you, I think that my way is a very good one!:)

[uPWarrior]

I obviously cannot comment on all the papers you've written as I've just read the 4 KO papers you posted, so you need to take my criticism with a rock of salt (not just a grain).

You missed my point on the second paragraph, which was the most relevant. When I said that these papers don't provide "a better understanding of the game" I didn't mean that players would not improve by reading them; they might, be it 1 point or 100. However, research is not about making a specific piece of knowledge accessible to the masses (that's what books are for) but rather about finding things that were not known before. Do you believe that professionals or top amateurs are not aware of the concepts you are describing (i.e. playing KOs to perfection in situations where KOs is all that's left)?
I think you confuse research with the writing of knowledge in a paper format. That being sad, I have no doubt that amateurs could benefit from these concepts - but I don't think this is the right/simplest form of communication.

[RobertJasiek]

Ah, ok, so you have meant that the 4 kos papers did not find things that were previously not known. However, this is exactly what those papers do: report findings of previously not known things! The first message of this thread tells you about that in a few keywords.

The external / internal ko paper is the simplest: all its theory is new. Previously those few professionals trying at all to explain the difference or general application of the supplementary ko rule all failed or even said during all of the WMSG 2008 official rules "explanations" that "now is not the time to pose questions".

Basic ko types paper: virtual-force and other fundamentals are new. Definitions of ko terms regardless of the used input ruleset is new and a major breakthrough. All the definitions of ko types are new. Previously, ko types could not always be distinguished unequivocally. The ko type Dead Ko was previously not clearly understood to be a ko type of its own right (except that I had made related preliminary remarks during previous years). The ko type Closed Death is my invention (I had anticipated it a few years earlier but had not been aware clearly of it being an extra ko type). (If you really want to see salt, compare the Ing ko rules' descriptions of ko-types;) )

Ko definition paper: From default restriction rules to answer-force, the fundamentals contain important inventions. (Compare the Ing rules' word (and almost nothing more) "invariation" to anticipate the default restriction rules.) The local- and global-ko-intersection types are new inventions. All their conditions are. (In 1996, Fred Hansen et al noticed that there was some relation between Ing rules and cycle-set, but could not figure out which relation.) In particular, the last condition of global-ko-intersection is a major breakthrough. In retrospect, it is "obvious"; that's how good it is! It was so "obvious" that previously nobody could fathom it. The ko definition is a new invention, and my masterpiece. It is also an extreme advance to the previous state of affairs, which was my late 1990s' attempt, which defined ko so that, as a consequence, each stone in each position was a ko stone. Yet earlier, Ing rules spoke of "stones that can be captured cyclically or repeatedly", which was not even a definition (but added unnecessary confusion of referring to two ko types). The examples collection is the by far most relevant selection for the sake of long cycle study at any one place.

Ko and Dame Endgames Paper: Little new because earlier during the year of publication I had done the mathematical proofs;) However, if you take both the proofs and the paper together, then everything is new from a research perspective. Previously, the basics or score calculations were known only informally. The paper also gives clear summaries and works out the case conditions carefully. Previously, it was all informal talk about "half the number of ko threats" - nothing like "smaller than or equal to half...".

Summary: the ko papers reveal many new findings and clarify many things accurately!

[Bantari]

The external / internal ko paper is the simplest: all its theory is new. Previously those few professionals trying at all to explain the difference or general application of the supplementary ko rule all failed or even said during all of the WMSG 2008 official rules "explanations" that "now is not the time to pose questions".

Sorry, Robert, but we have had this discussion before, and I guess it did not sink in. What you say is just logically incorrect.

The fact that pros were unable/unwilling to explain something (to your personal satisfaction) at any specific moment is by no means equivalent with the fact that they do not know it. The best we can say - and that's also a big IF - is that they might not have this knowledge catalogued and wrapped up in a series of definitions/theorems/lists the way you like. But more likely they might have had other reasons to brush you off.

To support your assumption (that the pros do not know the stuff you write about) you would have to demonstrate that they ROUTINELY misplay the kind of situations you analyze in your papers. It is important to demonstrate a TREND - like they misplay such situations A LOT - not just that every now and then this or that pro makes a mistake. It would also help if you could then also demonstrate than AFTER reading your papers the pros improve their playing, although this would be just icing on the cake.

Until you do the above, I think it is safe to assume that since the pros are usually playing ko situations and endgame sequences much more exact and 'correct' than you or me, they also understand it much better than you or me. Otherwise we have to assume that they somehow BY CHANCE stumble onto correct (or 'better than you or me' in any case) move sequences in their games - which I think we can all agree is absurd.

[RobertJasiek]

Bantari, you have not understood the paper! It is not about playing well, but is about always correctly distinguishing external from internal ko.

Ing overlooked a counter-example creating a conflict between his rules' concepts of balancing breaths and identifying unreal external breaths. http://home.snafu.de/jasiek/e2.html "A problem of definition".

A professional with whom I talked about WMSG Rules for about an hour told me that he did not understand in general the distinction between external and internal ko. So he explained his partial case by case understanding as far as he could.

The professional routinely giving almost word for word the same official WMSG Rules lectures (I listened to all of them at their rules of play parts) sticked closely to his prepared paper to ensure saying only what he understood. One could see very clearly how unhappy he was in his position having to explain exceptional rules he, as a known lover of exceptions-free rules of play, would rather have preferred not to be used at all.

Of other professionals, I have heard or read essentially nothing on the topic.

[Bantari]

Bantari, you have not understood the paper! It is not about playing well, but is about always correctly distinguishing external from internal ko.

I am not talking about a specific paper. I am talking about your claims that:
1. You describe something as 'new' and as contributing to playing strength, and
2. You validate the fact that its 'new' with the fact that a pro did not explain something to you during some event or other a few years back.

In particular, I take exception with your (implied) statement that pros do not understand something because they do not take trouble explain it to you.

If you did not make any of the above claims, please forgive, I misunderstood.

PS>
For clarification, I have indeed not understood the paper because I have not read it.
Regardless of which paper you are talking about.

[uPWarrior]

Bantari: I don't agree with you - contributing to playing strength isn't necessary. One simple example: one could discover (through mathematical arguments) the required komi to make a 19x19 game fair and that would not impact the playing strength of any player in the world. In the same way, Robert contributions could be new yet not particularly relevant to the playing strength. That's fine.

On the other hand, they do have to be new - i.e. nobody was able to understand the concept clearly before it was fully described. Isn't increasing the understanding that we have of particular phenomenons the goal of research after all? Robert seems to believe that professionals in general do not clearly understand these concepts. I'm skeptical but I guess that nobody can really tell.


This is akin to the problem of having an high school student doing research in Mathematics. Without understanding what the community of experts knows, it is impossible to be sure whether he is contributing to science or not. That is why peer review is important and unfortunately I think that the best peers are not only go researchers but also top go players.

[Samura]

As I understand, Robert is trying to classify things that arise from the rules and that are not self-evident. Despite the simple rules, Go looks to have a lot of emergent phenomena, and they obviously can be analysed mathematically/computationally.

I have a lot of interest in combinatorial and mathematical studies of Go, like endgames and kos. I sincerely don't understand what is all this fuss with someone trying to classify exhaustively this phenomena. I have the impression that this forum is missing Jasiek's point.