Robert Jasiek's Go Theory Research

[RobertJasiek]

Bantari, it is commonly agreed for research that something is called "new" or "invented" if before nobody has stated, explained or especially published it. Of course, theoretically there could exist somebody having solved the problem for himself but keeping this fact for himself forever. Did Fermat really prove the Great Fermat himself?:)

Samura, at times, I classify things arising from given rules. At other times, I allow every input ruleset. The former applies for the Supplementary Ko Rule, the latter applies to the external / internal ko distinction.

[Kirby]

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

I see. I think this is an important distinction. Many go players are interested in improving their ability to play well, and probably assume go instruction to be about... playing well.

Because of this, it can seem absurd that an amateur player claims to know more than pros.

But presumably this claim is not about "knowing the ideas and having the mental capacity to win more games", but rather about knowing particular types of classifications like, for example, the meaning of "external" or "internal" ko.

Since pros may not even care about some of these topics, I don't find this hard to believe. I'm sure that Robert has identified certain perspectives of the game that others haven't, yet.

After all, the game is a deep one, so there're many things to observe.

Still, when it comes to knowledge and strength that leads to winning games, I'm confident that pros are the best resource in this area (holding teaching ability constant).

(I guess I posted again :-p)

[RobertJasiek]

In reply to Bantari in viewtopic.php?p=129658#p129658

go playing competition is close to what you describe: (usually) a shallow knowledge of the whole beats the narrow but deep(er) knowledge of a particular aspect. However, systematic go theory research contribution is close to how you describe science: a large number of tiny steps, each made by a different researcher, and each ultimately contributing to the overall knowledge.

The phrase 'go theory' has - at least for decades - been used with different meanings from 'informal common knowledge' via 'practical example variations exploration study' to 'methodical, formal research'. You say that I would confuse people by using the phrase in these varying meanings. It is more correct to say that everybody has confused everybody else by continued usage of the phrase with very different meanings. Since the late 90s, I have contributed to the clarification that the phrase does have those different meanings in different contexts. You can criticise me for not spelling out the currently used meaning of the phrase each time I used it, but it is unfair to blame me for confusing different usage, who I have been one of those contributing the most to clarifying the fact that the phrase is used with varying meanings.

Dogmatic about axioms?! I presume axioms when necessary, but if others have other axioms, they can bring forward them, so that they can be discussed and compared.

[RobertJasiek]

viewtopic.php?p=148608#p148608

When RJ defines something, it is a formulation of a meaning. It's not theory.

When I define something, it depends on the purposes of creating the definition whether the definition is

- only a formulation of a meaning or
- both part of a theory and a formulation of the correct meaning for the theory.

Many definitions belong to the latter, if "theory" is used in an informal sense. Still quite a few definitions belong to the latter for a formal sense.

In fact, a few important theories of mine consist of (axioms and) definitions! (Except that such theories are also verified.) It is the mathematical approach to theories: Start from axioms, proceed by definitions, possibly use theorems. The interesting thing about go rules theory and go terms theory is that they can often be constructed only by axioms and definitions, while theorems are not even needed.

[pwaldron]

Has any of this "research" been cited by others in peer reviewed journals or conference proceedings? Has it been cited in any non-peer reviewed publications? If so, please point us to a few such papers.

[uPWarrior]

Robert, do you have any kind of background on scientific research, or a scientific education in general?
I would guess so, given your interest in this topic, but your last posts lead me to believe that you don't have such experience.

[RobertJasiek]

I do not know of citations in peer reviewed journals or conference proceedings. IIRC, a few such papers by others mentioned rules papers in the literature references, but I do not collect such, so I cannot tell you by heart which. What is a "non-peer reviewed publication"?

Why do you ask? Your questions sound as if you need third persons to judge about correctness or quality of my research results. Can't you judge by yourself?

Peer reviews? I motivated Chris Dams to prove my conjecture, then I peer reviewed his proof. That's how small the research community is for go theory research.

I find my own research so very useful that I apply it frequently and that it has accelerated my continued research. This applies to both theory in research papers and theory in books applicable for go players.

Examples of accelerated research:

Japanese 2003 Rules -> paper Types of Basic Kos -> paper Ko

New Ko Rules -> fixed-ko rule -> paper Types of Basic Kos -> paper Ko

...->...-> Japanese 2003 Rules ->...->...-> Simplified Japanese Rules

mobility difference -> influence stone difference -> territory and influence ratio -> value model for josekis

miai value of early corner stones -> value model for josekis

[RobertJasiek]

uPWarrior, I studied mathematics + theoretical informatics for 7 years at university. I needed 1.5 years to get the Vordiplom, but then studied and played too much go, go rules theory and go theory, so that I never made the Diplom.

[Boidhre]

A non-peer reviewed publication is a self published piece of original research that has not been vetted or screened by experts in the field. It varies from field to field how much respect such is held in.

[RobertJasiek]

Another mentioning in a literature list:

http://lie.math.brocku.ca/twolf/papers/semeai.pdf

IIRC, this paper did not appear in a journal, but I am not sure about that. Its acknowledgement lists a few of the usual suspects:)

[RobertJasiek]

pwaldron, your questions appear to overlook the problem inherent in go theory research. So far, there are mainly these mathematical go theory research fields (except for statistics):

1) computer go theory (popular among paid scientists, students and programmers)

2) combinatorial game theory and related fields (popular among paid scientists, partly involving students)

3) semeai theory (a few researchers)

4) rules theory (quite a few low level approaches to writing down rules and their terms at all, advanced studies by me, a bit above low level studies by a few other researchers)

5) definition-derived theory for intermediate or advanced terms and applied go theory (mainly by me)

IMO, currently there are these reasons why my (5) research is still scarcely cited in peer-reviewed media:
- (5) is not popular among researchers yet.
- In particular, (5) is not popular among paid scientists with easy access to peer-reviewed media.
- Paid scientists do not always have a go playing strength or knowledge enabling them to do research in (5).
- Maybe paid scientists have not always understood yet that they can circumvent their problem not only in (1) or (2), but also in (5), because bottom-up research starts from axioms.
- Maybe part of the paid scientists want, or are expected, to show a variety of advanced mathematical techniques, so that an almost only definitions approach is viewed below their level of education. (IMO, they would overlook the elegance of this approach.)
- Maybe my preference for semi-formal language instead of symbolic mathematical annotation lets part of the paid scientists create a prejudice of uninteresting contents.

[RBerenguel]

(a) Maybe part of the paid scientists want, or are expected, to show a variety of advanced mathematical techniques, so that an almost only definitions approach is viewed below their level of education. (IMO, they would overlook the elegance of this approach.)
(b) Maybe my preference for semi-formal language instead of symbolic mathematical annotation lets part of the paid scientists create a prejudice of uninteresting contents.

I've edited the quote to mark the reasons as (a) and (b) (I have ignored the previous 4.)

Answer to (a): A "paid scientist" (a term which I guess refers to a researcher, at least in this context) does not want or is expected to show "a variety of advanced mathematical techniques." At least in the field of mathematics, which is where I come from, researchers are expected to research and publish (the old motto "publish or perish.") The thing is, the problem has to be either "en vogue," interesting in its own right or at the very least, publishable (which is quite the shady object.) But in any case, I don't understand exactly what an "almost only definitions approach" stands for. In axiomatic constructions you usually try to derive something out of the axioms. But what is a "definitions approach"? Defining objects without a goal is rather pointless, as far as research goes. It may be useful to later "explain" something, but explaining is a reinterpretation of something according to the definitions. I can define A as the set of red ripe apples, B an ovoid shaped object used in hen reproduction and so on. Unless I explain the algorithm of an apple pie, I'm shaving yaks.

Answer to (b): If the result is interesting you could use ideograms like you were playing Pictionary and people would care. I've read plenty of proofs which are nonsense, stupid or uninteresting (I can easily put my own research in the not very interesting basket) and the formality of it doesn't help or hinder who is reading. Like in online media, "content is king."

[pwaldron]

IMO, currently there are these reasons why my (5) research is still scarcely cited in peer-reviewed media:

Since one opinion is as good as another, I'll put forward another reasons why your work isn't cited: your work hasn't made any advancement that is worth citing or building upon.

[RobertJasiek]

I don't understand exactly what an "almost only definitions approach" stands for.

Few or no theorems, proofs, use of existing theorems, algorithms, test programs, mathematical techniques - but just axioms and definitions.

Defining objects without a goal

Right. Goals are also needed. Such as "the definition of ko identifies all known ko shapes and does not confuse any other shape with a ko".

"content is king."

AFA I am concerned, I agree:)

[RobertJasiek]

your work hasn't made any advancement that is worth citing or building upon.

This is wrong, because I have cited and built upon my own work. It is very wrong, because advancements have been created and explained. It is extremely wrong, because in some cases my results are extraordinary advancements.

For example, to repeat the most obvious, Ing's idea "a ko is given due to repetition or recycling" (the [informal] research predecessor to (my) later work) was proven in 1997(?) by me to be dubious by means of the proposition that each stone in each position can be recreated in a cycle and replaced in 2010 by me by means of the definition of ko, which is applicable for all stones(!) in all positions(!) under every ruleset(!), so that NOT all stones in all positions are ko stones. This is an extraordinary advancement, e.g., because of the general applicability and the first meaningful definition distinguishing (for all kos) ko stones from non-ko stones.

Can you find any second researcher providing rules-dependent findings invariant under the choice of the [input] ruleset? Many papers by other researchers are rules-dependent, but show results for only one particular ruleset.