Your network blocks the Lichess assets!

lichess.org
Donate

Science of Chess: Subliminal Chess in the Expert Mind?

@dboing said in #10:

So I come from a more mathematical take on the nature of patterns, than from linguistic notions of chunks. Chunks as words and sequence chunks, work well with strings. But with spatial patterns, as in pattern recognition, with a well defined domain and perhaps ways to statistically extract "patterns", I wonder what would the relation be between template and pattern.

What makes the first figure you showed, pointing to, or containing either a "chunk", a "pattern" or a "template".
And what exactly on the board or generalized from that example could we use if trying to make this an automactic chess board information function.. That the chunk, pattern or template is present or not (or probably present or not....).

It seems to me that there coudl be systematic abstracting idea in there. But I do not get it. It seems that sometimes one puts the new words before the ideas, and hope the ideas will coalesnce if enough people reproduce the words in often enough similar context.

If I read that long article (which I will eventually) would I have a better discerning view between those 3 words?

I'm not distinguishing between those terms here, but that's just me. I don't make a distinction between a pattern being something mathematical and a chunk something linguistic, mostly because I don't see a reason for doing so. Again, this is just my perspective, but it seems to me that all of these words refer to a description of a position that is stored in memory, possibly with some description of next moves that accompany it.

I don't know what you mean by an automatic chess board information function, though, so I don't know how to respond to that.

Is "that long article" this blog post or something else? In any case, I'm not trying to define chunks in any formal way here, and the article I'm commenting on doesn't really do so either.

@dboing said in #10: > So I come from a more mathematical take on the nature of patterns, than from linguistic notions of chunks. Chunks as words and sequence chunks, work well with strings. But with spatial patterns, as in pattern recognition, with a well defined domain and perhaps ways to statistically extract "patterns", I wonder what would the relation be between template and pattern. > > What makes the first figure you showed, pointing to, or containing either a "chunk", a "pattern" or a "template". > And what exactly on the board or generalized from that example could we use if trying to make this an automactic chess board information function.. That the chunk, pattern or template is present or not (or probably present or not....). > > It seems to me that there coudl be systematic abstracting idea in there. But I do not get it. It seems that sometimes one puts the new words before the ideas, and hope the ideas will coalesnce if enough people reproduce the words in often enough similar context. > > If I read that long article (which I will eventually) would I have a better discerning view between those 3 words? I'm not distinguishing between those terms here, but that's just me. I don't make a distinction between a pattern being something mathematical and a chunk something linguistic, mostly because I don't see a reason for doing so. Again, this is just my perspective, but it seems to me that all of these words refer to a description of a position that is stored in memory, possibly with some description of next moves that accompany it. I don't know what you mean by an automatic chess board information function, though, so I don't know how to respond to that. Is "that long article" this blog post or something else? In any case, I'm not trying to define chunks in any formal way here, and the article I'm commenting on doesn't really do so either.

'Chunks' derive from information theory. George Miller was inspired by Claude Shannon, who predates Cognitive Psychology and Linguistics. Considered part of the 'Cognitive Revolution' of the 1950s, turning away from Behaviorism. Generally credited to Noam Chomsky, George Miller, Alan Newell and Herbert Simon after hearing Claude Shannon espouse his theory at a conference at MIT in 1956. For students of the history of Cognitive Psychology and Computer Science, Shannon's information theory is one of the prime theories creating both fields. Chunking comes from a unique period of history where computer science was a fresh field (still using vacuum tubes), and theories of the mind, where parallel to new theories of computer programing, where how the mind work was modeled in comparison to how the computer was programmed.

So a 'chunk' remains ill defined, a chunk is basically a unit of information, that can be defined mathematically, but the mathematical definition loses meaning when transferred to Cognitive Psychology. A 'Template' is also just a larger chunk, unit of information, that allows some variation. Chunking theory deserves more research, but is basically just a remnant of forgotten scholars, namely Herbert Simon, Allen Newell and George Miller, who have some students and followers, but not many, Fernand Gobet maybe most notably.

Related to chess, is that Simon and Newell (not to write a whole history book with the founders of computer science like Turing), where all club chess players, and chess played a disproportionate part of their research. That is not the case anymore, and besides for some memory researchers, most PhDs in the field have never heard of chunks, it is not an area of research. Even most chess players never heard of chunks, nor familiar with the decades of research on 'chunking' and chess. There are many reasons for this, like a Kuhnian paradigm shift, the direction of research in the 1980s & 1990s went away from chunks, and only someone who has read the historic literature is familiar.

So props for bringing chunking theory and these studies to greater attention, a theory that deserves more research, and a clearer definition.

'Chunks' derive from information theory. George Miller was inspired by Claude Shannon, who predates Cognitive Psychology and Linguistics. Considered part of the 'Cognitive Revolution' of the 1950s, turning away from Behaviorism. Generally credited to Noam Chomsky, George Miller, Alan Newell and Herbert Simon after hearing Claude Shannon espouse his theory at a conference at MIT in 1956. For students of the history of Cognitive Psychology and Computer Science, Shannon's information theory is one of the prime theories creating both fields. Chunking comes from a unique period of history where computer science was a fresh field (still using vacuum tubes), and theories of the mind, where parallel to new theories of computer programing, where how the mind work was modeled in comparison to how the computer was programmed. So a 'chunk' remains ill defined, a chunk is basically a unit of information, that can be defined mathematically, but the mathematical definition loses meaning when transferred to Cognitive Psychology. A 'Template' is also just a larger chunk, unit of information, that allows some variation. Chunking theory deserves more research, but is basically just a remnant of forgotten scholars, namely Herbert Simon, Allen Newell and George Miller, who have some students and followers, but not many, Fernand Gobet maybe most notably. Related to chess, is that Simon and Newell (not to write a whole history book with the founders of computer science like Turing), where all club chess players, and chess played a disproportionate part of their research. That is not the case anymore, and besides for some memory researchers, most PhDs in the field have never heard of chunks, it is not an area of research. Even most chess players never heard of chunks, nor familiar with the decades of research on 'chunking' and chess. There are many reasons for this, like a Kuhnian paradigm shift, the direction of research in the 1980s & 1990s went away from chunks, and only someone who has read the historic literature is familiar. So props for bringing chunking theory and these studies to greater attention, a theory that deserves more research, and a clearer definition.

@dboing All these terms are loosely defined, and from an era where the only data on the mind was psychometric, and the goal was to create a computer model for how the mind works. Some of these programs still exist like ACT-R.

Generally a 'Chunk' is a fixed pieces of information, while a template is a large unit of information that has some 'slots' that can be filled with unique pieces of information. So a chess chunk is a fixed pattern, and also explains errors, cognitive bias,..., as people falsely use a chunk that does not fit the situation. So chunking was updated with template theory, where chunks could also be large structures that allowed some variation.
But 'chunks' are basically just heuristics that cannot be mathematically defined, not psychologically tested, besides trying to correspond to psychometric data, or programmed into a computer to simulate the human mind. Chunking survives because it is a plausible method that corresponds to neuronal activity (check the original perceptron project). In the 1950s research into neurons and computer circuits were new, and researchers of the time thought neurons would work like logic circuits, similar to a microprocessor. Did not work out that was, so at best you have neural correlates of consciousness.

If you have ever taken a course in speed reading, chunking theory in speed reading is more straight forward and intuitive than in chess, and the path to being able to speed read, can be understood as the capacity to chunk.

But, no, 'chunk', 'pattern' & 'template' are not precise scientific terms capable of being reduced to mathematical equations, although they might be able to be defined in logical terminology. So 'chunking theory' did not survive the 'second cognitive revolution' of the 1980s and 90s, where neuro science took over. So you have the remnants of cognitive psychology based on psychometric data like theories of memory (working, short term, long term,...) verse the modern emphasis on neuro psychology.

@dboing All these terms are loosely defined, and from an era where the only data on the mind was psychometric, and the goal was to create a computer model for how the mind works. Some of these programs still exist like ACT-R. Generally a 'Chunk' is a fixed pieces of information, while a template is a large unit of information that has some 'slots' that can be filled with unique pieces of information. So a chess chunk is a fixed pattern, and also explains errors, cognitive bias,..., as people falsely use a chunk that does not fit the situation. So chunking was updated with template theory, where chunks could also be large structures that allowed some variation. But 'chunks' are basically just heuristics that cannot be mathematically defined, not psychologically tested, besides trying to correspond to psychometric data, or programmed into a computer to simulate the human mind. Chunking survives because it is a plausible method that corresponds to neuronal activity (check the original perceptron project). In the 1950s research into neurons and computer circuits were new, and researchers of the time thought neurons would work like logic circuits, similar to a microprocessor. Did not work out that was, so at best you have neural correlates of consciousness. If you have ever taken a course in speed reading, chunking theory in speed reading is more straight forward and intuitive than in chess, and the path to being able to speed read, can be understood as the capacity to chunk. But, no, 'chunk', 'pattern' & 'template' are not precise scientific terms capable of being reduced to mathematical equations, although they might be able to be defined in logical terminology. So 'chunking theory' did not survive the 'second cognitive revolution' of the 1980s and 90s, where neuro science took over. So you have the remnants of cognitive psychology based on psychometric data like theories of memory (working, short term, long term,...) verse the modern emphasis on neuro psychology.

@DIAChessClubStudies said in #12:

So a 'chunk' remains ill defined, a chunk is basically a unit of information, that can be defined mathematically, but the mathematical definition loses meaning when transferred to Cognitive Psychology.

I think this is especially important, especially when thinking about ideas from information theory and chess concepts that have some resemblance. If you're talking about Shannon and stuff like data compression, then "chunks" can mean something like the vocabulary you use to encode a signal efficiently (e.g. the Morse code alphabet, which is informed by letter frequencies). The trouble is that there's not to my knowledge any reason to think that the units people remember and store while gaining chess experience have a clear relationship to those principles. I'd suggest that there's no reason they should, either.

Broadly speaking though, I think plenty of my colleagues both know about and think about the idea of organizing information into larger units. In vision science (my own area of expertise), we think about this all the time in various ways: Perceptual organization, holistic processing, and a bunch of other topics in the field are all concerned with the nature of chunking visual information and principles that govern when and how the visual system does it. Likewise, I'd be surprised to find out that my colleagues who study memory don't use these ideas - probably not in the context of chess given that cognitive chess research is a bit of a niche interest, but I think they'd see the use of "chunking" there as a special case of a more general principle they're comfortable with.

Related to chess, is that Simon and Newell (not to write a whole history book with the founders of computer science like Turing), where all club chess players, and chess played a disproportionate part of their research. That is not the case anymore, and besides for some memory researchers, most PhDs in the field have never heard of chunks, it is not an area of research. Even most chess players never heard of chunks, nor familiar with the decades of research on 'chunking' and chess. There are many reasons for this, like a Kuhnian paradigm shift, the direction of research in the 1980s & 1990s went away from chunks, and only someone who has read the historic literature is familiar.

So props for bringing chunking theory and these studies to greater attention, a theory that deserves more research, and a clearer definition.

@DIAChessClubStudies said in #12: > > So a 'chunk' remains ill defined, a chunk is basically a unit of information, that can be defined mathematically, but the mathematical definition loses meaning when transferred to Cognitive Psychology. I think this is especially important, especially when thinking about ideas from information theory and chess concepts that have some resemblance. If you're talking about Shannon and stuff like data compression, then "chunks" can mean something like the vocabulary you use to encode a signal efficiently (e.g. the Morse code alphabet, which is informed by letter frequencies). The trouble is that there's not to my knowledge any reason to think that the units people remember and store while gaining chess experience have a clear relationship to those principles. I'd suggest that there's no reason they should, either. Broadly speaking though, I think plenty of my colleagues both know about and think about the idea of organizing information into larger units. In vision science (my own area of expertise), we think about this all the time in various ways: Perceptual organization, holistic processing, and a bunch of other topics in the field are all concerned with the nature of chunking visual information and principles that govern when and how the visual system does it. Likewise, I'd be surprised to find out that my colleagues who study memory don't use these ideas - probably not in the context of chess given that cognitive chess research is a bit of a niche interest, but I think they'd see the use of "chunking" there as a special case of a more general principle they're comfortable with. > > Related to chess, is that Simon and Newell (not to write a whole history book with the founders of computer science like Turing), where all club chess players, and chess played a disproportionate part of their research. That is not the case anymore, and besides for some memory researchers, most PhDs in the field have never heard of chunks, it is not an area of research. Even most chess players never heard of chunks, nor familiar with the decades of research on 'chunking' and chess. There are many reasons for this, like a Kuhnian paradigm shift, the direction of research in the 1980s & 1990s went away from chunks, and only someone who has read the historic literature is familiar. > > So props for bringing chunking theory and these studies to greater attention, a theory that deserves more research, and a clearer definition.

@DIAChessClubStudies said in #13:
So 'chunking theory' did not survive the 'second cognitive revolution' of the 1980s and 90s, where neuro science took over. So you have the remnants of cognitive psychology based on psychometric data like theories of memory (working, short term, long term,...) verse the modern emphasis on neuro psychology.

Hang on, though - cognitive science hasn't been taken over yet! Sure, we use tools from cog. neuro. to inform our thinking about cognitive processes, but we also still develop models of mechanisms for memory, recognition, etc. at different levels of description.

@DIAChessClubStudies said in #13: So 'chunking theory' did not survive the 'second cognitive revolution' of the 1980s and 90s, where neuro science took over. So you have the remnants of cognitive psychology based on psychometric data like theories of memory (working, short term, long term,...) verse the modern emphasis on neuro psychology. Hang on, though - cognitive science hasn't been taken over yet! Sure, we use tools from cog. neuro. to inform our thinking about cognitive processes, but we also still develop models of mechanisms for memory, recognition, etc. at different levels of description.

Thanks @NDpatzer , could be chunking can be revived. And chunking is likely the method the mind uses to recognize shapes, or spatial awareness. Template theory is also part of spatial awareness, memory for geography. Also part of famous memory techniques like the method of loci. I would say the main theories of mind close to chunking are 'predictive processing', 'active inference (Karl Friston). Just a mismatch of research where chunking theory was forgotten, besides for the relationship to short term memory. I once on stream asked Karl Friston about chunking and Fernand Gobet a few years ago, and Friston said he was not familiar (even though they are both in London).

I would like to see a complete theory of mind, built up from chunking theory, which ActR, Soar, CHREST, Clarion,..., are all attempts at, many directly related to chess studies. University of Michigan, where I attended, Cognitive Architectures is part of the Computer Science department, not psychology. 'Chunking' in general has largely been relegated to computer science, with just remnants still found in cognitive science literature. In a similar way that consciousness studies is part of philosophy or theology, and the few scientific centers to studies consciousness are in medical schools for anesthesiology. Consciousness is not a serious subject of scientific research.

A more multi-disciplinary approach is needed. I only am familiar with chunking theory, because I am a chess player, when I read 'Mind Over Machine' by Dryfus & Dryfus 35 years ago, and became fascinated by the subject. The 1st Cognitive Revolution was based on combining for the first time psychology, linguistics, computer science, anthropology, neuroscience, and philosophy.. Now the list is much longer, and would need cross collaboration that does not exist.

I have followed the 'Science of Consciousness Conference' for over five years, and don't think I have heard once 'chunking theory'. Think the closest theory to try to incorporate chunking theory would be Karl Friston's Free Energy Principle / Active Inference, based on Bayesian inference and Markov blankets.

Look forward to your future research.

Thanks @NDpatzer , could be chunking can be revived. And chunking is likely the method the mind uses to recognize shapes, or spatial awareness. Template theory is also part of spatial awareness, memory for geography. Also part of famous memory techniques like the method of loci. I would say the main theories of mind close to chunking are 'predictive processing', 'active inference (Karl Friston). Just a mismatch of research where chunking theory was forgotten, besides for the relationship to short term memory. I once on stream asked Karl Friston about chunking and Fernand Gobet a few years ago, and Friston said he was not familiar (even though they are both in London). I would like to see a complete theory of mind, built up from chunking theory, which ActR, Soar, CHREST, Clarion,..., are all attempts at, many directly related to chess studies. University of Michigan, where I attended, Cognitive Architectures is part of the Computer Science department, not psychology. 'Chunking' in general has largely been relegated to computer science, with just remnants still found in cognitive science literature. In a similar way that consciousness studies is part of philosophy or theology, and the few scientific centers to studies consciousness are in medical schools for anesthesiology. Consciousness is not a serious subject of scientific research. A more multi-disciplinary approach is needed. I only am familiar with chunking theory, because I am a chess player, when I read 'Mind Over Machine' by Dryfus & Dryfus 35 years ago, and became fascinated by the subject. The 1st Cognitive Revolution was based on combining for the first time psychology, linguistics, computer science, anthropology, neuroscience, and philosophy.. Now the list is much longer, and would need cross collaboration that does not exist. I have followed the 'Science of Consciousness Conference' for over five years, and don't think I have heard once 'chunking theory'. Think the closest theory to try to incorporate chunking theory would be Karl Friston's Free Energy Principle / Active Inference, based on Bayesian inference and Markov blankets. Look forward to your future research.

I see a parallel with verbal linguistic having been how De groot might have design the protocols. Could it be that the experimental protocols themselve were having natural chunks and templates being used frequently, and they would have become natural putative objects of their theories of the mind. Not a very connectionist point of view, but a most verbal seqeunce point of view. I fing the connection with CPU paradigm for thinking to be also of the verbal/linguistic framework.

It might be that the cognitive psychology might have shifter from that paradirm to the a more paralle processing or distributed notion of patterns in the brain substrate of the theories of the mind. I use words as if I knew their meaning but I am approximating from crumbs I might have read here and there. I appreciate the historical connections thanks.

I was talking about the blog. Thanks both of you for your take on how defined chunks and template might be.

I would add, that information might be in the eyes of the beholder.... as my signature obligatory derailing thought.

The domain one restricts oneself to in defining information may not be the full space of the problem that needs a theory of the mind, as in chess. One may only look at the move sequences as the information domain, and while the position information could be systematically and recursively derived, not deriving it and only looking at the move sequences, while in absolute could derive later the full position information seqeunce, that would still be a different domain of information definition. As information is about communication from the board signals or the written move sequence to the mind, that mind is not platonic, and will use the domain of information, if given only the sequence, it might never learn the position and might think that chess is only about move sequences..

I think that the neural networks statistical approaches to certain type of learning, while they are being construed through sequential operation a la CPU, are better thought of as configuration of activities rather than sequence of charcters being chunked.

Back to the figure. What elements of that position would be the "chunk" exactly. (by math and automatic I mean that we can take the full position information for many positions and try to figure out which would the chunk/template/pattern beholder consider of the same such class and which not, and if we allow the big stupid brain to have its continuum evaluation abilities the probailities that the chunk is there. What set of positions would fit in the class of the chunk for which the figure example is a representant. I could mutate the position and take pieces of the board, when will it stop being of the chunk or template or pattern.. There is the board set of position, and then there is the internal neurocognitive models that are nibbling away from ad-hoc models of cognition and trying to bridge to the brain based substrate of psychology. Did I mistreat any words?

I am not really knowleageable here. I should read. but I think some questions are rarely asked, and things can keep floating above such omission. I find in chess that there seem to be a gap that did not follow that paradigm shift. But I might not have read the recent science about it. and NDpatzer has been my window about what I might have missed in the past 15 years, I might not have been looking at all. About that neuroscience reaching psychology.

I do not vouch for the argumentative nature of above. These are threads and are more questions that arguments. They might not even be consistent which each other.

I see a parallel with verbal linguistic having been how De groot might have design the protocols. Could it be that the experimental protocols themselve were having natural chunks and templates being used frequently, and they would have become natural putative objects of their theories of the mind. Not a very connectionist point of view, but a most verbal seqeunce point of view. I fing the connection with CPU paradigm for thinking to be also of the verbal/linguistic framework. It might be that the cognitive psychology might have shifter from that paradirm to the a more paralle processing or distributed notion of patterns in the brain substrate of the theories of the mind. I use words as if I knew their meaning but I am approximating from crumbs I might have read here and there. I appreciate the historical connections thanks. I was talking about the blog. Thanks both of you for your take on how defined chunks and template might be. I would add, that information might be in the eyes of the beholder.... as my signature obligatory derailing thought. The domain one restricts oneself to in defining information may not be the full space of the problem that needs a theory of the mind, as in chess. One may only look at the move sequences as the information domain, and while the position information could be systematically and recursively derived, not deriving it and only looking at the move sequences, while in absolute could derive later the full position information seqeunce, that would still be a different domain of information definition. As information is about communication from the board signals or the written move sequence to the mind, that mind is not platonic, and will use the domain of information, if given only the sequence, it might never learn the position and might think that chess is only about move sequences.. I think that the neural networks statistical approaches to certain type of learning, while they are being construed through sequential operation a la CPU, are better thought of as configuration of activities rather than sequence of charcters being chunked. Back to the figure. What elements of that position would be the "chunk" exactly. (by math and automatic I mean that we can take the full position information for many positions and try to figure out which would the chunk/template/pattern beholder consider of the same such class and which not, and if we allow the big stupid brain to have its continuum evaluation abilities the probailities that the chunk is there. What set of positions would fit in the class of the chunk for which the figure example is a representant. I could mutate the position and take pieces of the board, when will it stop being of the chunk or template or pattern.. There is the board set of position, and then there is the internal neurocognitive models that are nibbling away from ad-hoc models of cognition and trying to bridge to the brain based substrate of psychology. Did I mistreat any words? I am not really knowleageable here. I should read. but I think some questions are rarely asked, and things can keep floating above such omission. I find in chess that there seem to be a gap that did not follow that paradigm shift. But I might not have read the recent science about it. and NDpatzer has been my window about what I might have missed in the past 15 years, I might not have been looking at all. About that neuroscience reaching psychology. I do not vouch for the argumentative nature of above. These are threads and are more questions that arguments. They might not even be consistent which each other.

@DIAChessClubStudies

I found your first post very interesting about the scientific trap that mastery might entail. I also think it is related to why theories of learning in chess might have not evolved a lot.

I do not agree though with the chunking theory that has not tried to dissect what those chunks might be at least on the board with respect to positions sets. The mathematical tools to adress such questions of position set classes, wihtout going enumerative about all the positions in such set, might be needed to start gonig beyond a counting of the number of chunks.

There is some nesting of the board logic, with concept as multipurpose moves the fork extension to many scales of chess. for example. I think studying more neurologically inspired internal models, say, as artificial neural nets might be machine models, a notch more neurologically based than working memory models, or other placeholder models that explain the empirical psychometrics consistently, but do not have a reductionist derivation to keep digging into.

But, still you posts have been interesting read giving some historical perspective. I also like the epistemological notions. I am not sure though that one should consider the past ideas are revolved, through a sequence of paradigm shifts, it might be nested paradigm chages. getting late. going to sleep. For example reinforcement learning is old but the biology of it is across all animal brains (at least the mammals are all functionning like that). And then combine that with connectionixt power of function expression, and you get a new theory of learning, that might not get stuck in trying to imitate masters, as a way to bypass certain logic of the board to be experinced, in all sort of plebian chess corners..

But I think you are rigth in there has been too much focus in the study of how the masters think, and not enough about how can learn from any point A to any point B. If that was a possibility in your first post.

@DIAChessClubStudies I found your first post very interesting about the scientific trap that mastery might entail. I also think it is related to why theories of learning in chess might have not evolved a lot. I do not agree though with the chunking theory that has not tried to dissect what those chunks might be at least on the board with respect to positions sets. The mathematical tools to adress such questions of position set classes, wihtout going enumerative about all the positions in such set, might be needed to start gonig beyond a counting of the number of chunks. There is some nesting of the board logic, with concept as multipurpose moves the fork extension to many scales of chess. for example. I think studying more neurologically inspired internal models, say, as artificial neural nets might be machine models, a notch more neurologically based than working memory models, or other placeholder models that explain the empirical psychometrics consistently, but do not have a reductionist derivation to keep digging into. But, still you posts have been interesting read giving some historical perspective. I also like the epistemological notions. I am not sure though that one should consider the past ideas are revolved, through a sequence of paradigm shifts, it might be nested paradigm chages. getting late. going to sleep. For example reinforcement learning is old but the biology of it is across all animal brains (at least the mammals are all functionning like that). And then combine that with connectionixt power of function expression, and you get a new theory of learning, that might not get stuck in trying to imitate masters, as a way to bypass certain logic of the board to be experinced, in all sort of plebian chess corners.. But I think you are rigth in there has been too much focus in the study of how the masters think, and not enough about how can learn from any point A to any point B. If that was a possibility in your first post.

Thanks @dboing , yes that was my point, from Michelene Chi, the original participant in the Chase & Simon study 1973 at Carnegie Melon, and the paradigm shift that occurred in the 1990s after Anders Ericsson. I plan to continue the series going through the various models of expertise, all getting from novice to master, which chess has fallen out from the studies due to difficulty in defining task performance other than rating and winning games against others in a rating level.
There have been attempts to measure chess skill other than rating, by tests.
I have also been interested in organizing youth chess festivals that focused less on competition, and more on puzzle solving, with little success.

Thanks @dboing , yes that was my point, from Michelene Chi, the original participant in the Chase & Simon study 1973 at Carnegie Melon, and the paradigm shift that occurred in the 1990s after Anders Ericsson. I plan to continue the series going through the various models of expertise, all getting from novice to master, which chess has fallen out from the studies due to difficulty in defining task performance other than rating and winning games against others in a rating level. There have been attempts to measure chess skill other than rating, by tests. I have also been interested in organizing youth chess festivals that focused less on competition, and more on puzzle solving, with little success.

@DIAChessClubStudies said in #19:

There have been attempts to measure chess skill other than rating, by tests.

Well, @NDpatzer might have given us a cognitive activities' series of blogs, not necessarily about rating, although it would be a high density sampling focus in the data, as mastery is the mystery. Contrasting groups of 2 classes, then 3 classes.

I am curious about such efforts (measured chess skills, tests). Perhaps those could unlock that few-bins point of view on the cognitive study side, by providing board information tangibles that can also be connected to many bins (even unfathomably many) of, and along, the chess learning trajectories.

Maybes. If I made some misconceived statement above, I would like to be informed, well anyway I would.

Circumstantial and tangential (self stuff)
Also, I see that my syntax ordering is messed up. Sorry. Kind of in some middle messy ground between french and english. But my language tool corrector got it, it put an apostrophe behind "activities", so it might not be so bad.

@DIAChessClubStudies said in #19: > There have been attempts to measure chess skill other than rating, by tests. Well, @NDpatzer might have given us a cognitive activities' series of blogs, not necessarily about rating, although it would be a high density sampling focus in the data, as mastery is the mystery. Contrasting groups of 2 classes, then 3 classes. I am curious about such efforts (measured chess skills, tests). Perhaps those could unlock that few-bins point of view on the cognitive study side, by providing board information tangibles that can also be connected to many bins (even unfathomably many) of, and along, the chess learning trajectories. Maybes. If I made some misconceived statement above, I would like to be informed, well anyway I would. Circumstantial and tangential (self stuff) Also, I see that my syntax ordering is messed up. Sorry. Kind of in some middle messy ground between french and english. But my language tool corrector got it, it put an apostrophe behind "activities", so it might not be so bad.