- Blind mode tutorial
lichess.org
Donate

The Axiom System - Part 4: Justification in Chess

#121

@DailyInsanity said in #114:

Yes true, but with the important caveat that this would be done with the idea of improving one's intuition/'feeling' in such positions, as opposed to acquiring knowledge/explanations which would then be used to deduce the correct moves in one's own games (which, as I argue in this post, is not possible).

I do thank you guys for continuing to post your thoughts here :) I will have some free time coming up soon, and so I hope to be active on lichess again.

Hey Jack, how has your free time been going lately?

@DailyInsanity said in #114: > Yes true, but with the important caveat that this would be done with the idea of improving one's intuition/'feeling' in such positions, as opposed to acquiring knowledge/explanations which would then be used to deduce the correct moves in one's own games (which, as I argue in this post, is not possible). > > I do thank you guys for continuing to post your thoughts here :) I will have some free time coming up soon, and so I hope to be active on lichess again. Hey Jack, how has your free time been going lately?
#122

@data_gatherer said in #120:

only just read this. Is it possible that what you cal motif at more of the current position trigger features (those can be described, withiout being attached to action rules, often people throw baby with the bath water too eargely). Those don,t need to have names already having been called dibs in some book you might not be aware of, and therefore, you might want to find your own natural language analogies that might do the self-talk job helper to ruminate in the board foresight problem later in new positions. It helps with the slow critical thinking part. It helps separating aspects of the pattern awareness I would say. Perhaps you can even confront your board experience with other heads have their own distinct board experience trajectory, and those differnt angles might welcome your complementary view and attempt to share some more abstract idea that the very specific case where you started seeing a pattern.

back to trigger feature I would call that a static pattern. Then my impression is that the woodpecker thing is about internalizing certain short sequences of moves into the "muscle memory" given the trigger pattern.. I would call those, dynamic patterns.. Now, those woodpekcy problems might have been rigged to serve you problems without the fog of the often neglected part of the foresight in game problem, which only records the move candidates from the currrent position being grown methodically from current position. But that is bypassing the whole very fast and hard to just use words, but Kotov did mention it in one book or was it chapter dedicated to thinking models in chess.

The imagiation and desirable search aspects.. It is currently often left to brute force experience.. But I think that is where the evalution question of yours might be coming from. The woodpecker problem are rigged from the movie script writer... (ok in my home version of French, we would say. Arrangé avec le gars des vues, how exotic!).

That means you have neve being train in the goal or plan imagination and desirable possible nearby position to aim at.

there might be many patterns on a real board at the same time, and even if not seeing them, or not so many, (of the trigger type) , in real game you can't count of on the plan having been selected for you as in a battery of woodpecker problems. (or puzzles in general).

That work of imagining possible futures within your current internal representation or skill set "map" or whatever chess vision horzion?.

whel you can't imagine yet what your improved future self might on the same position if revisting it in some better experienced future.. so you are stuch with that. The problem in real game is the multiplicity of such possible imaginable nearby or less nearby goals to imagine and then find ways from current position to achieve that.

ok. you could always go full bling brute force calculation from all legal moves starting from current position. but we all have lack of patience for that.. we like to try things but can we make our thinking more visible to our own mind.

ok. I bet I have expired my welcome period.. So I think the question is great and contrary to this blog reluctant to consider chess theory, or principles as justificaoni, I think one just have to dissect them for their missing logic or context precision.

that is actually fun work. debunking rules of thumbs. by finding their exceptions and then finding the missing logic and missing conditions board information in term of feature sets, often themselves nameable trigger pattern or building block toward such patterns. The problem with the usual third wheel learning crutches, beyond their initial get in the game deeper helping hand, is that they start becoming obstacles to learning the logic behind them.

exceptions to the rule. is kind of a let down for me. I deleted a post i made somewhere, as not appro;priate there. but I gyuess while imposing my rambling here I could mention an idea, perhaps sounding out there. .but really a hunch when I think of chess and what NN might be learning about board feature patterns in their latent space (input space augmented by the internal layers of that big monster of a function. The mythical black box entrails...

It basically is just that we learn coarse patterns in terms of position set volume where the action rule might be gainful, and then some boundary around that volume (not couting here... seeing potatoes with smaller or bigger sizes, I know, it might sound like blasphemy, but how? where.. the board is made of discrete finite squares or an 8x8 grid of points.. etc...

Well. call it neural netowrk magics. We all carry a wet gelatinous monstrous many folded flaps and slabs version of that in our heads. but that is a long story..

so fractal of nested patterns and caveats ooops.. I was saying that one person pattern caveat, is another person pattern to be discovered. The idea is that by decompsoing those moot and unrefiend action rules all in one principles into their logical components where we are careful to names the postion logic features antecedant where the trigger patterns for example are in the right rest of board context other possible features sets context that make applying the learned action rule of thumb, actualy logically gainful, not just by blind application of the action rule for its probaility virtues (pricniple here meanring that on average you might find gain more often than not, but that gets old as one improves).

ok. I diggressed. but this is about having a analytical language. That would be the baby of the bathwater. it is not because a blind principle was only a starter crutch is not not really a science of chess language..

that is can't be dissectied into a more debatbale language that is more cognent about the board logic. That is for me the dormant chess theory that is being left to sit in the dark. ok. done.

so the science of evaluationg future desirable imaginable positoins that the current position allows us to skip ahead and try mutated versions of juicy wishable position we might fooolishly try to reach with our already better trained part of the paly thinking models often well addresse by existing theory of learning. I should find that kotov book or chapter.

@data_gatherer said in #120: > only just read this. Is it possible that what you cal motif at more of the current position trigger features (those can be described, withiout being attached to action rules, often people throw baby with the bath water too eargely). Those don,t need to have names already having been called dibs in some book you might not be aware of, and therefore, you might want to find your own natural language analogies that might do the self-talk job helper to ruminate in the board foresight problem later in new positions. It helps with the slow critical thinking part. It helps separating aspects of the pattern awareness I would say. Perhaps you can even confront your board experience with other heads have their own distinct board experience trajectory, and those differnt angles might welcome your complementary view and attempt to share some more abstract idea that the very specific case where you started seeing a pattern. back to trigger feature I would call that a static pattern. Then my impression is that the woodpecker thing is about internalizing certain short sequences of moves into the "muscle memory" given the trigger pattern.. I would call those, dynamic patterns.. Now, those woodpekcy problems might have been rigged to serve you problems without the fog of the often neglected part of the foresight in game problem, which only records the move candidates from the currrent position being grown methodically from current position. But that is bypassing the whole very fast and hard to just use words, but Kotov did mention it in one book or was it chapter dedicated to thinking models in chess. The imagiation and desirable search aspects.. It is currently often left to brute force experience.. But I think that is where the evalution question of yours might be coming from. The woodpecker problem are rigged from the movie script writer... (ok in my home version of French, we would say. Arrangé avec le gars des vues, how exotic!). That means you have neve being train in the goal or plan imagination and desirable possible nearby position to aim at. there might be many patterns on a real board at the same time, and even if not seeing them, or not so many, (of the trigger type) , in real game you can't count of on the plan having been selected for you as in a battery of woodpecker problems. (or puzzles in general). That work of imagining possible futures within your current internal representation or skill set "map" or whatever chess vision horzion?. whel you can't imagine yet what your improved future self might on the same position if revisting it in some better experienced future.. so you are stuch with that. The problem in real game is the multiplicity of such possible imaginable nearby or less nearby goals to imagine and then find ways from current position to achieve that. ok. you could always go full bling brute force calculation from all legal moves starting from current position. but we all have lack of patience for that.. we like to try things but can we make our thinking more visible to our own mind. ok. I bet I have expired my welcome period.. So I think the question is great and contrary to this blog reluctant to consider chess theory, or principles as justificaoni, I think one just have to dissect them for their missing logic or context precision. that is actually fun work. debunking rules of thumbs. by finding their exceptions and then finding the missing logic and missing conditions board information in term of feature sets, often themselves nameable trigger pattern or building block toward such patterns. The problem with the usual third wheel learning crutches, beyond their initial get in the game deeper helping hand, is that they start becoming obstacles to learning the logic behind them. exceptions to the rule. is kind of a let down for me. I deleted a post i made somewhere, as not appro;priate there. but I gyuess while imposing my rambling here I could mention an idea, perhaps sounding out there. .but really a hunch when I think of chess and what NN might be learning about board feature patterns in their latent space (input space augmented by the internal layers of that big monster of a function. The mythical black box entrails... It basically is just that we learn coarse patterns in terms of position set volume where the action rule might be gainful, and then some boundary around that volume (not couting here... seeing potatoes with smaller or bigger sizes, I know, it might sound like blasphemy, but how? where.. the board is made of discrete finite squares or an 8x8 grid of points.. etc... Well. call it neural netowrk magics. We all carry a wet gelatinous monstrous many folded flaps and slabs version of that in our heads. but that is a long story.. so fractal of nested patterns and caveats ooops.. I was saying that one person pattern caveat, is another person pattern to be discovered. The idea is that by decompsoing those moot and unrefiend action rules all in one principles into their logical components where we are careful to names the postion logic features antecedant where the trigger patterns for example are in the right rest of board context other possible features sets context that make applying the learned action rule of thumb, actualy logically gainful, not just by blind application of the action rule for its probaility virtues (pricniple here meanring that on average you might find gain more often than not, but that gets old as one improves). ok. I diggressed. but this is about having a analytical language. That would be the baby of the bathwater. it is not because a blind principle was only a starter crutch is not not really a science of chess language.. that is can't be dissectied into a more debatbale language that is more cognent about the board logic. That is for me the dormant chess theory that is being left to sit in the dark. ok. done. so the science of evaluationg future desirable imaginable positoins that the current position allows us to skip ahead and try mutated versions of juicy wishable position we might fooolishly try to reach with our already better trained part of the paly thinking models often well addresse by existing theory of learning. I should find that kotov book or chapter.
#123

my previous post might be bullshit if the question of evaluation was not about the planning problem at all but about evalutating the current position. although I am not sure I would understand that question.

I think my own problem in real game si about evaluating possible imagined futures, and so not aiming a position that is not worth the tactical effort to get there... or if about getting somewhere, finding out intermediate positoin plausiblity, which is not about scoring type of evaluation but still about assessing the probability of own plan being reachable.

blabbla. disclaimer: you don't have to read.. I leave the blog author full license to tell me to go park my stuff elsewhere. no offense would be taken... well, I might have some ego bruise.. no.. just kidding. I will confront that situation when the times come. i do not think it woud hurt that much.. :).

my previous post might be bullshit if the question of evaluation was not about the planning problem at all but about evalutating the current position. although I am not sure I would understand that question. I think my own problem in real game si about evaluating possible imagined futures, and so not aiming a position that is not worth the tactical effort to get there... or if about getting somewhere, finding out intermediate positoin plausiblity, which is not about scoring type of evaluation but still about assessing the probability of own plan being reachable. blabbla. disclaimer: you don't have to read.. I leave the blog author full license to tell me to go park my stuff elsewhere. no offense would be taken... well, I might have some ego bruise.. no.. just kidding. I will confront that situation when the times come. i do not think it woud hurt that much.. :).
#125

@data_gatherer said in #120:

Hello. Any update on the rest of the 6 articles or so?
@chessTraian said in #121:
Hey Jack, how has your free time been going lately?

My free time hasn't been so free lately unfortunately, which was unanticipated :) I appreciate the reminders and for keeping me in check, but I can't guarantee the next post by a certain date etc.. will try my best.

@data_gatherer said in #120: >Hello. Any update on the rest of the 6 articles or so? @chessTraian said in #121: > Hey Jack, how has your free time been going lately? My free time hasn't been so free lately unfortunately, which was unanticipated :) I appreciate the reminders and for keeping me in check, but I can't guarantee the next post by a certain date etc.. will try my best.
#126

@DailyInsanity said in #125:

My free time hasn't been so free lately unfortunately, which was unanticipated :) I appreciate the reminders and for keeping me in check, but I can't guarantee the next post by a certain date etc.. will try my best.

You can't get any more axiomatic than that!

@DailyInsanity said in #125: > My free time hasn't been so free lately unfortunately, which was unanticipated :) I appreciate the reminders and for keeping me in check, but I can't guarantee the next post by a certain date etc.. will try my best. You can't get any more axiomatic than that!
#127

Hey everyone, today I'd like to share a quote I just found that's in keeping with the spirit of this blog series: "Just how many elements (a.k.a. positional motifs) could be named in total, I shall not venture to say. One hundred, two hundred, or more? What I do know precisely is that in every position there are certain elements of greatest importance, and that by unravelling them we enable ourselves to formulate the right plan of action." -Alexei Kosikov, Elements of Chess Strategy

Hey everyone, today I'd like to share a quote I just found that's in keeping with the spirit of this blog series: "Just how many elements (a.k.a. positional motifs) could be named in total, I shall not venture to say. One hundred, two hundred, or more? What I do know precisely is that in every position there are certain elements of greatest importance, and that by unravelling them we enable ourselves to formulate the right plan of action." -Alexei Kosikov, Elements of Chess Strategy
#128

@DailyInsanity said in #125:

My free time hasn't been so free lately unfortunately, which was unanticipated :) I appreciate the reminders and for keeping me in check, but I can't guarantee the next post by a certain date etc.. will try my best.
I do feel kinda awkward about giving you reminders, since I don’t want them to annoy you. As you seem to have implied before, you do have a life outside of chess that matters exponentially more than helping the few people who bother with such a non-dogmatic, currently-unproven way of improving at positional chess use it to beat more of their competitors.

@DailyInsanity said in #125: > My free time hasn't been so free lately unfortunately, which was unanticipated :) I appreciate the reminders and for keeping me in check, but I can't guarantee the next post by a certain date etc.. will try my best. I do feel kinda awkward about giving you reminders, since I don’t want them to annoy you. As you seem to have implied before, you do have a life outside of chess that matters exponentially more than helping the few people who bother with such a non-dogmatic, currently-unproven way of improving at positional chess use it to beat more of their competitors.
#129

you are keeping this thread from automatically archiving, which allows, future posts in that rare scope of topics that the op introduced. That is fine. You could say Bump... but having something to say on top is bonus. Fine with me. I am not the op though, but from relatively recent posts, it is fine with op as well. I think lichess still automatically closes (archive) after 3 weeks, right?

you are keeping this thread from automatically archiving, which allows, future posts in that rare scope of topics that the op introduced. That is fine. You could say Bump... but having something to say on top is bonus. Fine with me. I am not the op though, but from relatively recent posts, it is fine with op as well. I think lichess still automatically closes (archive) after 3 weeks, right?