Engine for human chess learning beyond mere tactical best next move blog series (overflow thoughts)

Article:

Finding Threats and Piece Manoeuvres with Engines by lichess user jk_182
https://lichess.org/@/jk_182/blog/finding-threats-and-piece-manoeuvres-with-engines/2Pm5O3JT

https://lichess.org/forum/community-blog-discussions/ublog-2Pm5O3JT#6

Finding threat featre is like a null-move pass your turn thing. If there one passes one's turn here might be what the one sided oppononent threat spectruym might be (might be nice to have PVs too, for the "idea" is rarely alone.

rest moved below.

In hindsight of the below ramble. I intended to give a header phrase, and end up making a self observing rationalization of my writing behavior. while describing what the other rambling was about.

the introductory rambling about the past rambling that is now below this introduction rambling (this is what happens when I look at what I just worte, a never ending-post-editing disclotaed possibly nested in worn order constellation of rambling, as I need to explain the actual writing chronoloical order not being necessary the best thematic ordering. and this post rambling was meant to act as introduction/conclusion. blablabla.

The many hidden ideas under best play. and puzzle being a good more than weak win play leanring experience. About that. And not just about error detection. But how we plan and change plans. etc.. (microscom of chess already a micorcosm of plenty cool stuff to wallow our curiosity onto, into, or withing. .whatever. need linguisic help, wihout dilution. So it might not just be a compressing job. it might be missing articulation of reasoning critical linguistic short strings, that as usual I did not have time to instruct my one character imcompressible or unlumpable (as in continuous pensmanship, we can think to write words while still thinking above, but here, the lowest even subsconsious autonomous thinking is stuck having to create one character at a time. then words sequences, etc (but that is another, more my own, about the conflict between the near subconscious parallel nature of pre-verbal thinking, and the note-taker, always running behing mind scribe that is being time pressured on the typing side as mentions, and on the stormy side of that cognitivo-emotionally guided logic grasping imagination trying to find some mechanistic understanding of the question at hand from others, and own threads of connected many thoughts, the minds eye 2D slice probable most wordable, but still like a 3D printer, another bottleneck of though communicaton that is not just the ASCII quanta of delivery.. That is my mental problem. And my math. training, while having solved that problem for existing math. shared, is not really helping when in the discovery mode, of I don't know yet what will solved. i.e. research. it might not be research to those how know already, but from this mind prison, well any thing not yet known, and motivating curiosity enough to put such sharing and thinking energy, well it is. Learning obvious zeroth premise often hidden assumption, is that one does not know yet. and might not even known what it does not know.

it means leaps of logic are hidden premises.. etc.. now. I ramble so much that I am making the spill over the next veritally spaced wall of text.

as promised the first draft rambling from the post blog discussion.

In puzzles the SF side is actually spectacular about this many ideas, as the puzzle solving side on the other hand is design to be mono-idea. But as patzer using puzzles not just to cranks ones rating up, one can use them on both learning sides, even when failing, going back after the trial and error (and last resurt givng up, I admit to have done 2 of those recently out of maybe 30 puzzles on hardest as my patchy skill set would randomly find such hard puzzle solvable, but as my rating increases I get exposed to longer puzzles of hard steps and like will end up failijng more and more, but some static position sometimes put me in front of blank imagination and I would give up, and learn about the very first position, but here I am talking about finally figuring it out with many mistakes so one can discovery the trickle of ideas leading to the solving (and often too early, lichess, but that is not just hard puzzles, it might be more intense the lack of satisyfing static position at solving point, for the solver, even if not having done a failed step, but let's put that Deep engine seeker problem (puzzle selection by puzzler).

The many ideas are more experiencable about the resources of SF to find all the tactical nooks and crannies that even a successful weak human solve, might not foresee, and often during the very solving the idea become a secondary idea or implict idea given the high level play on the other side, often finding a counter move of the nooks and crannies with my sub-ideas that might mitigae the prior human idea threat. do the best move is not the only idea in the position to lear about. And only looking at best play is like tricle down economy school of thought that is still fooling whole populatons (not explicitly in the language that fools but in the funding entities that make the ones with the fooling language be very visible, my democratic process contribution, right here, sorry off topic much...).

where was I. yes. best move. pale shadowm approximation. 3 PVS might help. but the idea world. and the learning context which this blog seems to have adopted. We are not going to use engines during a game, and if a game against engine, still we are not going to talk about ideas while playing against. So this is the bigger context of chess study. not chess performance. A friendly reminder of that huge world of chess that has been shrunk to some small subset in the lobby. Yes I need to plug that remembrance of a better chess world view of old.. I am a bit rigid sometimes. or maybe I can remember a lost good thing where I don'T see one anymore.

I think I circonvoluted enough about how chess and ideas are not just hindsight trickle main line to consider. And suggesting that puzzles being some imbalanced from the SF selection to the SF opponent side. that this might be a rish problem for this idea and all hidden ideas that the best move single line study is complete oblivious too (engine or human by the way, best chess, is a tiny part of the chess world, and if we took the bull of the real chess activity that dominates which is learning not just through tournament preformance games on histociral or current record, but all the other games now democratically or demographically all also on record. Those have more of the sub-ideas that even weak solving play (puzzles or single games won) contains.. This is the foresight problem of chess.

This is my thoughts when I get to have to change my plans in puzzles even if I find the solution, I remember my actual thinking, having to change so often that in the end I see I was deluded while doing the right move at the begging, because even if the "wrong" idea of future desirable, that desirable was stlil a lukring threat enough so, that even in the resourceful best play of SF, I would still get some gain relative to starting position, or maintain soemthing that was under danger of losing (but those all not really puzzle spectacular, I know from all the unsatisfying puzzle end being not on par with the tactical aborted weak solution ending often at multiPV plateaus).

Now go to read what the author could do with a single pass you turn other side threat. (I also sometimes use it post puzzle to verify I was not deluded in my piecewise plans. Such null move other side threat is indeed the local threat, and it won,t disapper because the ply. returned to owner before the passing what if, would have acted against it.

surely someone has thought about this. and could rephrase better

My only problem with this approach is that this only detects very direct threats. If one side would need to play two setup moves to create a threat, the idea might not be found after the first setup move (as long as the final threat could still be prevented).

Well that is a theme of a problem using an engine as oracle for the best move, and not being diligent about the subtree information that was compressed into such a mere one dimensional number. This the dominant only problem in general. It seems to sustain itself through ignorance bliss and some kind of assumption that nobody is curious about what it not shown.

or it may be at the very end point of the real optimization tangible goals and endpoints used to bring them to our attention and human position analysis tools, or learning helpers. It might also be a calc of how one seek advice from better players, or are taught that looking only at best moves at lot, and through long games, it how to optimally learn. Many things might be socially community wise converging to the oracle usage might be best theory of learning out there about chess. the chess forsight problem I would assume. But then I am not sure that this is what is being sought. The improvement implicit assumptions are that there is some kind of objective measure, otherwise why not be satisfied with chess learning. I thought initially that "adult improvers" was focussing on the adult, sill hoping to improve, alhtough having missed the train when it departed for their life span... Which is really the only way to reach the only ambition there can be in being attracted to chess.. The unique champion.

no matter how many applicants. There can be only one. And chess is hard work, I won't deny, so what are adults really wanting. just any "improvement".. ok i digress. i admit.

So. it that article about finding way around the lack of entrail sharing (and oracle might want to have some privacy with those).

or. some way to mix human thinking and sparse usage of the threat indicator. I guess my point above, could be helping that threat indicator, as being of plan nature. being a what if passing turn (or swith, in the alternating obligatory move turn-game)

now I have another reading carrot whether right or wrong. going back.

also the blog says "had". I saw it in puzzles. but maybe i have lichess tools doing it.. ohhh. now I should read that post by its author (the tool). It might have been from scratch in LT. and now, wowee, an indepedent take existing to consult in history on github?
doing the same thing. I would be curious myself. being stuck in only one mind, makes such things curiosity candies, I would project onto that other person. say exact substrate on lichess. Why would lichess have dropped that. not orable enough? Did people complain? or was it apathy for anything else than oracle best move. That threat is not even real.... if I can prevent it. Sorry, I hope nobody is thinking that. but I am exploring what I might not know at all. Such mysteries around chess. or in my glass of stormy water.

One idea I had was that one side would be allowed to make 3 moves in a row. The problem I ran into is that it's not possible to let Stockfish play as if it was allowed to make multiple moves in a row. So it might not even try to create the threat, since it sees that it can be prevented, even though we would allow Stockfish to move multiple times in a row.

Yes, expanding on the impossible. the engines are for horse races not analysis. but that is making the point very deliberately. Explaining from the human chess point of view the types of things we might want to look at from a single position moderate depth search for examples. and its cloud of evalations, perhaps before it all get squinched (?) into one single number to rule them all (other chess information it might have gleaned, but did not need for its well standardized end point selection pressure immutable definition (the same as weakly defined theories of learning under the vocable of improver, we want deeds, and I get same thing for engine.. none of that wishy washy imagination and pondering stuff... the deedds the one move... etc..).

caricatures. no offense, as it might not be anyone that could combine all those attitudes, at least non concsiously, and they might just go with the flow.. something my Don quichotte does not read the memos about...