Ok guys, learn your openings, the middlegame, the endgame... and of course, may the luckiest win!
maybe thermal noise? not even. Quantum uncertainty is still a practical random notion.
how much of the thermal noise random is actually about not being able to locate in its mechanistic state space all the interacting particles.
yes down to rather solid matter, we might be able to touch that uncertainty.
but even if i can't provide for an alternative. The uncertainty is still a mathematical function space attribution to a natural phenomenon that empirical data is not able to further restric (the basic wave-particly state space "position" uncertainty limit, "position" in state space would be 3D coordinates and velocity combined in a cartesian product, sorry for trying to shorten the verbiage of mine).
so from a "logical" or definition standpoint.. still practical. but well constructed from there on.
I am not updated about such concept past the last 20 or 30 years. But I don't think physics (the nature physics) would have changed since then. only how far we can extend the physics questions (the cultural physics).
So given I don't know about models of nature that would replace the practical random QM framework.
I might drop the "practical" for that.
I agree with the blog's reference to poker's term: "Variance". I mean any game could be considered to have a element of luck if you take into account every other possible factor that could influence the outcomes. But in terms of the game itself there's no luck because all information is right there in front of you to figure out.
Overall, I think this is more of a philosophical thing than anything as people's perspectives on this can vary greatly. So it's a pretty interesting thing to think about!
@PkGam said in #93:
> I agree with the blog's reference to poker's term: "Variance". I mean any game could be considered to have a element of luck if you take into account every other possible factor that could influence the outcomes. But in terms of the game itself there's no luck because all information is right there in front of you to figure out.
Yes complete information set to determine the next position given the half-move chosen and applied. However the win, lose, or draw outcome, is going to have so many decision points of the sort, which, yes at every step of the way are going to be what i call locally determined transitions from the previous complete set to its legal complete set successor position.
No human can tell at first move what the outcome is going to be with 100% certainty.
This is not only because of the uncertainty about the opponent choices (although that is a definite factor, of the unitary game).
I suggest doing either the real experience of the mind experience. of playing against yourself. Taking notes along each games about whether the game looks like a win for either side.
both players identical.... you would even know most of your next turn and previous turn intentions or plans.
This is called calculation horizon for humans. (the horizon is a metaphor for limited short term or working memory resources, i.e breadth of calculation). So, beyond that we have varying odds and intuition and chess theory taking the job of forecasting odds.
so the suggestion is about at what depth would identical players or a player doing its best for each side be able to tell the outcome. no position to position uncertainty. yet, not outcome certainty. somewhere in there we lose the "no luck" from the local transitions.
@dboing said in #94:
> No human can tell at first move what the outcome is going to be with 100% certainty.
And no machine can either, so what? If that was the case, the game would be solved, there would be no game.
It's a battle of minds: concentration, skill, memory, knowledge, preparation... You are not "lucky" because your opponent made worst moves, and it's pointless to simplify and gather all the skills under the umbrella of the word "luck".
@MarkIorio said in #95:
> It's a battle of minds: concentration, skill, memory, knowledge, preparation... You are not "lucky" because your opponent made worst moves, and it's pointless to simplify and gather all the skills under the umbrella of the word "luck".
Well, it means that the outcome of early decision in uncertain.
luck has already been dissected along the thread.. i was trying to wrap up back to the title.
The essence of random and advantage or disadvantage first are to be separated out of luck.
then what is random. it is ignorance or uncertainty.
then back to title conflation... leading to luck.
too much conflation does add some fog when observing things which have uncertainty.. (i.e. things to ascertain through learning, if we can).
@dboing
Decision points: Yes, but decisions aren't luck. Luck is basically a factor you have no control over such as rolling dice. So having decisions that are able to influence the game for better or worse is the exact opposite.
No 100% certainty: But that still wouldn't mean it's luck as it doesn't mean that you couldn't see the available moves as a possibility on your own turn as you try and visualize the lines. Plus if the opponent flubs, it wouldn't mean you got lucky either because it's not like they didn't have a choice in the matter. They just made a mistake and you seen a way to take advantage of it. People often call this luck because they don't expect a mistake or maybe think it's lucky if they see a vulnerability they didn't think they would. But there wasn't any dice that influenced moves because it's a do or don't decision.
Playing against yourself, identical players, calculation horizon: I'm not entirely sure what you mean here. You can definitely set up a position with yourself by playing moves with reaching that position in mind, including wins/losses for a side. But what I'm guessing you mean is that you end up playing out a game and hit a point where you find a better move than you thought you'd have in a position and switch to that or maybe just get an urge to explore a different line (equal or worse). It still wouldn't mean luck was a factor because it was a decision you made either way. If your exploration reaches a position you never seen before by playing your best calculated moves, you just continue to do your best to calculate it. Like if you do puzzles and start out at a position to decipher.
complete information set, legal move chosen. 100% determination of the next complete information set for both players. not argument there. This say there is no luck or uncertainty (i prefer this word as less loaded, no evaluation attached or in the back of mind, as luck seems in practice to be attributed to the player not the board, implying a goal direction).
the decision and its successor position is not luck, it is determined by previous position and the one thing moved. completely known all of these things. at the single transition level..
also lets take the evaluation connotation out of it.. random, uncertain, unknown instead, undetermined.
but chess is not a one move game. it is tight sequence of such things evolving in a huge space of legal positions.
interlude on branching degrees:
and at each position there is not only one legal moves but a variable branching degree that with lone king to king starting in the free interior of board have each a maximum of 8 legal moves.. and a minimum when in direct mutual restriction (same file or same rank, and only 1 squares between them (their zone of influence maximally intersecting), then 5 each.
with all pieces on board and if we did not care about occlusion if we could superimpose all the plaments and moves... 32 times each mobilities of the infinite board intersection the finite boad. (sorry i don't trust myself babbling this out).
one needs to substract then for legality of the superimposition...
end of interlude.
that big chess space is where the random or ignorance happens...the uncertainty is not about any pair of complete information sets whitin the player (even the self-player) horzion, it is about any pair betwee current position and any position outside of the horizon.
machines also.. they have more resources. but some new machines have been taking the problem from an inverted point of view.. Working with the inverse problem of sorts from a mesh of outcome data to and "dynamic" mesh of trial decisions. They have lots of resources too. we can't do that ourselves.. One is acknowledging uncertainty in its formalism, the other is struggling with it as its tentacules of exhaustive search go deeper and deeper (and arguably narrower while at it).
I am not talking about move-only-chess but whether you know with certainty if that move is a winning move.. or if you will still need to work it out..
Do you accept the notion of horizon. I think some people do not even accept its existence. If so, then we are in classical game thoery land, or all knowing extensive tree rational player. And there is neither transition luck neither outcome uncertainty.
neither truncated tree advantage amount uncertainty.. because having the full extensive tree, the normal form calculation of odds back to the truncated tree, would be the exact odds. you either know the full tree or you only know the truncated one
the complete exhaustive search machine would have the complete tree and would not need forecaster that we now have taken as true outcome currency.. maybe using engine gives a false sense of certainty.
do your best.. we are in agreement.. but that is not certainty. we learn to shape and reduce the width of uncertainty with many games.. we can also use the engine.. but i have fundamental doubts about prioritizing the per position engine opinion over my own rational abilities to find clues on board.. We use those clues to the best of our evolving abilities.
this is why chess is a fun lifetime endeavor.. (also even if strongly solved, we still don,t have the in-game resources to store that).
The problem with the title, and now a long thread, is that it can be dissected 2 ways (table). "luck" and "the game".
I never said the individual transitions had any luck. The tension between that and the big chess world of outcomes is what makes chess still a game to be played throughout ones life.
I do not find tic tac to very stimulating in the long run. Its extensive tree is likely to be in some human horizon.. and all sorts of things in game theory can be illustrated there to motivate applying the same theory to chess.
edit: about "inverse problem" of knowing game outcome to learning to play. I said we don,t have the resources... well not to the extent of silicon machines. But the longtime ignored par of chess learning that is using our many games experience based in our intuition itesf based on our long term memory processes (I think of associative type, so as to be clear about coined term from cognitive high level science i might not be aware of, i think more neuro than psycho, but i do extrapolate to psycho). I think that is why we tend to attribute more human qualities to leela types of engine. They are built as intuition learners. Statistical machines. So is our animal brain, the global analytical part of it, that might only approximate logic, but has some leaks (why we need our concsious logic filters, at least for that).
We are small scale hybrids of the 2 species of engine. or they are dissected models plugged into high voltage we would fry under. both having to deal with the big extensive legal tree to extract some better chess out of it..