But I think what the GM who wrote this and other strong players have tried to say is there is a misleading element in self analysis by 3300 elo computers,
where losing a center pawn is in their eyes a great sacrifice because they see some 3000 elo tactic down the road while your attention is elsewhere . and if your positional understanding is focusing only on “how do I get my queen in the game, it’s passive” , you might overlook the importance of a center pawn as the entire life line of the position.
Understanding that in some position the most important piece is a pawn which cuts circulation from opponents position, will affect your candidate move selection. (Just an example, of many possible examples) Doing a mediocre amount of puzzles daily and thinking all games at your level are won by tactics will affect your candidate move selection.
So the point he’s trying to make is that the games are not won or lost by tactics as the way it seems from engine analysis .
one thing about tactics obsessed analysis is that it is mostly about mistakes to avoid.
that will never bring the planning thinking mindset. of setting and assessing goals.. the positive view of chess. not just surviving one tactics time window to the next.
I guess if just having a mistake not to make view of chess, one is not going to be able to convert experience into a consistent set of components that should be increasingly well assessed in minds eye, toward planning and tactical execution confort.
So I don't think someone ought to view strategic advice as mistakes not to be made, but more as planning positive arguments (not that some of the arguments might be that one component seen in mind's eye might be detrimental, but that the psychology is not about: avoiding all the possible mistakes i have seen myself making, the catalog of mistakes not-to-make might get overwhelming.
Rather than tactics and strategy, I feel like the concepts of calculation and heuristics are more useful as a way to organize your study or focus on how to improve.
"Strategy" seems like it only exists in chess because it's too branching and long to feasibly calculate, and so you end up with a collection of heuristics in order to determine what might be a good move. And there's a nice collection of concepts that we have regarding this, such as controlling open files, controlling the center, bishop pairs, outposts and all that. These are heuristics we use to judge whether a move was good or bad. It's always good to try and understand what "good principles" are in chess, learn as much as you can, and try to actively run through them when deciding on a move where tactical calculation isn't yielding anything.
Engines also greatly rely on strategy, but their strategy is on the level of choosing which branches to calculate and evaluate, as they can't calculate them all. This is why it's good to always remember that just because stockfish thinks something is the best move, it may not be (hence some wild games against AlphaZero).
"Tactics" seems to contain the core of what chess is as a game at an analytical level: computation. There's not much to say here. You simply work on calculating things accurately and finding the line that yields the best result. This is where puzzles come in, though I'm not sure that puzzles are *just* tactics, such as finding a way to force a mate or trap a piece or force a draw or whatever. I've seen plenty where the goal was just a better position, with no clear route to anything decisive.
There's always the unsung heroes, though, like visualization and chess memory. One of the biggest differentiators between high level players and the rest of us is chess memory. It feeds into so much that goes on. It lets you remember opening theory. It allows you to more easily see tactics. It allows you to better judge heuristics by evaluating. It lets you completely ignore whole swathes of moves and positions intuitively. These skills are trained just by playing chess, but it's always useful to sit and try to remember move sequences after a game, or to sit and try to visualize a position.
Edit: I think this is my first post on these forums. How does solving a mate in 1 prove I'm human? ahaha
material counting is also a heuristic....
the capchas on the internet are rarely to prove being human. At least not here, it is to slow your posting down.
Or you can view it like a chess candy reward. They use to be boring, but i have found that maybe a year or more ago, they had more diversity. A little surprise about which piece to move not always being the most powerful one.. for example...
Really not about being a machine or not... What machine would want to post in a chess forum anyway?
Thank you for your first post. Welcome.
About exhaustive truncated chess tree exploration engine and "strategy". godepth, gonode or time limits. always truncated.
If assimilating strategy to beyond our human horizon of full min-max calculation, that kind of long term, only just being deep consequence of early decision, i would not say that the programming of the engine is to consider human strategic or positional clues** within human horizon.
It does amount to looking like that from our human eyes, because we are not looking at the positions from which such engine get their evaluation from.
These are deep beyond 16 plies to mention an arbitrary number out of the blue, and we never get to look at them.. so we have that score of the current next position best candidate, and we don't see anything like a material advantage in our human horizon or the few plies given, and then wow. keep going and well deep there we do get our visible explanation. extending the horizon by machine far but not always to terminal positions.
I don't know the proportions, in the SF sub-trees complete set of leaves, for typical go-depth parameter values, of leafs that are legal terminal versus those of heuristic evaluation (dominated**) by material count. But I assume we are far from depths ensuring we are looking at the complete AB pruned chess tree. (other reasons than depth too, but here not needed).
So I don't agree calling any of that strategic engine thinking. not in programming (by virtue of historical parameter optimization i claim**), and not about WDL odds. I don't think that our human positional/strategic signals are about predicting material imbalances later more than they are about estimation WDL odds.. or just winning odds for one player doing the assessment.
** its leaf evaluation programming function does use terminology as well as, in abstract of actual parameter values, potentially strategically predictive components (positional signals of the input board), but the history of parameter choices as those have been implemented, has been made sequentially for most of its history, resulting in many frozen coefficients to those possibly strategic types of clues, that were all dependent on previous sequence of chess theory components frozen coefficients one at a time, back to material imbalance counting heuristic, that is my understanding so far from what i could read, with NNue in the picture it is the whole classical evaluation matrix of components that is frozen, did i gather).
that is maybe off topic but i thought knowing the internal of this type of engine might be crucial to the discussion. as introduced by previous post. It would be nice for all to have access to those associations between current position and the deep tips being probed. I would call engine thinking tactical extensions. just that we never see their tips where they could have been doing strategic assessment.. maybe they do now. how could be know.?
of course there is some room for re-purposing terminology, to redefine heuristic material imbalance outcomes deep from current position decision and human maximal true calculation horizon, as strategic assessment of the current position receiving a score from that deep. I don't think that was the intent of previous post.
Some people used to think (and i think I was) that masters actually calculate as deep as they reap the rewards of their early decisions, but review Carlsen interviews, and he had the transparency of being clear about how he comes up with some of his good moves, I have been told, and my understanding is that it has all been internalized, and does not need to be calculated exhaustively. The imagination is already having the right suggestions and only need complementary calculations.. probably in positions that might have never been encountered in all their parts.. that would be nice to study, notions of similarity and how far does a position need to be from experience for calculation to feel needed. or is that more about how much room is being sensed. how very tactical a position might have become (closing a loop).
That it is necessary to modulate the importance given to tactics seems logical. However, it seems that the tactical level and the elo rating are correlated! it therefore seems normal that the tactic be worked on, and more generally the calculation. if the "tactical" rating on lichess, for example, does not change, it is to be expected that the FIDE elo will not move either.
Well. so... how to improve tactic ? the chessmood training tactic programm seems to be very basic.
Okay, but then, what is the best way to study strategy? I mean besides paying 600$ for the course linked in the article :))
@LeonardoJBX said in #8:
>
Winning chess ( by Sairawan). Read the whole series. you will be at least 1500 by the end of it.
Agreed That Strategy is The Best Weapon
The Best Blog In This Year I read
Mastering tactics is of utmost importance for almost all chess players but to achieve it is a really onerous and time-consuming task. Repudiating its significance and focusing on openings and strategy is a fanciful way to improve. It has to be trained in a systematic way. An insular truant ominously may never become a good chess player.