@OctoPinky said in #10:
> I was thinking about moves like this: I followed general idea "push passed pawns", even better when you can push them with check. Well, it seems one of these is a winning move, the other (actually played) ruined my advantage.
so seeking a human rationale about why SF says a5 while you did push the pawn?
> "push passed pawns".
I can't say that I am familiar with either this action thumb rule or when does it apply.
But given the connected pawn pair ou 2 pawn island. i have seen a better pawn technique of alternating phalanx with chain a sort of 2 pawn walk ratchet. That would have prevented the kind from blocking the pawn.
I do not see here that the check is really taking initiative as the defending reply is canceling the first intention of pushing the pawn (further as needed).
Ok i see the bishop controlling light square complex around that possible 2 pawn advance. I do not have a principled point here other than the more autonomous pawn ratchet, which might be just about move ordering.. you had the right idea but should have prepared against the blocking as check reply.
I am just using my current limite chess vision thinking here, i am not talking from autority. Thanks for sharing.
so in hindsight as too late for me to experience this in foresight.. and this is not the question anyway.
what is your question. Finding board features clues that might help reproduce the better move elsewhere.
so I do think the plan goal is the idea, but that it not the method of immediately possible. The bishop makes your own king pawn rachet protection job, needing a darrk square pathway.. which without having checked, and not typing without board even in peripheral vision, but kind of remembering the starting position.. I think the 2 pawn march (it might even be a thing... I should ask my mentor....because I recall having discussed this at some point, not on the rim, but near the extended center. Pawn technique is still a skill set gap in my internal representation of the chess board dynamics (etc..) which some might call chess vision, but I have ambitions to formalize this in some alternative time line (lol), and internal representation seems to need imply that it might be evolvable and related to our experience and how we learn from it.
So 2 pawn racthet is my go to words for a bunch of chess stuff. I see many ratcheting patterns.. most often those are in endgames.. I tend to call repeated dynamic sub-patterns that way. but it might be sometimes within a set of such sub-patterns. Like in the KNBk endgame. although it is not as clear what are the "atomic" subpatterns being ratcheted..
is that helping. I am just describing what the SF suggestino seems to accomplish. I think perhaps dissecting ideas into goal or method might help assemble the bulding blocks in order.
yes to wanting to push the pawn or its adjacent neighbor. Maybe that checking extra idea seems to be not as important to me, as it is easily escape while also blocking the pawn. I know you were going to bring the other pawn later.
so maybe this is really about learning the virtues of keeping pawn connected either as phalanx or chain. in one configuration it has tight slab interdiction forward against kind ever consider getting close.. and that might be room for the king to interweave itself in the 2 pawn ratchet. when a dark square appears..
it might just be luck here that this is allowed. I stop here. what are your thougths.?
so much rambling. I hope it is stimulating and not a chore.
@dboing said in #11:
> is that helping
Thanks for your thoughts. It helps a lot, actually. Somehow the ideal setup is placing pawns in the same rank (say a5-b5) which is no longer possible after b4+. This is already known against a King alone, but not that obvious with that Bishop still on the board... and the passed h-pawn.
But my point was, rather than discussing this particular move, to show that computer analysis alone is just wasting time when facing ideas beyond one's knowledge (this sets different thresholds for different people, of course).
So my advice would be to use computer analysis to find almost obvious mistakes and/or study some few non-obvious ones you can learn from, but not to punish yourself contemplating each and every blunder you made. Even more when computer rates the impact of them (-2, - 2.6, etc...) but not how difficult are they to find.
@OctoPinky said in #13:
>
The problem is that this ceiling only scope on the human player whole game, is not coming with an adapted conversion rule to the player internal representation current "accuracy" (not that ceiling one, we just lack vocabulary so we might as well all try to throw our 2 cents and expand, something might stick better after enough heads chew together on this..).
It is very noisy feedback for the learning problem (yes we also need to name a few more things that have been under the rug for too long, am I talking about chess only, I wonder, did not mean to at first, but now it seems it might be a general rule of thumb: introspection for all!).
err. so proud of that I forgot where I was going.. to be continued....
I was gonna say the noise comes from the score not telling us much about the human breadth foresight (and then the player "level" version) that would have been necessary for that ceiling point of view "mistake" **.
we would need to know the depth of the leaves (and those leaves, if learning was the intent, lichess and its mission of education reminder here), so that we can filter the mistake noise with those
1) that have human discovery potential, and
2) could further research how to within 1) find where the learner is really going to learn somehitng that can generalize (i.e. learning in chess is not about imitation, it is about improving generalization of experience and learnable, most appropriately). Here I think I have undestood your point. maybe. @OctoPinky
** whatever, inaccuracies, a point is the amplitude from ceiling is noise most of the times, given the level I am at and if I was playing with time pressure, playing correspondance is pruning on that noise, if I were even to use SF, but I can see my mistakes and their logic by myself, and if I don't well there is always another slow game, but usually I can see my mistakes from having made plans as hypotheses, that is what slow games allows.
Comparing chess engines to humans is utter nonsense. It's like demanding that Formula 1 drivers just sit in the seat of a Ferrari, with the pit crew driving the car using telemetry. Even though that's almost what's happening in practice, viewers watch it as if it were a real sport. Cheaters on lichess.com and chess.com act exactly like this, and I even assume they're proud of their cheating. How good are they at winning this way? The joy lies in deceiving others. I understand that the goal isn't to have an ELO of 3,500 just by calculating millions of lines of variations, just by being a giant calculator. The goal is to have an ELO of up to 1,500, but through your own effort. The sport is getting really weird...