5

First of all, I am explicitly not asking for a list of endgame principles, nor am I asking for resources to study endgames from.

My question is: How does one develop endgame theory? The main motivation of my question is that I have been trying to study endgame theory for a chess variant (atomic chess, as it were) and there is precious little of it. I am quite literally writing new theory the deeper I study.

I'm looking for answers that could address either or both of the following points, in regular chess or in chess variants:

  • How endgames (regular chess or any chess variant) were analysed, organised and codified over the decades until we have our present-day knowledge;
  • How to study endgames, but without access to any books or professional games (just a chess set, a brain, an engine, tablebases, and lots of patience.)
  • 3
    Start with 3-men endgames, then 4-men and so on... en.wikipedia.org/wiki/Endgame_tablebase – David Aug 1 at 9:56
  • I already have access to (6-man) tablebases, as mentioned. I'm asking how to study endgames as a human, generate rules of thumb and principles and classify endgames. – Remellion Aug 1 at 10:14
  • Possible duplicate: chess.stackexchange.com/questions/1702/… – Phonon Aug 1 at 10:57
  • 1
    @Remellion What makes you think that the "endgame classification principles" of regular chess apply to Atomic? – David Aug 1 at 10:59
  • 2
    @Remellion I made that change to counter a vote to close on the grounds that the question was unclear and because I thought you made it clear in the body that that was your main intention. I think it is a good question and hope that it remains open. – Brian Towers Aug 2 at 9:22
2

How do you want to do it?

I know that sounds flippant, but that's pretty much the only good answer to questions along the lines of "How do I create something that no one has created?"

Chess endgame theory developed by people starting from very simple positions and identifying commonalities among them. Typically, endgames get classified by the pieces involved (Rook endings, minor piece endings, etc.) so I'd start with elemental positions that are clearly won/lost, see what leads to them, try and see common maneuvers to create them, etc.

From simple positions you create the building blocks for understanding more complex positions.

Rinse. Lather. Repeat.

An example is pawn endings. We started with simple concepts like the queening square, moved on to techniques like opposition, and kept building until we ended up with a theory of co-ordinate squares, which covered all pawn endings. Opposition, Triangulation, all of that stuff boiled down to being special cases, subsets within that theory. But it all started with atoms, from which we built molecules, and so on until we eventually arrived at an impressive, comprehensive, edifice.

There's no single uniform best way to approach something that's never been done. You try something, if it seems helpful you keep using it; if not, try another way.

By definition, there is no mapped out course to get you safely through uncharted waters.

  • Flippant indeed, but a fair assessment. I might ask a separate question in the future to clarify what you said - is it really true that historically we started from the "square of the pawn" and then developed the opposition and other more complicated concepts, or were there complicated individual problems and studies composed before anyone had explicitly written anything about even basic ideas like "king opposition"? – Remellion Aug 21 at 7:51
  • Well, more likely we started from "I know this position is won," and built a library of won positions, at which point someone else said "memorizing all those is silly, there's got to be a better way!" And with each pass over known positions abstractions were refined, until we have what we have today. The beginning position is arbitrary, but I'd suggest the simpler the better. And not sure if they were composed, but the odds are they were. – Arlen Aug 21 at 19:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.