4

I was wondering if the only way to evaluate a position is by looking for next moves like a chess engine does.

Is it possible with only a static evaluation of the board to get a correct evaluation about which side has the best position?

4
  • That's not how engines evaluate a position. It's fairly trivial to set up, or even find, positions, where one side has more material, or other advantages, but cannot avoid a forced mate.
    – Herb
    Mar 4 '17 at 12:59
  • 1
    You don't think like an engine.
    – SmallChess
    Mar 4 '17 at 14:03
  • No i dont, but that Is not the question literally, not just engines look for next combinations moves, People also do it. Tha question Is if there's a way to know if Is a good position for White or Black without looking the next moves
    – pimi
    Mar 4 '17 at 14:08
  • People are normally scoring end positions of variations, so scoring positions is the most normal thing for chess players. We are usually trying to eliminate tactics as evaouation function is quite static. You can quickly evaluate how deep search is needed based on piece placement and possible tactics. By scoring starting position you make mistake in solid percentage of cases as you apply static function on too dynamic problem. That's point of calculating and you can hardly make it useless part of chess play.
    – hoacin
    Mar 5 '17 at 6:51
3

Evaluate without calculating.

Source: The wisest things ever said about chess by GM Soltis

So yes, we should evaluate a position before selecting candidate moves and calculating concrete variations. Considering whether the opponent's king is safe will help you decide whether you should look for a mate in N type move, which is very different to capitalising on a space advantage.

From Tune your chess tactics antenna: know when (and where!) to look for winning combinations by Emmanuel Neiman, he recommends the following steps to selecting a move.

  1. Global vision: think of this as walking past a game and your first impression being 'white is winning' or 'it's about equal'
  2. Analysis of the position: This is where you look at king safety, material, space etc.
  3. Looking for the theme: is there the potential for a pin, fork, back rank mate etc. in the position?
  4. Looking for candidate moves: based on steps 1-3, generate a small list of candidate moves (usually 2 or 3)
  5. Calculation of variations: work out the candidate moves as far as you can, and evaluate that final position

I would argue that steps 1-3 fall into the category of evaluating a position without looking ahead.

2
  • Can you develop a Little more the fiesta Threesome points?
    – pimi
    Mar 4 '17 at 16:56
  • @pimi: I've fleshed out the points in a bit more detail.
    – user1108
    Mar 6 '17 at 9:02
1

How to Evaluate a Chess Position would cover up a Book or two . There would be some infinite Chess Advices from different GMs in respective Positions .It is also a Time factor parameter . It really matters how much time you have on Clock to determine the various elements .

Some of the Simplest way to determine is

  1. How many Pieces are Active on respective side . You can count the number of squares occupied by your Opponent's pieces .

  2. Who has an Attack ? I mean whose pieces are covering up on a particular side of the Board .

  3. Is your King safe and Opponent's Bishops , Rooks are not pointing towards your King .

  4. Do you have a Passed Pawn and Opponent's Pieces are tied up to block the parser .

  5. let's assume you are in Middle game and you have a chance to exchange the Pieces and go to End Game directly . You have two Pawns one on either side of the flanks and your Opponent has the same material (two Pawns) and on same side of your King where it can Blockade easily .

So above are some of the factors where Position can be evaluated and you can make the Critical decision .

0

Is posible only with a static evaluation of the board to get a correct Score about who side Is in best position?

Not sure what you consider correct Score. You can certainly evaluate a score for the current position basically in the same way that engines do, giving points for material, piece activity... And that's what humans do first when looking at a position: they look for material imbalances, piece activity, space advantage, king safety, etc and end up with an evaluation of the position including an idea about who is better.

Based on this evaluation you would develop a general plan: do I need to improve piece activity, do I need to defend my king,...?

Only then you would think of concrete ways to implement this plan, candidate moves and calculate variations. Because of the initial analysis of the position you would already have limited the number of variations that you want to calculate.

Obviously if you calculate a score from the current and possible future positions like engines do, the score will be more reliable than if you only calculate a score from the current position.

I can think of two exceptions to all this and that is

  1. forced lines, where you don't have much choice (e.g. re-capture after a capture, getting out of check,...).
  2. highly tactical concrete positions, where an evaluation on general principles will be useless and you won't be able to get a meaningful score without considering possible variations.

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.