Comparing chess annotations.

  • #1

    I am familiar with two types of annotations that are used to evaluate a certain position.  Computers tend to give a numeric assessment of who is ahead by how many 1/100s of a pawn.  For example, a position may be evaluated as 2.75 meaning that white is ahead by 2 and 3/4 of a pawn.  Some evaluations give informator symbols like =  +-  ect. to indicate who has some degree of winning advantage.  My question is what is the correlation between the numberic and symbolic evaluation?  Is there some type of grid or chart that would convert from one to the other?  For example, if the position is evaluated as 1.33, does that convert to an equal, slight winning, clear winning etc.  Thanks for any input. 

  • #2

    Two things I know:

    a) The Informator symbols have words that go with them. Like = is even; +/= is slight advantage, etc.

    b) Fritz and other programs have actual break points that are equated with these symbols, though I can't remember what they are exactly without looking at the program running. So, for example, less than +0.32 might be =, from 0.33 to 0.72 might be +/=, +0.73 to some other number might be +/-, etc.

    So, the answer is YES, but I don't have the numbers in front of me, and it's possible that each program (Fritz, Junior, Rybka) may actually use different values. I'd like to know the answer to this question, too.

  • #3

    I’m not sure there is a direct correlation. I just did a quick experiment with Rybka and came up with the following:

    0-.25 (=)

    .25-.60 (+/=)

    .60- up (+/-)

    The evaluations engines give are based mostly on material. A GM’s experience may tell him that positional factors outweigh any material advantage. In fact top this is why top level CC players routinely beat players who rely on engines alone to select their moves. The top level players can direct the game into positions that engines do not evaluate correctly.

    As far as individual engines go top level CC players recommend analyzing with at least 2 engines because of the difference in their evaluations. When I check out games with Chess Tiger it’s evaluations are often way more optimistic (or pessimistic) than Rybka’s.

    I saw a book in which Lev Alburt recommended using a slightly modified system to evaluate a position. He listed 5.0 as being equal. At 5.5 for White vs. 4.5 for Black then White stands slightly better and at 6.5 to 3.5 White is clearly better, etc. His recommendation was laughable because unless you’re a GM you can’t evaluate the position accurately enough to make those kinds of numeric judgments. When I saw that crap it made me wonder about anything Alburt recommends.

or Join

Online Now