I'm not sure I understand the OP. But maybe a simple example would help (or reveal that he means something different). Suppose we wish to evaluate a position with respect to the side-to-move. For simplicity, suppose we only consider "mobility" and "king safety". Say each is assigned an (int) value between 0 and 15 inclusive. Suppose we decide that mobility is twice as important as king safety, so we weight the former with coefficient 2 and the latter with coefficient 1. Suppose we have a position where the (base) mobility eval is 10 and the (base) king safety eval is 5.syzygy wrote:Some kind of Euclidean distance between "white's position" and "black's position" would only measure the dissimilarity between the two. That is of no use whatsoever as a measure of how good the position is for white or for black. (Well, if white and black's positions are identical, the evaluation should indeed be close to zero, but that's really it.)zullil wrote:But maybe I'm missing something ...
Then one way to compute the total eval is 2 * 10 + 1 * 5 = 25. This amounts to calculating the taxi distance from (20, 5) to the origin. I think the OP's suggestion is to calculate the total as something like sqrt(20^2 + 5^2) instead, which would be the euclidean distance.
I still don't see any advantage to the OP's approach ...