In the Texel tuning method, it is the eval that is tuned.Laskos wrote: ↑Sun Jul 28, 2019 2:48 pmEven in these very particular conditions, how many parameters are there in Texel tuning method, aside eval? Another goal of that method is to make the mapping of the eval to score for the whole game via a logistic. And it fails to do the same mapping for openings and endgames, if my experiments were not that off. And as soon as you deviate from a "Texel engine at a particular time control and hardware", those too few parameters have to be changed. It is good for self-training and self-testing, but one can hardly generalize.
In an A/B engine the eval is a (usually) linear combination of features and in the Texel tuning method the coefficients are tuned in such a way that over a given large body of positions the resulting eval matches as well as possible the observed outcome of those positions (win, draw, loss) (via a logistic function, there are a number of variants, depending on the loss function, scoring of draws, etc...).
Instead one can also ask for "draw eval", a linear combination of features which matches as well as possible the draw/non-draw outcomes on the training positions. This can be tuned in the same way. Features that are likely relevant are game stage, king safety, passed pawns...