Author: José Carlos
Date: 17:54:15 08/05/04
Go up one level in this thread
On August 05, 2004 at 20:01:29, Cesar Contreras wrote: >I'm trying to improve my move ordering, with a little success, principal due to >correct my understanding and implementation MVA\LVV. I had not sucess with SEE. >Anyway, i think my tree it's still too big. > >I need to know if i can do any more about it. > >currently i do: > >1.- Moves on transposition, ordered by evaluation (some one suggested me order >by node count, why do you think?) I guess he meant _at the root_. In the root position of the search, you want to try the PV move first (because your best move so far) and then all the rest, which you expect to quickly fail low. As the world isn't perfect, you'll fail high on a non pv move sometimes. Due to alphabeta nature, you want that to happen as soon as possible. Node count is a good general idea, because more nodes means "in the last iteration, it was harder to refute (fail low) this move, so it might be good". I use node count, but I've found better (for my program) to try previous pv moves before. For example, in the initial position, e4 is your pv move until iteration 7. Then you fail high on d4. At iteration 10 you fail high on Nf3. In this case, for iteration 11, I first search Nf3, second I search d4, then e4, and after that, all the other moves, sorted by node count. >2.- Killer moves. >3.- History heuristics. >4.- All promotion moves. >5.- All capture moves. Try winning captures and promotions before killers and history. A winning capture is one with a positive SEE or, if you don't do SEE on internal nodes, one where the captured piece is bigger than the capturer. Same for promotions. This way, you can skip these moves from killers, and store more information in them. Winning captures are very likely to fail high (if the node is a fail high node, of course). >6.- All remaining non-capture-non-promotion moves. And now, the losing captures and promotions. >What i see on this list it's that the last "non-capture-non promotion moves", >comes last, but i think it's to frecuently that the best move it's not a capture >or promotion move. I also do (in my private program, not in Averno) escapes, which I define as "a piece which is attacked by a smaller piece moves to a safe square". These moves are sorted after killers (I don't do history). >I'm thinking of make some Mix, first try the best promotion move, then the best >capture move, and then the best non-capture-non-promotion move, then the next >best promotion move, and so on. > >What do you think about it? Experiment with my proposals and then, come up with the order that works best for you. That's the best I can say. José C.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.