Author: Robert Hyatt
Date: 12:50:57 09/25/98
Go up one level in this thread
On September 24, 1998 at 21:36:30, John Coffey wrote: >On September 24, 1998 at 21:12:46, Robert Hyatt wrote: > > >>You are overlooking a serious issue... the tree search grows *exponentially* >>with depth. Which means that any effort expended to order the best move >>first at any ply will greatly pay off in reduced tree search space. No >>matter how deep in the tree it is, capture or regular search. With alpha/ >>beta, move ordering is not just an issue, it is *the* issue. You can >>*easily* make your tree 10X larger with sloppy move ordering, and you will >>be faster than everyone in terms of NPS, but you will be searching 2-3 >>plies less deeply than everyone else. It is the *depth* and not the *NPS* >>that is important... > > >No I understood this issue. I am just asking if there is a dimished return, >say in an N ply search the last 2 to 3 ply (N-1, N-2 or even N-3) then the move >ordering would not seem to me to be worth the cost. > >john think about this: if we only take the last ply before the q-search, and we sort moves into worst-to-best order, rather than best-to-worst... what does it do? It makes your search W times slower, where W is the average number of moves at that depth. IE 30-40. This is what happens if you only misorder one ply (and I am assuming this ply is not an "ALL" ply where ordering is useless). If you screw up at two plies from the Q-search, *and* at one ply from the q-search, your tree is 30^2 times bigger. So *yes* it is important to order *everywhere* even in the capture search.
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.