Author: martin fierz
Date: 07:49:55 04/08/04
Go up one level in this thread
On April 08, 2004 at 09:38:01, Daniel Shawul wrote: >On April 08, 2004 at 09:14:05, martin fierz wrote: > >>On April 08, 2004 at 07:10:29, Daniel Shawul wrote: >> >>>Hi >>>Here are some questions >> >>[snip] >> >>>5.How do i measre my move ordering performance?Is it by counting only fail highs >>>of the first moves searched. It seems to me doing this measures performance of >>>hashtable since the hash move is the first searched. Or is it counting fail >>>highs of first x moves?? >> >>it's not measuring hashtable performance. you have more positions without hash >>moves than with hash moves... > > ok that's right. What I am getting is 16.7% fail highs at the first move >searched. I have heard many people report 80-90% move ordering efficiency. >Are they counting fail highs at 2nd,3rd,... moves too? My efficiency seems to >be too low compared to what i do in ordering. 16.7% is horrible - if you count the same way as everybody does: you compute FH-ratio as 100* (#fail highs on first move) / (total fail highs). so you don't count all cases where you don't fail high, which is sensible since you search all moves there anyway - it doesn't matter how you order there. the number as defined above should be ~90% or better, or your move ordering is poor. if that doesn't sound right to you, remember that this does NOT mean that you find the best move first in 90% of all cases, only that you find "a good move" first in 90% of all cases - often many moves lead to cutoffs. cheers martin > >> >>cheers >> martin
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.