Author: Stephen A. Boak
Date: 23:16:16 02/04/01
Steve,
I began to send a private eMail, but I think many Fritz6 users may have the
same questions.
When analyzing an OTB game I have played, entering it or playing it back
move by move, I normally set Fritz6 to show the 3 top variations.
1. Why doesn't a 2nd or 3rd best move, for example, get sorted to show it
has become the 1st best move, after its eval has sufficiently improved?
Many times a changed eval score shows the move should move up in the
multi-variation ranking, but it does not--neither during the ply analysis (when
the score changes), nor when the analysis of that ply is completed.
I haven't checked this yet, to see if the 'mis-sorting' affects the reported
best move/eval in an overnight Game Analysis, but I wonder...
Two examples:
After 1. e4 c6 2. Nc3 d5 3. Qf3:
rnbqkbnr/pp2pppp/2p5/3p4/4P3/2N2Q2/PPPP1PPP/R1B1KBNR b KQkq - 0 1
Analysis by Fritz 6 (after 16/44 plies completed):
1. = (0.06): 3...e6 4.e5 Nd7 5.Qg3 Ne7 6.Nf3 Nf5 7.Qg4 d4 8.Ne2
2. = (0.09): 3...dxe4 4.Nxe4 Nd7 5.d3 Ndf6 6.Be2 Nxe4 7.dxe4 Nf6 8.Qe3
3. = (0.00): 3...d4 4.Bc4 e6 5.Nce2 b5 6.Bb3 a5 7.c3 a4 8.Bc2
1A. Why didn't Black's 3...d4 move (Variation 3) get promoted to become the 1st
(best) variation? It shows a 0.00 eval (even score) whereas the other two
variations show a slight plus score for white.
Continuing the same game with:
3...dxe 4. Nxe4:
rnbqkbnr/pp2pppp/2p5/8/4N3/5Q2/PPPP1PPP/R1B1KBNR b KQkq - 0 1
Analysis by Fritz 6 (after 16 ply):
1. = (0.09): 4...Nd7 5.d4 Ndf6 6.c3 Nxe4 7.Qxe4 Nf6 8.Qe2 Bg4 9.f3
2. = (0.00): 4...Nf6 5.Nxf6+ exf6 6.Bc4 Nd7 7.Be2 Bd6 8.Qe3+ Be7
3. = (0.00): 4...e6 5.Ne2 Nd7 6.N2g3 Ngf6 7.d4 h5 8.Bg5 h4 9.Ne2
1B. Why didn't both Black's moves in Variation 2 and Variation 3 get promoted
above Variation 1, since both show an even score, whereas Variation 1 is
slightly favorable for White?
1C. Is there something intentional in the programming that makes a slightly
inferior move be ranked as best?
2. Does it have anything to do with odd/even ply results? I've read that there
is something suspect about even-ply results that may lead one to avoid using the
even-ply analysis results. Would you please summarize the odd/even ply issue,
also, so I understand it better?
3. Does this noted problem affect overnight Game Analysis? [I may try these two
exact positions, with a mandated n-ply search depth to see].
I noticed this mis-sorting behavior with my original Fritz6 right off the CD, as
well as with Fritz6a, and my current Fritz6'e' (English) version with engine
date of Jan 23 2001.
Just got around to sending in the question.
Thanks in advance for your reply to this puzzling Fritz6 behavior.
--Steve Boak
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.