HIARCS Chess Forums Forum Index HIARCS Chess Forums
World Championship winning computer chess software program & downloads for chess database, analysis and play on PC, Mac and iPhone
 
 QuestionsQuestions   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   RegisterRegister 
 ProfileProfile   Private MessagesPrivate Messages   Log inLog in 

engine problems with closed positions ?

 
Post new topic   Reply to topic    HIARCS Chess Forums Forum Index -> Programming Discussions
View previous topic :: View next topic  
Author
gsgs
Member


Joined: 26 May 2016
Posts: 37
Full Name: guenter stertenbrink
Location: hilter

PostPosted: Tue May 31, 2016 12:38 pm    Post subject: engine problems with closed positions ? Reply with quote

[closed positions = fixed pawn structure, pawn chains, no/few open lines]


I was looking for the opening with smallest draw-rate fr engines > 3000Elo and came up with :
1.d4 c5 . Unfortunately this is also good for white, with a score of ~60%, while the
normal score for white is only ~55%.
So it is rarely played by strong engines.
Now, it wasn't my "feeling" that Ben Oni is that bad, so I tried to examine a closed position
d4,c5,d5,e5,c4,d6,e4,Be7,Nc3,Bg5 with stockfish.

The score goes down from +0.68 at depth 10 to +0.47 at depth 29
(average for the 8 best moves)

It seems that the longer stockfish thinks about the position, the less favourable for white it considers it.

Similar for komodo. +0.50 at depth 10 , +0.35 at depth 21

it was suggested, that engines also underestimate the black kingside attacking chances
in the classical closed main variation of the King's Indian.


---------------------------------------------------------

Code:

echo off
echo uci
echo setoption name MultiPV value 8
echo position startpos moves d2d4 c7c5 d4d5 e7e5 c2c4 d7d6 e2e4 f8e7 b1c3 e7g5
echo go movetime 2000000
set xx=1
:xx4
set /A xx=%xx%+1
if %xx% EQU 10000 goto xx5
goto xx4
:xx5
echo quit



vbo3.bat | stockfish > vbo3.mo1
stockfish 7

depth , cp 1,2,3,4,5,6,7,8 , moves 1,2,3,4,5,6,7,8

10 , 67.9 , 93,68,68,68,68,68,60,50  ,  c1g5,f1e2,g1f3,f1d3,h2h3,g1e2,d1a4,d1b3
11 , 69.6 , 98,80,77,75,59,59,55,54  ,  c1g5,g1f3,h2h3,f1d3,f1e2,a2a3,d1a4,g1e2
12 , 64.4 , 98,72,67,65,59,54,53,47  ,  c1g5,f1d3,h2h3,a2a3,g1f3,g1e2,d1a4,f1e2
13 , 57.0 , 77,59,58,56,54,54,50,48  ,  c1g5,f1d3,a2a3,f1e2,h2h3,g1e2,d1a4,g1f3
14 , 60.2 , 74,67,65,57,57,56,53,53  ,  c1g5,f1d3,a2a3,f1e2,g1f3,g1e2,d1a4,h2h3
15 , 59.1 , 69,69,68,57,56,53,51,50  ,  c1g5,a2a3,f1d3,g1e2,h2h3,g1f3,f1e2,d1a4
16 , 56.9 , 73,60,58,55,53,53,52,51  ,  c1g5,f1d3,a2a3,d1c2,g1f3,f1e2,h2h3,g1e2
17 , 55.4 , 67,58,58,54,53,51,51,51  ,  c1g5,f1d3,a2a3,f1e2,f2f3,h2h3,d1c2,a1b1
18 , 55.0 , 62,58,56,56,53,53,51,51  ,  c1g5,c1d2,f1d3,a2a3,g1e2,f2f3,f1e2,a1b1
19 , 57.6 , 72,64,60,58,53,53,52,49  ,  c1g5,f1d3,a2a3,c1d2,g1e2,a1b1,f2f3,f1e2
20 , 55.5 , 68,60,59,54,53,53,50,47  ,  c1g5,f1d3,a2a3,b2b3,c1d2,f2f3,f1e2,d1c2
21 , 56.2 , 64,62,62,55,53,53,51,50  ,  c1g5,f1d3,a2a3,b2b3,c1d2,f2f3,f1e2,d1c2
22 , 52.9 , 61,57,57,54,52,48,48,46  ,  c1g5,f1d3,a2a3,d1c2,b2b3,f1e2,d1a4,g1f3
23 , 53.1 , 61,61,57,50,50,49,49,48  ,  c1g5,f1d3,a2a3,d1c2,b2b3,f1e2,d1a4,g1f3
24 , 51.5 , 60,54,51,51,51,50,48,47  ,  c1g5,f1d3,f2f3,a2a3,d1c2,b2b3,f1e2,g1e2
25 , 50.2 , 60,56,51,50,49,46,45,45  ,  c1g5,f1d3,a2a3,d1c2,f2f3,f1e2,b2b3,g1e2
26 , 50.5 , 59,56,52,51,47,47,46,46  ,  c1g5,f1d3,a2a3,b2b3,g1f3,c1d2,d1c2,f1e2
27 , 49.2 , 61,53,51,48,46,46,45,44  ,  c1g5,f1d3,c1d2,a2a3,b2b3,f1e2,g1f3,d1c2
28 , 47.6 , 56,55,47,47,45,44,44,43  ,  f1d3,c1g5,c1d2,f1e2,d1c2,a2a3,b2b3,g1f3
29 , 46.7 , 56,50,47,45,             ,  c1g5,f1d3,a2a3,c1d2




Code:

komodo 7

10 , 50.4 , 79,54,49,48,47,47,40,39  ,  c1g5,d1a4,g1f3,c1d2,f1d3,h2h3,f1e2,g1e2
11 , 50.1 , 81,60,51,47,46,43,37,36  ,  c1g5,f1d3,g1f3,g1e2,f1e2,h2h3,d1a4,c1d2
12 , 55.2 , 81,77,60,47,47,45,43,42  ,  d1a4,c1g5,f1d3,g1e2,h2h3,g1f3,a2a3,f1e2
13 , 48.2 , 69,54,51,45,44,43,42,38  ,  c1g5,f1d3,a2a3,g1e2,d1a4,f1e2,f2f3,g1f3
14 , 49.0 , 69,54,52,49,46,43,41,38  ,  c1g5,f1d3,a2a3,g1f3,d1a4,f1e2,g1e2,d1c2
15 , 44.5 , 64,49,45,42,41,41,40,34  ,  c1g5,f1d3,g1f3,f1e2,d1a4,g1e2,a2a3,c1d2
16 , 41.6 , 59,49,46,41,39,36,33,30  ,  c1g5,f1d3,d1a4,c1d2,a2a3,f1e2,g1f3,g1e2
17 , 34.7 , 49,41,39,33,31,30,29,26  ,  c1g5,d1a4,a2a3,c1d2,f1d3,g1e2,f2f3,d1c2
18 , 35.7 , 54,45,35,35,34,31,27,25  ,  c1g5,d1a4,f1d3,f2f3,g1e2,a2a3,a1b1,d1c2
19 , 34.6 , 51,38,37,34,33,30,28,26  ,  c1g5,f2f3,d1a4,f1d3,g1e2,a2a3,c1d2,f1e2
20 , 34.6 , 53,36,33,32,32,32,31,28  ,  c1g5,d1a4,g1f3,g1e2,f1d3,c1d2,f2f3,a2a3
21 , 34.9 , 52,36,36,33,32,31,30,29  ,  c1g5,d1a4,f1e2,g1f3,f1d3,a2a3,g1e2,c1d2




maybe someone can extent this, to see to which value it cnverges
or maybe someone can do the same for the King's Indian, see above
Back to top
View user's profile Send private message Visit poster's website AIM Address
gsgs
Member


Joined: 26 May 2016
Posts: 37
Full Name: guenter stertenbrink
Location: hilter

PostPosted: Thu Jun 02, 2016 8:41 am    Post subject: Reply with quote

maybe this is a more general phnomenon, not limited to closed positions,
just positions with many good moves.
I observed it in a few non-closed positions too.
This is probably already well known, but noone bothers to tell it ...
There could be a name for it ...
increased-depth-analysis-computer-chess-advantage-lowering-phenomenon
idaccalp

die berechnete Groesse des Vorteils ist reziprok proportional zur Rechentiefe
Back to top
View user's profile Send private message Visit poster's website AIM Address
Display posts from previous:   
Post new topic   Reply to topic    HIARCS Chess Forums Forum Index -> Programming Discussions All times are GMT
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum


Powered by phpBB © 2001, 2005 phpBB Group
Protected by Anti-Spam ACP