Article View: gnu.chess
Article #5184Checkmate eval problem
From: Dr Michael J. White
Date: Sun, 10 Apr 1994 23:36
Date: Sun, 10 Apr 1994 23:36
109 lines
4775 bytes
4775 bytes
Hello!
This position shows that counting behavior I mentioned in an earlier
post. It's not the same position that I mentioned, but this one is weird
enough. The fix from Kong Sian gets rid of the funny evaluation sequence.
Now, no mating lines show at all in the evals (I don't know if that
fact is bad or good). The same final move is chosen in both cases.
A small detail is that I had HISTORY off in the first run, but HISTORY
was on in the second...
The fix from Kong was in search.c:
Change this line:
if (flag.hash /* && ply <= Sdepth */ && *rpt==0 ...etc...
to:
if (flag.hash && SAVEHT && *rpt==0 ...etc...
Here's the position, before and after output.
Have fun,
Mike
THE POSITION:
----------------------------------------
1 1 post book force
e2e4 g8f6 e4e5 f6d5 d2d4 d7d6 f1c4 d5b6 c4b3 c8f5 d1f3 d8c8 g1e2 e7e6 b1c3
d6e5 d4e5 b8d7 e2d4 f5g6 h2h4 c7c5 h4h5 c5d4 h5g6 f7g6 c3b5 f8b4 e1d1 d7e5
f3e2 c8c5 c1f4 b6c4
depth 6 bd go quit
R - - - K - - R
P P - - - - P P
- - - - P - P -
- n Q - N - - -
- B N P - b - -
- b - - - - - -
p p p - q p p -
r - - k - - - r
BEFORE IMPLEMENTING "&& SAVEHT" :
------------------------------------------------------------
Move# 18 Target= 0 Clock: 1019845
2. -108 0 2131 f4e5 c4e5 h1h7
3& -69 1 12622 f4e5 c4e5 b3e6 a7a6 a2a4
3+ -11 2 17078 b3c4 c5c4 e2c4 e5g4 c4b4 g4f2 d1c1
3& 255 3 25114 b3c4 c5c4 f4e5 c4e2
3. 255 4 30676 b3c4 c5c4 f4e5 c4e2
4& 297 8 54214 b3c4 c5c4 f4e5 d4d3 e2d3 a8d8 b5d4 c4d3
c2d3
4. 297 12 78587 b3c4 c5c4 f4e5 d4d3 e2d3 a8d8 b5d4 c4d3
c2d3
5& 298 17 113527 b3c4 c5c4 f4e5 d4d3
5. 298 19 125766 b3c4 c5c4 f4e5 d4d3
6- 157 403 3007551 b3c4 e5c4 e2e3
6- 157 415 3098430 f4e5 c4e5 b5a3
6- 161 419 3120227 b5d4
6- 162 425 3158645 h1h4
6- 162 556 4138759 e2e3
6& -9979 1667 12327536 e2e3 c4e3 d1c1 c5c2 b3c2
6& -9975 1925 14026014 h1h4 c4b2 d1c1
6& -9971 2046 15137962 b5d4 c4b2 d1c1
6& -9967 2371 17312422 f4e5 c4b2 d1c1
6& 169 2411 17576384 b3c4 h8g8
6. 169 2436 17741788 b3c4 h8g8
1. ... b3c4
Gen 64221213 Node 17741788 Tree 1056 Eval 756197 Rate 7283 EC 634078/121032 RS hi 1 lo 2
Hin/Hout/Tcol/Coll/Fin/Fout = 2204454/184832/391/0/0/0
My move is: b3c4
AFTER IMPLEMENTING "&& SAVEHT":
------------------------------------------------------------------------
Your move is?
Move# 18 Target= 0 Clock: 101847
2. -108 0 2095 f4e5 c4e5 h1h7
3& -69 1 12980 f4e5 c4e5 b3e6 a7a6 e2c4
3+ -6 2 16598 b3c4 e5d7 h1h7 h8h7 c4e6
3& 310 4 28994 b3c4 c5c4 f4e5 c4e2 d1e2 d4d3 e2d3 e8g8
3. 310 5 35634 b3c4 c5c4 f4e5 c4e2 d1e2 d4d3 e2d3 e8g8
4& 313 7 52090 b3c4 c5c4 f4e5 d4d3 e2d3 c4g4 d3f3 g4f3
g2f3
4. 313 11 75691 b3c4 c5c4 f4e5 d4d3 e2d3 c4g4 d3f3 g4f3
g2f3
5& 314 16 104920 b3c4 c5c4 f4e5 d4d3
5. 314 17 116116 b3c4 c5c4 f4e5 d4d3
6- 171 186 1362647 b3c4 e5c4 e2e3
6- 171 198 1446589 f4e5 c4e5 e2e3
6- 174 202 1465677 b5d4
6- 175 206 1490860 b5a7
6- 176 207 1500850 h1h4
6- 176 224 1608229 e2e3
6& -1538 22395 17793175 e2e3 d4e3 a2a4 e3f2 d1e2 c4a3 b2a3 f2f1
a1f1 c5b5 a4b5
6& -412 23174 23041665 h1h4 c5b5 d1c1 b5c5 e2e4 e8c8 f4e5 c4e5
b3e6 c8b8
6& -35 23562 25577787 f4e5 d4d3 c2d3 c4e5 d3d4 a8d8 b3e6 e8f8
6& 183 23581 25701602 b3c4 c5b6
6. 183 23622 25964357 b3c4 c5b6
1. ... b3c4
Gen 92572985 Node 25964357 Tree 1041 Eval 1272710 Rate 1099 EC 1003330/268330 RS hi 1 lo 2
Hin/Hout/Tcol/Coll/Fin/Fout = 3423531/327298/2347/0/0/0
My move is: b3c4
------------------------------------
THE END
Technical Details
Message-ID:
<2oa2ia$nlr@vixen.cso.uiuc.edu>Path:
arc1-ca.pugleaf.net!gmd.de!Germany.EU.net!EU.net!howland.reston.ans.net!vixen.cso.uiuc.edu!osiris.cso.uiuc.edu!mwhiteShow Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 10d71d,fbe9e17015bd0101,start X-Google-Attributes: gid10d71d,public X-Google-ArrivalTime: 1994-04-10 16:36:42 PST Path: arc1-ca.pugleaf.net!gmd.de!Germany.EU.net!EU.net!howland.reston.ans.net!vixen.cso.uiuc.edu!osiris.cso.uiuc.edu!mwhite From: mwhite@osiris.cso.uiuc.edu (Dr Michael J. White) Newsgroups: gnu.chess Subject: Checkmate eval problem Date: 10 Apr 1994 23:36:42 GMT Organization: University of Illinois at Urbana Lines: 109 Message-ID: <2oa2ia$nlr@vixen.cso.uiuc.edu> NNTP-Posting-Host: osiris.cso.uiuc.edu