gnugo-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[gnugo-devel] arend_3_5.9: more influence breaking through


From: Arend Bayer
Subject: [gnugo-devel] arend_3_5.9: more influence breaking through
Date: Thu, 4 Jul 2002 14:35:03 +0200 (CEST)

 - bugfix in break_through_helper()
 - influence pattern tuning

The breakage is 6 PASSes and 3 FAILs compared to CVS.  trevorc:1480 is
acceptable, trevorc:1530 is another problem (missing reverse followup value
for A9), trevord:310 is the only somewhat serious FAIL, at which I will
look another time -- for now, I think the PASSes weigh considerably more
than those FAILs.

Arend


(Full results of CVS+this patch:)

./regress.sh . reading.tst
169 unexpected FAIL: Correct '0', got '1 F13'
./regress.sh . owl.tst
66 unexpected PASS!
./regress.sh . nngs1.tst
28 unexpected PASS!
./regress.sh . viking.tst
7 unexpected PASS!
./regress.sh . lazarus.tst
16 unexpected FAIL: Correct '!L17|J5|K5|K6', got 'L17'
./regress.sh . trevorb.tst
620 unexpected FAIL: Correct '!J10', got 'J10'
660 unexpected FAIL: Correct 'D3', got 'H5'
./regress.sh . strategy2.tst
73 unexpected PASS!
./regress.sh . nicklas1.tst
502 unexpected PASS!
./regress.sh . nicklas4.tst
1201 unexpected FAIL: Correct 'Q9', got 'S7'
./regress.sh . trevor.tst
15 unexpected PASS!
380 unexpected PASS!
./regress.sh . trevorc.tst
770 unexpected PASS!
890 unexpected FAIL: Correct 'J10', got 'G12'
1480 unexpected FAIL: Correct 'A8', got 'F6'
1530 unexpected FAIL: Correct 'A9', got 'G5'
./regress.sh . global.tst
46 unexpected PASS!
./regress.sh . arend.tst
17 unexpected PASS!
./regress.sh . 13x13.tst
65 unexpected PASS!
./regress.sh . semeai.tst
32 unexpected PASS!
37 unexpected PASS!
./regress.sh . trevord.tst
310 unexpected FAIL: Correct 'S17', got 'Q13'
./regress.sh . owl1.tst
268 unexpected FAIL: Correct '2 A1', got '0'


Index: engine/reading.c
===================================================================
RCS file: /cvsroot/gnugo/gnugo/engine/reading.c,v
retrieving revision 1.65
diff -u -r1.65 reading.c
--- engine/reading.c    27 Jun 2002 18:10:14 -0000      1.65
+++ engine/reading.c    4 Jul 2002 12:27:10 -0000
@@ -817,6 +817,13 @@
                return CUT;
              }
            }
+           else {
+             /* Lost everything. (Note we ignore ko at the moment.) */
+             popgo();
+             popgo();
+             popgo();
+             return 0;
+           }
            popgo();
          }
          popgo();
Index: patterns/barriers.db
===================================================================
RCS file: /cvsroot/gnugo/gnugo/patterns/barriers.db,v
retrieving revision 1.36
diff -u -r1.36 barriers.db
--- patterns/barriers.db        20 Jun 2002 22:32:22 -0000      1.36
+++ patterns/barriers.db        4 Jul 2002 12:27:13 -0000
@@ -1613,10 +1613,11 @@
 :8,B,value(30)

 ?Xx
-Qa.
+ba.
 o!X

 ;safe_omove(a)
+>return (!xplay_attack(a,b));


 Pattern Intrusion40
@@ -1679,7 +1680,7 @@
 ;lib(A)==2 && !attack(A) && !oplay_attack(b,c,d,d)


-Pattern Intrusion44
+Pattern Intrusion44a
 # gf New pattern. (3.1.20)

 .XQ
@@ -1696,6 +1697,26 @@
 >return (!xplay_attack(a,f) && (o_somewhere(e) || (!xplay_attack(e,f))));


+Pattern Intrusion44b
+# ab New pattern. (3.3.5)
+
+..X?
+.XQo
+!!o?
+
+:8,B,value(30)
+
+.dG?
+eFho
+cab?
+
+
+;safe_omove(a)
+;&& (o_somewhere(b) || oplay_attack(a,b,b))
+;&& !oplay_defend_both(a,c,d,e,F,G)
+>return (!xplay_attack(a,h));
+
+
 Pattern Intrusion45
 # gf New pattern. (3.1.20)

@@ -1808,7 +1829,7 @@
 ;!oplay_attack_either(a,b,c,c,d)


-Pattern Intrusion51
+Pattern Intrusion51a

 ?Q?
 X.X
@@ -1816,12 +1837,32 @@

 :|,B,value(30)

-?Q?
-cad
+?e?
+CaD
 .b.

-;oplay_break_through(a,b,c,b,d)
->return(!xplay_attack(a,b));
+;oplay_break_through(a,b,C,b,D)
+>return(!xplay_attack(a,e));
+
+
+Pattern Intrusion51b
+# ab New pattern. (3.3.5)
+# This is a bit incorrect, as usually X can choose on which side the
+# intrusion is allowed to happen.
+# Due to reading depth limitations, oplay_break_through == CUT is not
+# sufficient (see nngs1:40).
+
+?O?
+XQX
+!X!
+
+:|,BX,value(20)
+
+?O?
+aQc
+.b.
+
+;(oplay_break_through(a,b,c)==WIN)


 Pattern Intrusion52
@@ -1832,6 +1873,12 @@
 .X.

 :8,B,value(30)
+
+X.x
+ba.
+.X.
+
+>return (!xplay_attack(a,b));


 Pattern Intrusion53
Index: patterns/endgame.db
===================================================================
RCS file: /cvsroot/gnugo/gnugo/patterns/endgame.db,v
retrieving revision 1.34
diff -u -r1.34 endgame.db
--- patterns/endgame.db 3 Jun 2002 15:01:00 -0000       1.34
+++ patterns/endgame.db 4 Jul 2002 12:27:16 -0000
@@ -131,7 +131,7 @@
 ----

 ; marginal_eye(*) && proper_eye(a) && proper_eye(b)
-; && xplay_attack(*,a,c)
+; && xplay_attack(*,a,c) && !xplay_attack(*,c)


 Pattern EE2




reply via email to

[Prev in Thread] Current Thread [Next in Thread]