Ticket #48: gunnar_7_8.11.diff

File gunnar_7_8.11.diff, 9.8 kB (added by gunnar, 3 years ago)

Fixes for a few tests each in several files.

  • regression/9x9.tst 

    RCS file: /cvsroot/gnugo/gnugo/regression/9x9.tst,v
retrieving revision 1.21
diff -u -r1.21 9x9.tst
     
    216216 
    217217loadsgf games/bitti-gnug36.sgf 24 
    218218440 restricted_genmove black D6 C6 B7 
    219 #? [!D6]* 
     219#? [C6|B7]* 
    220220 
    221221# Not A8 as in the game 
    222222loadsgf games/bitti-gnug36.sgf 30 

  • regression/century2002.tst 

    RCS file: /cvsroot/gnugo/gnugo/regression/century2002.tst,v
retrieving revision 1.33
diff -u -r1.33 century2002.tst
     
    1919 
    2020# B15 and D13 are both locally better. C13 fills a dame. 
    2121loadsgf games/century21-2002/golois-gnugo.sgf 239 
    22 35 reg_genmove black 
    23 #? [!C13] 
     2235 restricted_genmove black B15 D13 C13 
     23#? [B15|D13] 
    2424 
    2525loadsgf games/century21-2002/goint-gnugo.sgf 37 
    262640 reg_genmove black  
    … …  
    117117#? [F6] 
    118118 
    119119loadsgf games/century21-2002/gnugo-katsunari.sgf 58 
    120 210 reg_genmove white 
    121 #? [!A15] 
     120210 restricted_genmove white A15 C16 
     121#? [C16] 
    122122 
    123123# a move in the center is bigger than this monkey jump 
    124124loadsgf games/century21-2002/gnugo-katsunari.sgf 76 

  • regression/dniwog.tst 

    RCS file: /cvsroot/gnugo/gnugo/regression/dniwog.tst,v
retrieving revision 1.12
diff -u -r1.12 dniwog.tst
     
    99#? [!S5] 
    1010 
    1111loadsgf games/dniwog.sgf 42 
    12 3 reg_genmove white 
    13 #? [!B5] 
     123 restricted_genmove white B5 B2 
     13#? [B2] 
    1414 
    1515loadsgf games/dniwog.sgf 62 
    16164 reg_genmove white 

  • regression/gunnar.tst 

    RCS file: /cvsroot/gnugo/gnugo/regression/gunnar.tst,v
retrieving revision 1.56
diff -u -r1.56 gunnar.tst
     
    2121 
    2222# L17 is horribly overconcentrated 
    2323loadsgf games/nngs/gnugo-3.3.10-rcde05-200210300235.sgf 27 
    24 4 reg_genmove white 
    25 #? [!L17] 
     244 restricted_genmove white L17 O18 
     25#? [O18] 
    2626 
    2727# G1 is the wrong way to defend G2. Both F2 and K2/K3 are locally better. 
    2828# But what's really urgent is a move around F12. 
    … …  
    3939 
    4040# O14 adds nothing at all to the position. 
    4141loadsgf games/nngs/niko-gnugo-3.3.11-200211201651.sgf 216 
    42 7 reg_genmove black 
    43 #? [!O14] 
     427 restricted_genmove black O14 M19 
     43#? [M19] 
    4444 
    4545# O13 is completely wasted. 
    4646loadsgf games/nngs/niko-gnugo-3.3.11-200211201651.sgf 222 
    47 8 reg_genmove black 
    48 #? [!O13] 
     478 restricted_genmove black M19 O13 
     48#? [M19] 
    4949 
    5050# Critical to keep black split up. 
    5151loadsgf games/gunnar/gunnar1.sgf 34 
    … …  
    113113# superior. A move around K16 may be even better to start with. 
    114114loadsgf games/nngs/gnugo-3.3.16-rubus-200302281805.sgf 17 
    11511521 restricted_genmove white P17 Q17 K16 K17 
    116 #? [!P17]* 
     116#? [Q17|K16|K17]* 
    117117 
    118118# Same problem as 21 but with slightly different surroundings. 
    119119loadsgf games/nngs/gnugo-3.3.16-rubus-200302281805.sgf 19 
    12012022 restricted_genmove white E17 D17 K16 K17 
    121 #? [!E17]* 
     121#? [D17|K16|K17]* 
    122122 
    123123# Must not play empty triangle at O5. N5 superior local shape. 
    124124loadsgf games/nngs/gnugo-3.3.16-rubus-200302281805.sgf 69 
    … …  
    361361# The game move at M1 is just weird. R2 is locally better than Q2. 
    362362loadsgf games/kgs/dogo-TheGNUGo.sgf 44 
    36336364 restricted_genmove black M1 Q2 R2 M5 R5 
    364 #? [!(M1|Q2)]* 
     364#? [R2|M5|R5]* 
    365365 
    366366# P19 is one point reverse sente. There are many larger moves. 
    367367loadsgf games/kgs/dogo-TheGNUGo.sgf 146 
    36836865 restricted_genmove black P19 Q12 B11 S8 
    369 #? [!P19]* 
     369#? [Q12|B11|S8]* 
    370370 
    371371# A10 was an overplay. Black should capture it and not fall back. 
    372372loadsgf games/kgs/dogo-TheGNUGo.sgf 162 
    … …  
    476476# B6 unreasonable. 
    477477loadsgf games/kgs/haiku-GNU.sgf 93 
    47847886 restricted_genmove black C8 B6 
    479 #? [R8]* 
     479#? [C8]* 
    480480 
    481481# O5 solves the problem. P4 backfires. 
    482482loadsgf games/kgs/melonhead1-GNU.sgf 127 

  • regression/kgs.tst 

    RCS file: /cvsroot/gnugo/gnugo/regression/kgs.tst,v
retrieving revision 1.22
diff -u -r1.22 kgs.tst
     
    77#? [C10]* 
    88 
    99loadsgf games/kgs/yagr2.sgf 65 
    10 30 reg_genmove black 
    11 #? [!P18|P17|O17] 
     1030 restricted_genmove black N19 P19 L18 P18 P17 O17 
     11#? [L18] 
    1212 
    1313loadsgf games/kgs/yagr3.sgf 78 
    141440 restricted_genmove white A7 A6 
    … …  
    9797# Don't kill your own group. 
    9898loadsgf games/kgs/yagr-ituyosi.sgf 160 
    9999200 restricted_genmove white K4 M4 L4 
    100 #? [!K4]* 
     100#? [M4|L4]* 
    101101 
    102102# An "owl attack captures tail" problem 
    103103loadsgf games/kgs/yagr-gab9.2.sgf 158 

  • regression/neurogo.tst 

    RCS file: /cvsroot/gnugo/gnugo/regression/neurogo.tst,v
retrieving revision 1.43
diff -u -r1.43 neurogo.tst
     
    2626#? [D6]* 
    2727 
    2828loadsgf games/FSF-neurogo.sgf 50 
    29 5 reg_genmove white 
    30 #? [!F8] 
     295 restricted_genmove white F8 F9 
     30#? [F9] 
    3131 
    3232# Tenuki is perhaps acceptable. 
    3333loadsgf games/FSF-neurogo.sgf 86 
    … …  
    4747#? [A18]* 
    4848 
    4949loadsgf games/FSF-neurogo.sgf 136 
    50 10 reg_genmove white 
    51 #? [!A3] 
     5010 restricted_genmove white A3 B4 
     51#? [B4] 
    5252 
    5353# Semeai problem. 
    5454loadsgf games/FSF-neurogo.sgf 148 
    … …  
    6262loadsgf games/FSF-neurogo.sgf 176 
    6363# This may be hard to fix at this point but since we 
    6464# can't win the ko we shouldn't start it. 
    65 13 reg_genmove white 
    66 #? [!P7] 
     6513 restricted_genmove white P7 M7 M8 
     66#? [M7|M8] 
    6767 
    6868loadsgf games/FSF-neurogo.sgf 178 
    696914 reg_genmove white 

  • regression/nngs.tst 

    RCS file: /cvsroot/gnugo/gnugo/regression/nngs.tst,v
retrieving revision 1.84
diff -u -r1.84 nngs.tst
     
    4040 
    4141 
    4242loadsgf games/nngs/AdaGeek-gnugo-3.1.18-200201070826.sgf 72 
    43 140 reg_genmove black 
    44 #? [!M3|M4] 
     43140 restricted_genmove black M3 M4 O11 
     44#? [O11] 
    4545 
    4646 
    4747loadsgf games/nngs/AdaGeek-gnugo-3.1.18-200201070826.sgf 80 
    … …  
    139139#New failure GNU Go 3.7.3 
    140140loadsgf games/nngs/AdaGeek-gnugo-3.1.18-200201071333.sgf 114 
    141141302 restricted_genmove black C12 F15 M7 B14 
    142 #? [!C12]* 
     142#? [F15|M7|B14]* 
    143143 
    144144 
    145145#New failure GNU Go 3.7.3 

  • regression/olympiad2004.tst 

    RCS file: /cvsroot/gnugo/gnugo/regression/olympiad2004.tst,v
retrieving revision 1.9
diff -u -r1.9 olympiad2004.tst
     
    11# Q12 might be playable after all but it's very unorthodox and it 
    22# would be preferrable if GNU Go played some more normal fuseki move. 
    33loadsgf games/olympiad2004/19x19/gnu-ind.sgf 5 
    4 1 reg_genmove black 
    5 #? [!Q12] 
     41 restricted_genmove black Q12 E3 E4 
     5#? [E3|E4] 
    66 
    77# G14 looks more natural than the played move at G12. 
    88loadsgf games/olympiad2004/19x19/gnu-ind.sgf 65 
    … …  
    6565 
    6666# L9 is zero points gote. 
    6767loadsgf games/olympiad2004/19x19/gnu-ind.sgf 223 
    68 12 reg_genmove black 
    69 #? [!L9]* 
     6812 restrcited_genmove black L9 F7 
     69#? [F7]* 
    7070 
    7171# A7 solves the tactical problem as well as B10 but also splits the 
    7272# eyespace into two parts. 
    … …  
    124124 
    125125# K4 is overconcentrated. Which moves more than D14 are good here? 
    126126loadsgf games/olympiad2004/19x19/int-gnu.sgf 42 
    127 106 reg_genmove white 
    128 #? [!K4] 
     127106 restricted_genmove white D14 K4 
     128#? [D14] 
    129129 
    130130# With the black stone at R9, R6 is locally better than Q6. 
    131131loadsgf games/olympiad2004/19x19/int-gnu.sgf 46 
    … …  
    264264# F6 really bad shape. 
    265265loadsgf games/olympiad2004/9x9/gnu-dbg.sgf 11 
    266266601 restricted_genmove black F7 F6 H4 G3 
    267 #? [!F6] 
     267#? [F7|H4|G3] 
    268268 
    269269# G6 about as bad shape as one can get. E6 has more promise. 
    270270loadsgf games/olympiad2004/9x9/gnu-dbg.sgf 23 

  • regression/strategy5.tst 

    RCS file: /cvsroot/gnugo/gnugo/regression/strategy5.tst,v
retrieving revision 1.48
diff -u -r1.48 strategy5.tst
     
    9191# Black has a combination attack which must be defended correctly against. 
    9292# In particular L5 fails. 
    9393loadsgf games/strategy48.sgf 
    94 274 reg_genmove white 
    95 #? [!L5|M5|N5] 
     94274 restricted_genmove white L5 M5 N5 M6 
     95#? [M6] 
    9696 
    9797# Semeai problem. White can make seki, black can kill. 
    9898loadsgf games/strategy49.sgf 
    … …  
    113113# Really, anything is better than A18. (B13 and B12 are quite 
    114114# ineffective too but can at least not be repeated...) 
    115115loadsgf games/reading41.sgf 130 
    116 279 reg_genmove black 
    117 #? [!A18] 
     116279 restricted_genmove black A18 A16 B13 B12 E11 
     117#? [A16] 
    118118 
    119119loadsgf games/nngs/takeda-gnugo-3.3.23-200307302010.sgf 10 
    120120280 reg_genmove black 

  • regression/tactics1.tst 

    RCS file: /cvsroot/gnugo/gnugo/regression/tactics1.tst,v
retrieving revision 1.19
diff -u -r1.19 tactics1.tst
     
    77# Apparently, T9 gets overvalued because the effect on S7 and T6 is 
    88# counted twice (territorially and strategically) 
    99loadsgf games/tactics05.sgf 
    10 10 reg_genmove white 
    11 #? [!T9] 
     1010 restricted_genmove white T9 T8 S9 S8 
     11#? [S8|S9] 
    1212 
    1313loadsgf games/tactics06.sgf 
    141420 owl_defend C19