| 1 | |
| 2 | Some remarks concerning the files collected: |
| 3 | |
| 4 | seki1.sgf |
| 5 | |
| 6 | The "classical" seki. gnugo plays correctly, but the score is wrong. |
| 7 | |
| 8 | seki2.sgf |
| 9 | |
| 10 | Each side has an eye |
| 11 | |
| 12 | seki3.sgf |
| 13 | |
| 14 | Another one. |
| 15 | |
| 16 | seki_stupid.sgf |
| 17 | |
| 18 | Here i wanted to see, if gnugo plays the stupid move C 1. Well it doesnt. |
| 19 | But maybe it is good to keep the file, just to make sure that future |
| 20 | "improvements" dont spoil it. |
| 21 | |
| 22 | sekichallenge.sgf |
| 23 | |
| 24 | Thats what it is. gnugo makes mistakes and loses everything. But i get |
| 25 | the impression that the reading depth is just a little bit too short. |
| 26 | |
| 27 | seki_score.sgf |
| 28 | |
| 29 | a seki. unless i counted it wrongly, assuming an equal number of |
| 30 | prisoners |
| 31 | the japanese rules give: |
| 32 | |
| 33 | territory b 96 w 97 |
| 34 | prisoners b 0 w 0 |
| 35 | seki b 0 w 0 |
| 36 | |
| 37 | the chinese rules give: |
| 38 | |
| 39 | territory b 96 w 97 |
| 40 | stones b 79 w 79 |
| 41 | seki b 8.5 w 1.5 |
| 42 | |
| 43 | so there is a difference of 7 points, depending on which rules you take. |
| 44 | |
| 45 | twoko_seki1.sgf |
| 46 | |
| 47 | A double ko seki. with the present program and playing japanese rules |
| 48 | the eternal loop can only be avoided, if both players have enough |
| 49 | intelligence to see that it is impossible to win. i said already |
| 50 | before that i think that the program is wrong. that is after two |
| 51 | passes, i thought the game is finished, but do_pass allows gnugo |
| 52 | to play another move, so the loop continues. if we change do_pass, |
| 53 | it would be good enough to have one intelligent player to avoid |
| 54 | the loop. Since we know nothing about the intelligence of our |
| 55 | opponent, the intelligent player would have to be gnugo. |
| 56 | So if we stick to japanese rules, this needs a detailed analysis |
| 57 | of the situation. |
| 58 | In principle the japanese rules define in their precedents |
| 59 | that this position is seki, should gnugo know the precedents ? |
| 60 | chinese rules avoid the loop and give the interesting result, that |
| 61 | whoever insists on playing, loses everything. maybe not everybody |
| 62 | knows that already by heart, so it goes like this: |
| 63 | |
| 64 | w takes ko 1 |
| 65 | b takes ko 2 |
| 66 | w tenuki |
| 67 | b takes ko 1 |
| 68 | w takes ko 2 |
| 69 | b tenuki |
| 70 | |
| 71 | finally there will be no playable point left on the board. |
| 72 | lets assume that b has played the last possible move. |
| 73 | then if w insists: |
| 74 | |
| 75 | w takes ko 1 |
| 76 | b takes ko 2 |
| 77 | w passes |
| 78 | b takes ko 1 |
| 79 | |
| 80 | now w would like to take ko 2, but that is forbidden, because it |
| 81 | reproduces a position, which has already existed before. |
| 82 | so |
| 83 | |
| 84 | w passes |
| 85 | b takes all. |
| 86 | |
| 87 | If there is another ko somewhere else on the board, then this |
| 88 | double ko would be an example of what Ing calls a disturbing ko. |
| 89 | |
| 90 | |
| 91 | twoko_seki2.sgf |
| 92 | |
| 93 | Another double ko seki. twoko_seki1 has two groups with an eye each, |
| 94 | this one has one liberty common to both groups. |
| 95 | |
| 96 | twoko_live.sgf |
| 97 | |
| 98 | B is alive. |
| 99 | |
| 100 | |
| 101 | twoko_dead.sgf |
| 102 | |
| 103 | W is dead. |
| 104 | |
| 105 | |
| 106 | fourko.sgf |
| 107 | |
| 108 | A quadruple ko. I have never heard about that, but i guess the |
| 109 | japanese will have to cancel the game. with chinese rules one |
| 110 | can play it of course, but it is rather tricky. |
| 111 | |
| 112 | onadare_ko.sgf |
| 113 | |
| 114 | The double ko variation of the onadare joseki. |
| 115 | |
| 116 | onadare_ko2.sgf |
| 117 | |
| 118 | Here i have added a ko in the lower right corner. This is completely |
| 119 | different from twoko_seki1.sgf plus another ko. here the black group |
| 120 | in the double ko is alive. however white has an unlimited supply |
| 121 | of ko threats, so it will win any other ko. If i understand this |
| 122 | correctly, the japanese do not consider this to be a triple ko |
| 123 | I thought that Ing should call the double ko a "fighting ko", |
| 124 | but apparently he doesnt. So the Ing rules would have to treat |
| 125 | the two cases twoko_seki1.sgf plus another ko and the present one |
| 126 | the same way. |
| 127 | gnugo does not know how to handle this. |
| 128 | |
| 129 | bent4.sgf |
| 130 | |
| 131 | gnugo handles this correctly. when i said maybe for the wrong reasons |
| 132 | Gunnar said, what do i mean. well first one should look, if it does |
| 133 | all possible variations correctly. But what i really was afraid of, was |
| 134 | maybe it treats it the same way it handles seki. only for seki it is |
| 135 | wrong and here it is right. anyway, whenever the seki has been fixed, |
| 136 | one should look, if this is still ok. |
| 137 | |
| 138 | bent4_2.sgf |
| 139 | |
| 140 | Another one. This one is also handled correctly. |
| 141 | |
| 142 | |
| 143 | seki_bent4.sgf |
| 144 | |
| 145 | The combination of a bent4 with a seki, which gives an unremovable |
| 146 | ko threat. The japanese rules kill the bent4 anyway. gnugo has |
| 147 | various difficulties. |
| 148 | |
| 149 | |
| 150 | doublebent4.sgf |
| 151 | |
| 152 | I guess the japanese rules say that both groups are dead ??? |
| 153 | Apart from that gnugo handles it correctly. |
| 154 | |
| 155 | threeko1.sgf |
| 156 | |
| 157 | The "classical" triple ko. i guess, if we play japanese rules, |
| 158 | gnugo should realise that the game should be cancelled. |
| 159 | Playing chinese rules one can usually neutralise two ko's |
| 160 | and then play the one which is remaining as before. |
| 161 | |
| 162 | threeko2.sgf |
| 163 | |
| 164 | This is a fake. white can never win this. |
| 165 | |
| 166 | threeko3.sgf |
| 167 | |
| 168 | Similar to threeko2.sgf, with a slight difference. If it is |
| 169 | white's turn it can play B 5 and then it becomes a double ko |
| 170 | which black cannot win. |
| 171 | If it is black's turn, the japanese will cancel the game and |
| 172 | for the chinese white can never win. |
| 173 | |
| 174 | chosei.sgf |
| 175 | |
| 176 | A chosei( long life / eternal life ). |
| 177 | Apart from the fact that the game should be cancelled, |
| 178 | gnugo makes a lot of mistakes and loses everything. |
| 179 | It seems everybody who writes about that, shows the same diagram. |
| 180 | Maybe this is the only one ? |
| 181 | If people think that this is interesting, I would have to do |
| 182 | some work. |
| 183 | |
| 184 | junkan-ko.sgf |
| 185 | |
| 186 | A junkan-ko (round robin ko). Everything I wrote about chosei |
| 187 | is valid here as well. i thought a little about these positions, |
| 188 | but i dont want to bore everybody, unless somebody really wants |
| 189 | to hear about it. |