Re: [gnugo-devel] atari_atari problems

[Trevor Morris]

Yes, I'd already deleted that problem locally - I would consider
L4 a success of atari_atari now.  Yes, there are some complicated
variations, but not worth regressing over.

-Trevor


At 05:56 PM 12/19/2001 +0100, you wrote:
>Trevor wrote:
>> I believe Gunnar said he's working on some bugs in atari_atari.
>> Here are some good problems for that:
>> 
>> trevorb:780
>> trevorb:800
>> trevorb:880
>> 
>> As you can see from these, the current atari_atari is broken pretty badly.
>
>I think trevorb:800 is incorrect, or at least not worth regressing
>over. White certainly loses the game after playing L4 so considering
>that J4 gives more potential for black mistakes, it may be the best
>move. However, if black plays correctly, it seems to me like L4 is
>bigger than J4. Of course I may have missed some variation and there
>is a ko complicating matters, but I don't think the ko is good enough
>for white. See sgf below for the main variations in my analysis.
>
>(;GM[1]FF[3]
>SZ[13]HA[0]KM[5.5];B[dc];W[jj];B[cf];W[ke];B[dj];W[jc];B[hk];W[kh];
>B[jl];W[kk];B[kl];W[gc];B[fj];W[ck];B[cj];W[dk];B[ek];W[bj];B[bi];
>W[el];B[fl];W[dm];B[fm];W[bk];B[aj];W[bm];B[fd];W[ll];B[gd]C[!E12];
>W[eb];B[db]C[!H10];W[hd];B[fb];W[gb];B[hi];W[fc];B[gf]C[!E10];W[ed];
>B[fe]C[E11];W[if];B[hg]C[E11];W[da];B[ec];W[fa];B[cb];W[ak];B[lm];
>W[ai];B[bh];W[ml];B[he]C[J9];W[ea];B[id];W[hc];B[ig];W[ie];B[ca];W[jg]
>;B[hf];W[ih];B[em];W[dl];B[ii]C[K6|K5];W[hh];B[ji];W[ki];B[jh];W[je];
>B[jk]C[J4]
>(;W[kj];B[ij];W[ah];B[ag];W[aj];B[bg];W[lk];B[km];W[mm];B[tt];W[tt]
>TW[ga][ha][ia][ja][ka][la][ma][fb][hb][ib][jb][kb][lb][mb][ic][kc][lc][mc][id][jd][kd][ld][md][le][me][jf][kf][lf][mf][kg][lg][mg][lh][mh][li][mi][lj][mj][mk][al][bl][cl][am][cm]
>TB[aa][ba][ab][bb][ac][bc][cc][ad][bd][cd][dd][ed][ae][be][ce][de][ee][ge][af][bf][df][ef][ff][cg][dg][eg][fg][gg][ch][dh][eh][fh][gh][hh][ih][ci][di][ei][fi][gi][ej][gj][hj][fk][gk][ik][gl][hl][il][gm][hm][im][jm]
>C[The game is over.  Final score:
>   White = 44 territory + 3 captures + 5.5 komi = 52.5
>   Black = 53 territory + 3 captures = 56
>Black wins by 3.5.
>])
>
>(;W[ij];B[lk]
>(;W[kj];B[mk];W[ik];B[hl];W[il];B[im];W[hj];B[gi];W[gj];B[gk];W[lj];
>B[mm];W[ah]
>C[This might not be sente, but let's assume it is for simplicity.
>];
>B[ag];W[aj];B[bg];W[kg];B[mj];W[mi];B[tt];W[tt]
>TW[ga][ha][ia][ja][ka][la][ma][fb][hb][ib][jb][kb][lb][mb][ic][kc][lc][mc][id][jd][kd][ld][md][le][me][jf][kf][lf][mf][lg][mg][lh][mh][li][al][bl][cl][am][cm]
>TB[aa][ba][ab][bb][ac][bc][cc][ad][bd][cd][dd][ed][ae][be][ce][de][ee][ge][af][bf][df][ef][ff][cg][dg][eg][fg][gg][ch][dh][eh][fh][gh][hh][ih][ci][di][ei][fi][ej][fk][gl][ll][ml][gm][hm][jm][km]
>C[The game is over.  Final score:
>   White = 39 territory + 3 captures + 5.5 komi = 47.5
>   Black = 48 territory + 5 captures = 53
>Black wins by 5.5.
>])
>
>(;W[lj]C[Ko, but does white really have sufficient threats?
>])
>)
>
>)
>
>_______________________________________________
>gnugo-devel mailing list
>address@hidden
>http://mail.gnu.org/mailman/listinfo/gnugo-devel
>