[gnugo-devel] ladder issue

[Portela Fernand]

The appended game features an interesting phenomenon : GnuGO playing out a
ladder for about 20 moves. Important to notice is the time taken by GG for
playing the ladder (moves 31 to 59). It seems, such kind of positions really
"kill" the engine.

Else, I have no clue what is happening here. I hope someone will shed some
light...

/nando

[For the record, the black player is the genuine 3.3.16, while the white
 one is a heavily patched mutant.]

(;GM[1]FF[4]
SZ[19]
PB[GNU Go 3.3.16 (random seed 462287504)]
PW[GNU Go 3.3.16 (random seed 462287504)]
RU[Japanese]HA[0]KM[5.5]RE[W+51.5]
;B[qd];W[dc];B[dp];W[oc];B[pe];W[lc];B[nd];W[pp];B[de];W[fd];B[cc]
;W[cd];B[dd];W[bc];B[cb];W[bb];B[db];W[ec];B[bd];W[ce];B[be];W[cf]
;B[bf];W[bg];B[cg];W[df];B[ef];W[dg];B[ch];W[bh];B[dh];W[eg];B[fg]
;W[eh];B[ei];W[fh];B[gh];W[fi];B[fj];W[gi];B[hi];W[gj];B[gk];W[hj]
;B[ij];W[hk];B[hl];W[ik];B[jk];W[il];B[im];W[jl];B[kl];W[jm];B[jn]
;W[km];B[lm];W[kn];B[jo];W[ej];B[ko];W[di];B[ln];W[ci];B[hg];W[hm]
;B[qj];W[kk];B[ih];W[in];B[qn];W[cn];B[nq];W[pn];B[pm];W[on];B[qo]
;W[qp];B[kq];W[oq];B[oj];W[np];B[hp];W[ic];B[fp];W[bp];B[cq];W[ck]
;B[nr];W[rp];B[bq];W[or];B[rn];W[nc];B[pb];W[gf];B[bo];W[li];B[jj]
;W[oh];B[kh];W[mh];B[qh];W[ll];B[le];W[lo];B[mo];W[mp];B[lp];W[mq]
;B[je];W[bn];B[md];W[mc];B[cp];W[ob];B[pc];W[fn];B[gg];W[kd];B[hd]
;W[hc];B[ff];W[lr];B[kr];W[pa];B[qa];W[oa];B[gd];W[gc];B[pg];W[qb]
;B[qc];W[rb];B[rc];W[ks];B[js];W[ls];B[jr];W[nj];B[nk];W[mk];B[ni]
;W[mj];B[nl];W[oi];B[pi];W[ee];B[fe];W[jd];B[sb];W[ra];B[id];W[ke]
;B[jf];W[go];B[gp];W[ro];B[sn];W[kj];B[ki];W[no];B[an];W[am];B[ao]
;W[mn];B[mm];W[lo];B[nm];W[om];B[ol];W[po];B[qm];W[od];B[oe];W[lq]
;B[kp];W[eo];B[co];W[bm];B[io];W[ho];B[ep];W[do];B[og];W[ng];B[nf]
;W[mf];B[ph];W[nh];B[lg];W[lf];B[kf];W[ld];B[ne];W[me];B[mg];W[le]
;B[lh];W[mi];B[so];W[sp];B[mo];W[nn];B[lo];W[fo];B[pd];W[ml];B[sc]
;W[qa];B[sa];W[];B[])