Some searching bug

[NelloHo]

look at this xiangqi fen:

4k4/4a4/3a2N1C/9/9/9/9/3A4B/9/2BAK4 w

through puzzler.py it's #5:

pv i8i4,e10d10,i4d4,d10e10,i3g5

but in pychess analysis, it's #2:

this kind of case has happend several times when I load puzzler.py on fens

[Fabian Fichter]

Likely that is just a matter of depth. Which depth did you use, and can you reproduce it at higher depth?

[NelloHo]

these are all #2 puzzles, so I think -d > 4 will be sufficient.
previously I use -d = 8 , now I use -d = 12 with -r = 3 , the problem is still there, also the results are different respectively , which is very strange.

[Fabian Fichter]

Probably same as #30.