[Bug-gnubg] Snowie error rate bug?

[Scott Steiner]

Here are the overall stats of an imported and analysed fibs match:

                                      Opponent                You
Overall error rate (total)        -3,658 (-31,893%)    -2,907 (-19,647%)
Overall error rate (per decision) -0,018 (-0,157%)     -0,016 (-0,105%)
Equivalent Snowie error rate      -11,7                -12,3

MWC against current opponent     37,75%               62,25%

As you can see, my overall error rate is better than that of my
opponent.  However, the equivalent snowie error rate rates me worse.

Attached is the original fibs match in JF format and the imported and
analysed match in sgf format.

snowie_error_rate_bug.zip
Description: Zip compressed data