Run engine tests with no opening books.

[Cubeman]

Just looking at that position RKNBNQBR, I can't see anything special that makes blacks position hard.Those game results are based on less than 80 games.To get a true idea would be to pit identical engines to play many games from that position and report back with results.Do CCRL test these FRC positions with identical software and hardware for both sides, otherwise the results would give false impressions.

[Edmund]

Just looking at that position RKNBNQBR, I can't see anything special that makes blacks position hard.Those game results are based on less than 80 games.To get a true idea would be to pit identical engines to play many games from that position and report back with results.Do CCRL test these FRC positions with identical software and hardware for both sides, otherwise the results would give false impressions.

Indeed 64 games is low, but still a score of 68% suggests a superiority of white with a very high significance (los: 99.94%, elo:130.7)

CCRL tests with both opponents on the same hardware and on top of that each opening position is played twice per pair with mirrored colors the second time. So this equals out the fact that not identical engines are used.