question about the new training process

[l1t1]

for now, when a new net is promoted, it only plays very little self-play games(1 or 2), then it starts to play test games against some nets trained with prior nets. but thus one net won't play with net trained with its own self-play games, is it a problem?