Assuming that five seed experiments are run now, is the result reported in the paper the highest of each experiment and then averaged?