Sample a ton; perform factor conditioning; then count the observation you’d like.by rejecting those who don’t match