I guess I am just missing something here. How can you compare expected blocks of different miners when different miners are on this pool. The work sent from kano to the miners is random right? So if a A6 hits a block. It is most likely because s9's did a lot of the work. Or A7's,s5's,s3's,or whatever else miners are out there. The only way you could truly compare the expected blocks is if they were each in their own pool.
We can identify most mining hardware from the mining protocol information so we can associate shares with the mining hardware it came from and derive statistical data from that. That's done in addition to checking the mining data on a per-user and per-worker basis.