well my 24h+ testruns are over and i might say those 25-30% higher profitabilities
reported on zpool are non-existent, i get very similar btc from nicehash and zpool
the same algo (qubit, quark, x11).
so in the end block % is much closer to reality than shares %, not technically but for practical calcs.
unfortunately it seems impossible to construct a formula or even a good estimate
since zpool and NH use so different approaches
Based on a recent short multi-algo test on NH, I'd say zpool far outstrips NH, but maybe that can change with supply and demand of orders.
There are definitely some figures that don't seem to add up on zpool, whereas hashpower didn't seem to have the same problem. The latter pool though looks like it needs putting out of its misery at times.