I've been thinking a lot about AW and it's a great software that requires a lot of attention, but it has a huge weak point.
They are the Benchmark that are made by protocol. I can see that I have a different HASH in two coins of the same protocol. It is very noticeable, I mino much neoscrypt, the same rig for a coin can give 5mhs, but for another currency neoscrypt can give a rate of 7 mhs.
Then thinking one realizes that it does not make sense that the switch is done only by the hash of the protocol, that's fine for the pools type Zpool and auto change.
But if I have 80 coins added, I need to know the hash for each of them, because although many are from the same protocol, they get different hashes. This difference can be by the ping, it can be by the stratum, by configuration of the Vardiff etc ...
Consequently, it would not be unreasonable PAtrike that you are considering in your programming that we can also save the hash for currency, and that this is the one chosen to make the Auto Switch.
I think the problem is perfectly understood, and is the culprit that some currency changes are worse instead of better, something that should not happen, and that only happens because of what I just commented.
A program of this price and this level, with so many added functions (and many other bugs), should be closer to the reality of each currency, being able to obtain that hash per currency / pool, instead of only Protocol HASH, which in the end it is not precise at all.
Hi, it actually makes sense to have benchmarks per coin instead of per algo. It may be a bit of hard work to tweak this in the programming, but in the long run us (users) will have a more down to earth expectation of what we would be earning and the profit switching would be more accurate.
Hopefully we could have this in a future update.