Are any albert units usable as benchmarks?

Before the albert units came out, nearly all work units seemed to be sized in a very tight band. That made it easy to compare various computers merely by comparing the times to complete a single unit. If something went wrong with a computer, it could be quickly identified.

I know Albert units can vary widely in the amount of computing power required to complete each one.

My question is whether it is possible to examine the ID numbers of albert units & find collections of units that can be expected to all require nearly the same amount of computing power?

At present I know nothing about the significance of the unit ID codes.

Thanks,

ADDMP.