This is very neat.
Would it be much extra effort to do the same thing but with a "leading edge" proth test, i.e. something that might have been assigned in the last day or two?
Obviously with tests from multiple k-values being assigned, this wouldn't be an exact science, but something that indicates a sort of mean, or average proth test size.
Anyways, cool stat.
I think someone just took the #4 or #5 spot from GIMPS with a General Fermat prime, didn't they? I thought I heard that somewhere......