Yes, this would be an excellent option to support in miners.
There is a slight tradeoff there, since shorter cycles tend to have higher fidelities, so while we want take some shorter lengths into account to reduce variance, we don’t want to include much shorter lengths that would move the expectation too much. I propose to do as in
which obtains a fidelity over the 10 cycle lengths from 24 up to 42.
For clarity let’s call this fidelity_24_42.
If miners could provide this metric that would be very helpful indeed.