A few days ago it was clear for one night, so I spent a couple of hours outside. At -22°C, a couple of hours was enough.
After a rough polar alignment, I did a TPoint automated sampling run, gathering about 25 calibration points around the sky. This allowed TPoint to calculate polar alignment error and provide me with correction recommendations. Then it clouded over and I was near-frozen.
So, I’ve applied the recommended polar alignment corrections, but have not yet had another clear night to do another pass of TPoint calibration. Because I had to do quite a large correction the first time, I imagine another round of corrections – probably down to a few arc-minutes – will be necessary before the alignment is sufficiently dialled in.