Yes Andrew. I realised that after a few attempts at ‘tinkering’ with PHD2!
At first, on guiding on the target star, I noticed a few zig-zags in the PHD trace as it hunted for the star in the slit. I tried to train that out with an increased hysteresis value but that just caused decreased ‘reactiveness’ either side of the slit and led to worse guiding (it would).
So I went back to my original values when I realised that it was, as you say, sampling the whole disc by default and thereby negating (to some extent) the chromatic aberration.
I wouldn’t say ‘bad guiding is good’ but “iffy” guiding certainly has it’s benefits!
I still guide on a nearby star given the opportunity.