Dave,
I asked more or less the same question to the RO for Derbyshire and he went back to the guidelines which just say to extrapolate based on the numbers you counted in the TTVs. I agree that this can give a result which you know is wrong - so you could leave the estimate blank, and I would then add a roving record for your sighting outside the TTV. I guess that there will be loads of instances of this sort of thing and the combination of TTVs and roving records will between them lead to something approaching an accurate figure. I'm sure that's the theory anyway!
Thanks Simon.
In the end I asked Dawn Balmer, and this is (part of) her reply :
"I suggest you put in a Tetrad Population Estimate based on your experience of the numbers present on other visits over the winter."