In reply to Beardyman:
We are trying that experiment now. We've got a 1930's house which is drafty and leaks heat (ongoing issue we are working on). If the house fully cools down then it takes 24-36 hours to get back to temperature, so we thought we would experiment.
Hence this winter we tried keeping the heating on continuously, and last winter we tried the on-off technique.
More accurately, this winter we heated the upstairs continuously (heating never off but controlled by roomstats), and downstairs was continuously heated 6am-11pm (also controlled by roomstats).
TRV's on all radiators.
Anyways, we got the gas bill for the winter a couple of days ago, and the result appears to be that it's about the same to heat the house continuously rather than on-off, however its *certainly* more pleasant to be in a house at a constant (warm) temperature. This rather surprised me as its counter-intuitive that on all the time is the same.
Total energy used for both winters (Nov-Feb) was around 15000 kwh.