Because theoretically there is no heat loss if you have another flat on top.

I do say theoretically....

er -- no !! they are usually full of outside air so not even as well insulated as an outside wall! No idea how building regs allow them either.

cheers, reading the SAP documentation this is what it says about party walls...

HEAT LOSSES
The areas of building elements are based on the internal dimensions of surfaces bounding the dwelling.
Window and door area refers to the total area of the openings, including frames. Wall area is the net area of walls after subtracting the area of windows and doors. Roof area is also net of any rooflights or windows set in the roof. Losses or gains through party walls and floors to spaces in other dwellings or premises that are normally expected to be heated to the same extent and duration as the dwelling concerned are assumed to be zero (and these elements are therefore omitted from the calculation of heat losses).

you can also insulate the party wall, and to a robust detail you would insulate the ends to stop flanking - if you have a partial fill cavity.

Posted By: 10100Losses or gains through party walls and floors to spaces in other dwellings or premises that are normally expected to be heated to the same extent and duration as the dwelling concerned are assumed to be zero (and these elements are therefore omitted from the calculation of heat losses).

Thats the bit that is difficult to assume one way or the other. If is an old house with no cavity then next door is likely hetaed so no losses.
If its a cavity how do you know one way or another? In my experience only the ends and tops of such walls are insulated, and if there are any gaps.......

SAP2010 is likely to include some kind of acknowledgement that uninsulated cavity party walls are potential thermal bypasses and can have significant effective U-values. Quite what form this will take I do not know. At Stamford Brook (cavity wall detail was robust details WM4) the measured party wall effective U-value ranged from 0.4 to 0.7 W/m2K. Even with a cavity sock at the sides and top of the cavity the effective U-value was still of the order 0.2 W/m2K. We have also measured party wall U-values in a timber frame construction of the order 0.4 W/m2K. In this case there was no atticspace and the heat loss was via all the edges of the cavity - including the underfloor void which we found linked with the party wall cavity. Our experince is that relying on the vertical acoustic barriers to seal the edges of the party wall cavity will not work - you will need properly constructed air barriers to do that. Cavity party walls in their various guises in the UK are designed for acoustic performance not thermal performance.

Also bear in mind that cavity party walls are not the only thermal bypasses in typical UK construction - we have identified many others - including some forms of knee wall cavity in room-in-roof construction , dropped soffits below ridge beams and many other instances of construction voids and cavities which break the rules of good thermal insulation. It is essential that the air barrier and insulation layer are continuous and in contact with other around the entirity of the thermal envelope.