Heat loss is the sum of everything leaking out through the fabric — ΣU·A·ΔT, each surface's area times how readily it conducts, times how much colder it is outside — plus the heat carried away by air changing over, 0.33·n·V·ΔT. Add them and you have the kW the house needs on the coldest design day. That number sizes the plant, whatever the plant is.
A radiator's rated output assumes water about 50 K hotter than the room. Run it cooler and output falls with roughly the 1.3 power of that difference — not linearly. A boiler at 70 °C gives a radiator about 85% of its rating; a heat pump at 45 °C gives it about 33%.
So the same room needs roughly two and a half times the radiator. Heat pumps in badly performing installations are usually not undersized pumps in a cold house — they are ordinary pumps pushed to high flow temperatures because nobody changed the emitters, which wrecks the efficiency that was the whole point.
The design heat loss, rounded up to the nearest available unit — not a rule of thumb per square metre. Oversizing costs efficiency because the unit cycles instead of modulating, and it costs money up front. A well-insulated 110 m² house is often 4–6 kW; a solid-walled Victorian one of the same size can be double that.
Rarely all of them. Radiators are frequently already oversized for their rooms, and the requirement is per room, not per house. Work out each room's loss and compare it with what the existing emitter gives at your intended flow temperature — this tool gives you that multiplier.
Twice, in opposite directions. It sets how much output your emitters give, and it sets the heat pump's efficiency — roughly 2.5% less COP per extra °C. Running hot to compensate for small radiators is exactly the trade that makes a heat pump expensive to run.
No. A real MCS survey measures every room, every surface and the actual construction. This is for understanding the shape of the problem before anyone quotes — and for sanity-checking a quote that looks wrong.