Why MCS and CIBSE calculate ground-floor heat loss differently
Solid floors, suspended floors, U-values, and the differences that matter when you design a heating system.
Published November 2026 · 10 min read
Two heat-loss calculations for the same house can give different answers without either being wrong. Ground floors are the clearest example. Their heat loss depends not only on the floor's area and insulation but on its shape, its construction, and the temperature taken on the other side, and the two methods a UK heating engineer meets, the CIBSE Domestic Heating Design Guide 2026 and MIS 3005-D for an MCS heat pump design, answer some of those questions differently.
We investigated those differences while building Heatworx, comparing our results against the published methods and against exports from the official MCS Heat Load Calculator. This is what we found.
Why a ground floor is different
For an external wall or a window, the heat loss starts from its area, its U-value and the temperature difference across it, and the two methods treat those elements in broadly the same way. A ground floor needs two more things settled first.
The temperature on the other side. A wall faces the winter design temperature, a few degrees below freezing in most of the UK. The ground under a slab is taken at the annual mean air temperature for the location, a calculation reference rather than a thermometer reading, which is roughly 8 to 11 °C across the UK. Taking a solid floor against the winter air instead would make its loss two to three times larger.
A U-value that depends on the room's shape. Heat leaves a ground floor mostly near its edges. A long, thin room with a lot of outside wall loses more per square metre than a square room in the middle of the house. Where the U-value is worked out from the construction, both methods take the room's area and the length of its edge that meets an outside wall into account, so two rooms with the same floor can have different U-values. A U-value you type yourself is used as typed, with no shape in it.
Solid and suspended floors
A solid floor sits on the ground. A suspended timber floor sits over a void that is ventilated to the outside, so cold air moves beneath it and, in some methods, through it. The two methods agree on what to do with a solid floor and differ on what to do with a suspended one.
How CIBSE calculates it
Under the Domestic Heating Design Guide 2026 (version 2026-02, incorporating the July 2026 corrigendum), every ground floor, solid or suspended, is taken against the ground reference temperature for the location, which is its annual mean. Section 2.5.2.3 is the clause.
The U-value comes from the guide's own ground-floor tables, indexed by the floor's construction, the soil beneath it, the insulation thickness, and the room's exposed perimeter divided by its area. The tables are built on a standard 300 mm external wall and carry no wall-thickness dimension of their own, so a measured wall thickness cannot enter a CIBSE lookup. For a suspended floor the tables also take the building's exposure to wind.
One further CIBSE rule affects ventilation rather than fabric. A suspended ground floor counts as part of the air-leakage envelope, because air moves through it; a solid floor does not. That is section 2.5.4.2.
How MCS calculates it
An MCS heat pump design follows MIS 3005-D Issue 3.0 (December 2025), which hands the calculation to BS EN 12831-1 and, for ground floors, says two things in clause 5.5.1. A solid floor in contact with the ground is taken against the local annual average external air temperature, much as CIBSE does. A suspended floor is taken at the full winter design temperature difference, like a wall.
For the U-value of a floor described by its construction, the official MCS Heat Load Calculator (version 3.0.1 when we checked) does not use the CIBSE tables. Its documentation says it uses RdSAP 10 section 5.12, the ground-floor calculation from the energy assessment procedure, together with RdSAP's Table 3 wall thickness for the age band. RdSAP itself specifies that calculation per building part, the main dwelling or an extension. In our comparisons with the calculator's own exports, the same formula was applied per room, using each room's area and exposed perimeter, which gave different U-values for rooms sharing one floor construction. Until a floor is placed in a room, the calculator's list shows an approximate figure worked on a standard 50 m² over 20 m stand-in shape; that is not what the rooms are calculated at.
The RdSAP formula fixes the soil at the equivalent of clay, 1.5 W/mK. The CIBSE tables carry a soil type of their own. That is a small difference in itself, but it is a good illustration of the principle later in this article: a method's built-in assumptions belong to that method.
One more MCS behaviour is worth knowing. If you type a U-value for a solid floor yourself, the calculator asks whether it already includes the ground-contact factor. If you leave that unticked, it reduces your value by the equivalence in Annex E of BS EN 12831-1. On our reference house the reduction came out at about 0.8 times the entered value, which is one observation, not a rule.
Worked example: one floor, two rooms
Our reference townhouse has a kitchen and a hall on the same solid ground floor, described the RdSAP way: solid, 25 mm of insulation, a measured 200 mm wall. The RdSAP formula, applied per room as the MCS calculator applies it, gives:
| Kitchen | Hall | |
|---|---|---|
| Floor area | 18 m² | 12 m² |
| Exposed edge | 12 m | 10 m |
| Exposed edge ÷ area | 0.67 | 0.83 |
| U-value, RdSAP 5.12 | 0.53 W/m²K | 0.58 W/m²K |
| Loss against the annual mean (18 °C room, 11.3 °C ground) | 64 W | 47 W |
Same construction, different U-values, because the hall has more edge per square metre. The MCS calculator's export printed the same two figures, and so does the Heatworx audit report. With no insulation and the 300 mm wall RdSAP's table gives for a filled cavity, the same two rooms come to 0.83 and 0.95; those are the figures the calculator printed for that variant, and the formula reproduces them.
Now make the floor suspended timber, everything else unchanged. Under MCS the kitchen floor is taken at the full winter difference of about 21 K, and its loss goes from 64 W to around 200 W. Under CIBSE it stays against the annual mean and its loss barely moves, but the floor joins the air-leakage envelope and the room's infiltration loss rises. Two methods, two different reasons for the number to change.
Where the methods disagree
For a solid floor described by its construction, the two methods agree more closely than their different lineages suggest: both rest on the ground-floor physics of BS EN ISO 13370, both take the floor against the annual mean, and on the reference house the CIBSE table value and the RdSAP value for the same room differed by a few hundredths. The differences that matter are four.
- A suspended floor is taken against the ground under CIBSE and against the winter air under MCS, which can make that floor's loss two to three times larger between the methods.
- A suspended floor joins the air-leakage envelope under CIBSE. Under MCS every ground floor is in the envelope whatever its form, so the switch moves nothing.
- A measured wall thickness enters the RdSAP formula under MCS; the CIBSE tables assume 300 mm.
- A typed U-value on a solid floor is reduced by the Annex E equivalence in the MCS calculator, depending on how the box is answered; under CIBSE it is used as typed.
How Heatworx handles both
The principle we work to is that evidence crosses methods and default tables never do. What you confirmed on site, the floor is solid, the house is from 2000, is evidence, and both methods are entitled to it. The U-value a method's own table or formula turns that evidence into belongs to that method alone.
Two honest qualifications. First, an RdSAP description often carries an assumption with it: if you did not confirm the insulation, the 25 mm in the example is RdSAP's default for the age band, and it travels into the CIBSE lookup as the best available fact. The audit marks such a floor's inputs as not yet confirmed. Second, some evidence cannot cross: the measured 200 mm wall is a fact, but the CIBSE tables have no wall dimension to receive it.
So a floor is described once, in one of three ways, and each method uses it its own way:
| How the floor is described | Under CIBSE | Under MCS |
|---|---|---|
| Picked from the Heatworx catalogue (form, soil, insulation) | The guide's table, per room by exposed edge over area | RdSAP 5.12 from the same form and insulation, per room, at the catalogue's 300 mm wall |
| Described the RdSAP way (form, insulation, age band, wall) | The guide's table from the same facts, per room; the wall thickness does not enter | RdSAP 5.12 from the description, per room, with the wall as entered |
| A U-value you entered | Used as entered | Used as entered |
The last row is the one known gap. Heatworx currently applies a typed solid-floor U-value directly under MCS. The official calculator can apply the Annex E adjustment, depending on whether the entered value already includes ground-contact effects, so for such a floor the two can differ. The Heatworx audit report flags the floor on its room page and in a method note, and until the adjustment is implemented those designs should be checked against the calculator before an MCS submission.
Both methods also need to know the floor's form, because each has a rule for suspended floors. Heatworx takes the form from the catalogue type, from an RdSAP description, or from the construction you set on a floor whose U-value you typed. A floor whose construction says nothing is taken at the full design temperature difference under MCS, which is right for a suspended floor and an overstatement for a solid one, and the audit says so and tells you how to fix it.
What the audit shows you
Every time the calculation uses a figure for a floor other than the one shown on its construction, the Heatworx audit report says so in three places: a sentence under the room's fabric list naming the method, every input, the room's shape, the value used and the value not used; a "per room · derived" line in the Construction sections appendix listing each room's figure; and a numbered method note at the back stating the rule and what the reference calculator does. If the floor's inputs are still assumptions, the appendix says that too.
What to check
Describe the floor you saw. Solid or suspended, the insulation if you know it, the age of the house: those facts go to both methods, and both will work the U-value out per room. Enter a U-value yourself only when you have one from a measurement or a product sheet, and know whether it already allows for ground contact. If a job needs both an MCS submission and a CIBSE calculation, run the survey under each and read the floor rows, but design the system to one set of room loads, not a mixture of the two.
Sources
- MIS 3005-D, The Heat Pump Design Standard, Issue 3.0
- MCS Heat Load Calculator documentation, U-value lookup
- RdSAP 10 specification (June 2025)
- CIBSE Domestic Heating Design Guide 2026