My only comment is well done.  You have come at this several different ways and got, perhaps a bit fortuitously, very similar answers.  If only heat pump installers did this as standard.

I can't see how you will do much better without making lengthy measurements specifically designed for the purpose, but neither do you need to.  

I don't know if you are contemplating a heat pump or already have one but if you are contemplating one this should stand you in good stead. 

Many installers will refuse to take any notice, more enlightened ones will note the results and adjust their assumptions accordingly (they have to do a heat loss calculation by mcs rules).  I don't think you need us to tell you who to go with if the first group come up with figures that are significantly different, as they may.

My calculations were different, but essentially my heat pump was sized based on measured consumption under conditions known to me and any installer coming up with a materially different figure was politely advised that their offer would not be considered.

Hats off, can´t fault it. 

In our case, I was most confident with the fact the surveys fell between the gas usage calculated loss. Well, HG´s did after some "tweaking"..

Posted by: @batpred

↑

In our case, I was most confident with the fact the surveys fell between the gas usage calculated loss. Well, HG´s did after some "tweaking"..

Was the tweaking adjustment of assumptions to fit the facts?

Posted by: @jamespa

↑

Posted by: @batpred

↑

In our case, I was most confident with the fact the surveys fell between the gas usage calculated loss. Well, HG´s did after some "tweaking"..

Was the tweaking adjustment of assumptions to fit the facts?

yes and no. apparently HG’s AI was automatically assuming massive ventilation losses (and I noticed other posters mentioning the same in their case. So that needed tweaking. But the second iteration of proposals and tweak seemed to be like an override, to input the lose value we had calculated from gas…

I just followed your suggested method. Ask questions until the right answer is written on the heat loss survey! 

Apart from that, we were very happy with the HG install and commissioning ..  

As others have said, well done, and thank you for posting such a detailed account of your methods and results.

Posted by: @rob-of-york

↑

The first thing that struck me was that the figures from the two completely different heat loss calculations are surprisingly close. As they are worked out completely differently, this gives me some confidence that they are about right.

This is perhaps one of the most significant findings, that two totally independent methods produce, give or take, the same result. As a scientist, you don't need to be told that this is replication of the best kind (because the methods are independent), and it greatly increases confidence in your results.  

Thank you all for the positive comments.

I have made a simple spreadsheet that implements the practical heat loss calculation that I hope will be of use to others - see below.

I agree that the close alignment of the theoretical and practical heat loss calculations that I did for my house may be rather fortuitous. 

However, the biggest area of uncertainty that I had was in the air changes per hour (ACH) value in the theoretical calculation. The house has retrofit double glazed doors and windows installed to a good standard. I have also been diligent in eliminating any drafts that I could find; they were easy to spot with a thermal camera when it was sub-zero outside. I therefore chose ACH = 0.5 as that seemed to be the default for well-sealed modern properties.

Every 0.1 increase or decrease in ACH equates to 268W of heat loss according to the calculations for my house. This is really significant; if a heat loss survey used ACH = 1.0 instead of 0.5, then the overall heat loss would turn out 1340W higher, about 21%.

Digging into this further, it appears that the issue with high ACH values being applied to older properties has been recognised. The Easy MCS website has some information on this stating that: 

New air change defaults

The guide will now align with BS 12831-1:2017, which sets simple defaults:

• 0.5 ACH for habitable rooms
• 0.0 ACH for non-habitable rooms

These apply to all homes, regardless of age or construction.

See https://www.easy-mcs.com/2025/12/02/what-the-updated-cibse-guide-means-for-your-heat-loss-calculations/

This seems to be a positive step with older homes no longer penalised with unrealistic ACH values that add kWs to the calculated heat loss.

Previously, pre-1996 homes used the following defaults for ACH depending on room type: bedroom 1.0, dining room and living room 1.5, kitchen, hall, and landing 2.0, and bathroom 3.0. I tried these values in my heat loss calculation, and they gave a value of 4415W for ventilation heat loss alone and 9312W in total. This compares to 1340W for ventilation losses and 6420W in total with ACH =0.5. With those previous defaults, I'd need a separate 5kW heat pump just for the ridiculous ventilation losses! I'm glad that issue with the MCS heat loss calculations has/is being fixed.

@rob-of-york — thanks for posting this, all very useful and accessible stuff. I particularly like your demonstration of the effects of different ACH values. 

There is also some published research (SPAB) which suggests that older houses with solid stone walls can be penalised by wall U-value estimates being too high.

Here is a further method of calculating heat loss, with thanks to @jamespa and @cathoderay for a lively discussion on another thread. (I've posted it here, as it fits with the other heat loss estimates that I have looked at).

This approach is again a practical one, and is based on daily averaged measurement data that it is possible to obtain with a smart meter and a home weather station.

The idea is that in an equilibrium state, the energy going into the house balances the heat leaving the house in any given time interval. Since energy divided by time is power, the equation is:

PM + PO = HTC(IT - OT)

where PM is the Power determined from Measured data, PO is the Power from Other sources, HTC is the Heat Transfer Coefficient and IT and OT are the average Indoor Temperature and Outdoor Temperature respectively over the time interval considered.

HTC is the power required to keep the house 1-degree above the outside temperature, and is measured in Watts per degree (W/C).

The equation can be rearranged to read: PM = HTC(IT - OT) - PO.

If we plot data for PM against (IT-OT) on a graph, then the equation is that of a straight line, with a slope of HTC that crosses the y-axis at a negative value equal to PO.

If we can obtain values for HTC and PO, then we can easily calculate the heat loss for a reference temperature difference, e.g. 24-degrees, representing -3C outside and 21C inside.

Given the daily variation in temperatures and the fact that houses typically have a lot of thermal inertia, a reasonable time interval for each data point is a day.

I have collected daily data for a 6 month period from Oct-24 to Mar-25. This includes gas consumption, electricity consumption, and average indoor and outdoor temperatures from a weather station, with a temperature sensor in my back garden.

The boiler is assumed to work at 90% efficiency, as it always runs condensing except when heating hot water, and all of the electrical energy is assumed to turn into heat with 100% efficiency. I totaled up the daily energy input from gas and electricity in kWh and converted that to a power figure by dividing by 24 hours.

The graph below shows a scatter plot of the 182 data points, along with a regression line.

The regression line gives the slope, i.e. HTC, which is about 268W/C and the y-axis intercept, i.e. PO, the power from other sources, which is about 455W.

There is one caveat with this data. I am unable to separate out the gas used for hot water (the heat from which just drains away) from that used for heating and so balances the heat loss from the house. On average, about 150 units of gas are used for hot water in a 31 day month, corresponding to about 180W continuously.

Assuming a 24-degree reference temperature difference, the heat loss works out at (268 x 24) + 455 - 180 = 6260W. This is close to what I obtained via a detailed room-by-room heat loss calculation, described in the first post in this thread.

It took some work to get the smart meter and weather station data in the right format in an Excel spreadsheet, however, there was nothing particularly sophisticated in the information needed or in the processing required.

It would seem that a good estimate of heat loss can be obtained from standard smart meter data, combined with average indoor and outdoor temperatures that simple home weather stations can provide.

[Note, as I have solar PV, I took the electricity consumption from the monitoring app for that, rather than from the smart meter data, since that shows we hardly draw anything from the grid]. 

It is interesting to note that while the above graphical method worked very well with a large number of data points, i.e. 182 data points for 6 months, there was too much variation to get good results when using smaller data sets corresponding to individual months.

The graph below shows exactly the same data points as used in the previous post, but now they are divided into individual months and 6 regression lines drawn, one for each month. It is apparent from the large variation in the slope and y-axis intercepts of these lines that 30 or so data points are simply not enough. Where there is only a small variation in temperature differences over the month, for example Feb-24, the extrapolation becomes particularly inaccurate. This problem is avoided with the full data set as the range of temperature differences is much larger (from 5 to over 20-degrees).

In an ideal world, I would be able to look at the same 6-month period over multiple years; however, unfortunately this is the only year where meaningful data exists. During 2023 and summer 2024 we comprehensively upgraded our house's insulation, and then at the end of summer 2025, we had an AC system (air-to-air heat pump) fitted that was used in winter 2025-25 to supplement the gas central heating. The app for the AC system only provides daily data for the past few weeks and rolls previous data up into monthly totals. Hence comparisons with earlier years are meaningless due to changes in the fabric of the building and a comparison with last winter is not possible due to a lack of data.

@rob-of-york 

Great data and analysis.  Im personally convinced that, intelligently used, this is a sound way to assess heat loss.

FWIW here is the plot I primarily used to size my heat pump.  I didn't have IAT but was pretty confident that it was sufficiently constant.  I didn't have OAT either, so used degree days from a nearby weather station.  Obviously its much better if you have both!  This is two years worth of data and daily resolution derived from half hourly smart meter data.  A very similar picture to yours.  The positive intercept suggests my base temperature for heating (ie the temperature at which the loss is balanced by extraneous sources) is actually 17.5 not 15.5, which is plausible as there are only 2 in the house and so perhaps only 600W gained from electricity baseload plus people.