@jamespa This does make me wonder though, how much more energy does the typical household use to heat a house constantly at 21C (heat pump) versus the conventional boiler approach of turn it on for an hour in the morning before you get up and go to work, and again for 3h in the evening when you get home. Does this make heat pumps more (or only) suitable for those who are home all day and would otherwise have had the heating on anyway.

Remember that the house continues to lose energy when the heating is off so the answer to the question depends how much it cools when its turned off.  If it doesn't cool too much (which many dont) then the reduction in energy lost to the outside world is small, so the reduction in energy required to replace the energy lost (which is what you have to do) is also small.  You can easily do the very rough maths.  So for example if a house is heated half time to 20, and at OAT=0 drops to a min of 14 when its not heated, the average temperature of the house is roughly 18.5 (using linear ramp up and down, an approximation of course).  So the saving in this example is 18.5/20 ie 7.5%.

If we are struggling to get heat pump installs to compare favourably on a price per kWh basis (given the 'spark gap'), if we are then using twice the amount of energy in providing low and slow heating to keep the house at a constant 21C (when it otherwise might have spent half it's time at 15C),

The typical saving is nothing like that much as the calculation above shows.  Even if, when its 0 outside, it magically goes from 20 to 15 and back to 20 instantaneously (which happens only if it has the thermal characteristics of a tent) the average temp is still 17.5 in this case for a saving of 12.5%.  Obviously if it cools this much when its warmer outside then the savings are more, but most houses wont do this because of the high heat capacity (often called 'thermal mass').  For example the heat capacity of my house is very roughly 14KWh/ degree C and the loss at -2 is 7kWh.  This means that it takes 2 hours to cool by 1 degree C when its -2 outside.

When I had my gas boiler I changed from on/off heating in the 'normal' pattern to heating 24x7 and at the same time reduced the flow temperature, resulting in a reduction in the amount of gas I used annually.  This was, I am pretty certain, the slight extra energy required being offset by the increased efficiency because the boiler was actually condensing, and also not cycling as much because I was heating at lower temperatures

Noted.  In fairness some people only have scrappy records of gas consumption, its geeks like us that now keep everything!

Clearly if you dont have records and dont have much idea how the house was heated its going to be difficult.  You still have an enormous increase in consumption however, you must have been living in something close to a freezer!

@jamespa This does make me wonder though, how much more energy does the typical household use to heat a house constantly at 21C (heat pump) versus the conventional boiler approach of turn it on for an hour in the morning before you get up and go to work, and again for 3h in the evening when you get home. Does this make heat pumps more (or only) suitable for those who are home all day and would otherwise have had the heating on anyway.

If we are struggling to get heat pump installs to compare favourably on a price per kWh basis (given the 'spark gap'), if we are then using twice the amount of energy in providing low and slow heating to keep the house at a constant 21C (when it otherwise might have spent half it's time at 15C), no wonder people notice their bills have doubled, and then try to turn down the heat pump to save electricity and then complain their house is permanently cold. I love my heat pump, but I'm not sure I'd love it so much if we were out at work all day.

you must have been living in something close to a freezer!

The years between oil use and the heat pump (roughly 2019 to 2021) when I used standard mains electricity to heat only the room(s) I was in were like living in a freezer in the colder months. The kitchen could get down to 7°C overnight. Maybe that put a rosy glow on the oil years, when in fact the oil years were also rather chilly at times. Plus as I said, we were hardier when younger!   

Noted.  In fairness some people only have scrappy records of gas consumption, its geeks like us that now keep everything!

Clearly if you dont have records and dont have much idea how the house was heated its going to be difficult.  You still have an enormous increase in consumption however, you must have been living in something close to a freezer!

We did discuss the oil to heat pump jump in apparent heat delivery some time ago in this thread. We didn't really reach a firm conclusion (i was still very new to heat pumps back then), more something along the lines that a lot of small things can add up, plus, crucially, the different use pattern, timed vs continuous running.

What I am reasonably sure about is that the overall picture (mostly my financial records are good enough to mean that i probably did include all oil purchases over the years, what is less clear is how and when the oil was used, and of course with the heat pump I have much better records) over several years is that energy delivered to the house did show a huge jump when I had the heat pump installed, and I think most of the explanation lies in the timed vs continuous heating pattern of use. My house probably did cool overnight and during the day when I used oil, I vaguely recollect it was only on for a small number of hours in the morning, and few more in the evening during the week, maybe more at the weekend. We were younger and fitter back then!   

Most of those comments also apply to a gas system, the material differences are the fact that you cant get daily/hourly data from an oil system and (perhaps) a greater uncertainty in efficiency.  The former does prevent you doing a plot vs OAT, but most of the matters you point out will affect this plot so equally affect one based on gas consumption.  In summary I think what you are saying applies almost equally to a gas based system and therefore your argument is tantamount to saying that sanity checks based on measured consumption are not practical.  In fairness some argue that this is the case, but you aren't one of those!

Obviously any sanity check depends on certain assumptions, mostly that you actually heat the house to a reasonably comparable level and that it stays reasonably warm throughout the day, whether or not it is heated.  Some you can estimate a correction for, others you cant.  Perhaps these assumptions weren't the case with yours.  Im really struggling to understand how you got away with 1000l of oil (10.3MWh/year) but now need 23MWh/year; that's a factor of 2 even ignoring efficiency, suggesting that the average difference between the house temperature (which is the main variable) and outside temperature was half what it is now.  Perhaps that is in fact the case.  Something quite dramatic must have changed to explain this, its very unlikely to be due solely to uncertainties in the calculation.  

I wouldn't say its a non starter.  If you can get annual consumption somehow it is still a sense check that will point out any survey results that are way out (particularly those that are grossly overestimated).  Very far from ideal, but perhaps still better than no measurements at all.

I beg to differ. I really could not make any useful predictions from my past heating oil use. Common sense based checks will kick in before oil use based checks kick in. Consider (as @old_scientist suggests):

1. you only have annual consumption, if that. How do you account for that late Feb top up, and the tanks still had oil in it (how much?) when you turned the heating off? How did you measure the volume in the tank? 

2. How efficient is my 10 year old boiler? 20 year old boiler?

And then:

3. when did you turn the heating on and off, both overall (for the season) and during the day? How and when did any daily timing's change? Did you keep records, or is it just a vague recollection? Did usage patterns change, and you have simply forgotten what the changes were?

I think it is worth looking in more detail at what happens if you try to do an oil use based assessment. Let's say it is likely I used around 1000L per annum. That was almost all heating, as I used an immersion heater for DHW in the summer and even sometimes in the winter. Given 10.35 kWh per litre, that is 10,350 kWh per annum. Boiler efficiency? Haven't got a clue, really. Lets say 75%, 10,350 becomes 7762 kWh. Winter DHW use? Who knows, lets say 5%, heating energy delivered is now 7374 kWH. Was the system on a timer, meaning an unstable IAT? Almost certainly. How do I account for this? Mean daily IAT? Do I know what that was? No. And so on. All I can say is I might have delivered around 7400 kWh of heat to the house over a year, but what that actually achieved beyond 'warming the house and some hot water' I have no way of knowing. There is no sensible way I can turn that into a kW loss at my design temperature, which is what I need to do a heat loss sense check on an installer recommendation.

Bear in mind we are focusing in energy delivered, not used. What was the figure on my quotation (the one used for the installation)? 26,389 kWh per year. By any sane sanity check, 7,374 vs 26,389 fails. Where do I go from here? That is why I suggest oil based sanity checks are a non-starter. If you do start doing them, you will very likely end up more confused that had you not started doing them.

What is the actual heat energy delivered to the house, now that the heat pump is installed? 23,079 kWh in 2025, 19,821 kWh in 2024. This is a LOT more than the oil use predicted. Those with long memories may recall we discussed this some time ago. The most likely explanation we came up with is that the increase came from changing patterns of use, from timed use, perhaps 12 out of 24 hours per day, maybe even less (2 hours in the morning, 6 in the evening ie 8 hours per 24 hours?) to 24/24 hour running. With the heating on for three times as long, one might expect it to deliver rather more energy. Or maybe not (oil boilers and heat pumps are very different beasts). 

If the oil heating was only on for 8 hours a day, perhaps we should multiply it by 3 to get a 24 hour equivalent. If we do that 7374 becomes 22,122 kWh per annum, much close tor actual heat pump usage. But that's with hindsight. At the quotation stage, I had no way of knowing what that factor might be, unless I had a meticulous record of how I had run my oil based system in the past, which I didn't have. I could only guess.

The empirical methods of determining heat loss either need careful measurements of heat delivered to the house at design OAT and a stable IAT (rarely done) or a series of measurements over time at various OATs and a stable IAT, which are then plotted to get the design OAT (by far the most common method). The '1000L of heating oil used over a season' just doesn't have enough other data attached to it to be of any practical use.      

I didn’t have the OAT for the full range of data so that’s probably why.

@travellingwave 

Ah yes, that would explain it. 

I didn’t have the OAT for the full range of data so that’s probably why.

however it is not so easy when replacing an oil boiler. When replacing my oil boiler I did not think to time on/off usage over a fixed period as suggested earlier in this thread, and estimating annual usage is not trivial when the oil tank is filled on different dates and the amount of oil in the tank is never accurately known. Then combine this with guesstimates of boiler efficiency anywhere between 60-85% and you end up with a figure with large margins of error.

Again, agreed. There is also the likely change from timed to continuous heating, which further complicates things. For all practical purposes, using past oil consumption to determine heat loss is in effect a non-starter. But with a gas boiler, which is what most people have, and either a smart meter or another way of accurately measuring energy use over time, it can be done. 

I wouldn't say its a non starter.  If you can get annual consumption somehow it is still a sense check that will point out any survey results that are way out (particularly those that are grossly overestimated).  Very far from ideal, but perhaps still better than no measurements at all.