Projections of running Heat Pump on Cosy Octopus tariff [UK]

I’ve been looking into which electricity tariff to switch to after my fixed one runs out next month, and so I’ve been doing detailed simulations using historical data from emoncms in various scenarios. I thought it might be useful to share my findings (with caveats).

Property: a 1950s bungalow with mediocre insulation, 11 kW heat pump, 5 kWp solar panels on east facing roof and 8.65 kWh of (usable) battery storage. No electric car. (I’ve not shown the effect of PV generation here)

Methodology: request hourly consumption for the past year, and simulate grid import and battery usage over time.

Tariff: Cosy Octopus provides two cheap periods between 4am - 7am, and 2pm - 5pm, and an expensive period 5pm to 8pm. Current rates for Yorkshire are 17.56p / 29.26p / 46.82p for cheap / main / peak periods. (Correct for August 2023, likely to be 2p lower in October)

I have plotted the total cost of 1 year’s worth of electricity, excluding standing charge and export:

Flexible Tariff: This is the cost to run my house, with heat pump, for one year, on the standard rate which is mostly dictated by Ofgem’s price cap - this is the benchmark by which all other scenarios are compared to. Ideally, I’d be looking for something that would be cheaper than this.

No Battery: Cost to run the house on Cosy Octopus, with no PV and no Battery, with consumption charged as whatever the rate is for hours of the day. This works out to be more expensive that the fixed rate would have been.

Demand Shift: This attempts to simulate shifting the consumption out of the peak period to when it’s cheaper. So, moving the heating periods to be earlier in the day, and heating the water in the cheap period. Simulation of demand shift is pretty crude, and I’m not entirely convinced it’s correct, however the impact seems to be small.

With Battery: This is where Cosy really starts to be economical. By filling the battery twice a day, we can avoid buying any electricity during the peak period. In fact, more than 75% of consumed electricity is charged at the cheap rate. Shifting heat pump into the cheaper period makes very little difference, as it mostly runs of the battery anyway.

Now, it could be that my simulations are way off, or my usage is atypical, but it’s hard to see how one could benefit from this tariff without the aid of a battery. Perhaps a newer or smaller property with better insulation would work out better, or where the heating is mostly off during the daytime. I’d be interested to hear from anyone else who’s on this tariff or is considering it.

Other observations:

  • adding in the generated solar plus export brings this down massively, unsurprisingly
  • having a bigger battery provides very little benefit, and isn’t an economical upgrade (for me)
  • I ran the same simulation for Octopus Flux, and found it to be a poor choice for my setup

Obligatory referral link with £50 bonus just in case anyone’s looking to join Octopus.


Did you look at Agile? For me (and probably typical non-HP setup) Average over the last 30 days is just below 20p per unit (from the emoncms Agile App) even with some stupidly high peaks this week.

I have been wondering if a Battery only setup would be beneficial. Difference between normal and peak is as a minimum ~25p it seems. For me, it is all about ROI.

1 Like

And your point is?

Hi Tim,

Great post, I’ve been mulling over the different tarrifs. I’m currently on agile (May 2023). Same heat pump as you, 1910 property with mediocre insulation, solar panels (2.4kw system) and 5.8kw battery.

Since May 2023 I’ve averaged a unit rate of 14.19p - see report below. I’ve been charging battery twice a day (picking cheapest 2hr rates both night and day) and then accepting I’ll export on some days on the fixed export rate of 15p.

What’s difficult to predict is the combination of winter rates, heat pump use and unit rates - last year’s look frightening at times given long runs of heat pump. I’m going to stick with agile ATM and try and tweak a bit more with timed hot water runs and see if that helps even further.

Not sure if that helps - I have considered additional battery storage but not sure on cost vs benefit ATM when heat pump could be using 1.2kw/hr without even taking into account background house load.

1 Like

I had not, but your reply prompted me to look. My gut feel was the the price is cheapest when I don’t need it (i.e. solar PV), and expensive when I need a lot (due to heat pump), so I’d assumed it wouldn’t be worthwhile. This is perhaps more true of the daily Tracker tariff, which doesn’t allow demand shifting and cheap charging.

Considering the last 12 months of my own consumption vs. historical Agile rates with a battery (charging up in cheapest 3 hours of 12 hour window), I found that average unit rate worked out at about 26p, which is somewhat better than the ~31p from government’s Energy Price Guarantee.

  • Whole year on Agile without battery = £2468 (30p / kWh)
  • Whole year on Agile with battery = £2138 (26p / kWh)

Now, I’m mindful that the prices last winter were ridiculously high, and Agile has been much more reasonable throughout 2023. Considering only usage and prices since then (240 days), the unit price came out about 20p. However, there’s no telling what will happen to Agile prices this winter…

The battery results in an annual saving around £300 in the analyses above, with a ROI over 15 years. Could maybe improve this with more intelligent charge management, like Kevin’s been doing. Compared to my analysis with Cosy, which yields a saving of £700 and ROI of 7 years.
Caveat: numbers likely to be very different if without a heat pump, so apply salt liberally.

That’s very useful, thanks. I think charging twice a day is the magic bit that makes the battery really worthwhile, making it effectively double the size for the same installation cost.

1 Like

It takes a few seconds with the Octopus Watch (SmartHound) app to work out for both day and night. I just manually set charge times via my ESS app and base it on 3kw/hr charge rate.

Obviously depends on what prices do this winter. As @Zarch reminded me, you can always switch tariffs if it’s too volatile. Although I don’t think you can switch back for another 6 months.

You can switch between their ‘smart’ tariffs every 30 days.
It’s only the tracker tariff that has some 6 month limitation onit.

1 Like

I have long wanted more battery storage, as it often runs out before sunrise*, but whenever I simulate that extra capacity I find the savings gained are way smaller than I’d hoped, making me think it’s just not worthwhile (for my numbers at least).

(*I’m aware that with Cosy or Agile or Go, I’d only need the battery to last until the cheap period).

So, I instead consider the Grid to be a battery with at least 50% efficiency: export at 15p, buy it back at 30p (or less). This “battery” costs nothing to install, has infinite capacity, and no self-consumption that a BCM would have. At the end of June I had 75 kWh “stored” in the grid, which lasted though a very wet July.


That is interesting.

With a battery only system you could charge twice a day. Cosy would make the returns easy to calculate as well.

1 Like

This might be stating the obvious so I say with caution - but in addition to charging the battery you would also be heating etc and may be able to modify household use to benefit from the dips. Although the 3-7am one is more tricky (aside from heat pump use). One thing I’m not sure on is whether if the battery is set to charge from grid, and then drops to say 50% of charge during the charging window, it keeps attempting to charge to 100% so could get additional marginal benefits there?

I guess cosy you have to make the decision based on a more predictable rate albeit affected by overall wholesale price.

Cosy seems to be working pretty well for me so far. I’ve set my thermostat to be warmer during the two cheap periods, so the house only needs a few short cycles in between.

I’m also charging the battery up during those times, which almost lasts until the next morning:

93% of the electricity I buy is at the cheap rate, and almost none at the peak rate. Average rate is 16.7p (inc VAT).



You know, I’m seriously considering Cosy for the winter.

I’m gonna have a ramble here… so I can get it down on paper/print.

I’ve been on Go for about 2 years now. Firstly Go Faster 5 hours (01:30 to 06:30) and now Standard Go (00:30 to 04:30).

Now that I’ve got 16kWh of batteries, I can’t fill them all in the 4 hours.
4hrs x 3.6kW charge rate per hour = 14.4kWh

As detailed in my ‘first winter with the heat pump’ blog I averaged around 14p per unit by charging the batteries, EV, dishwasher, ASHP hot water runs in the cheaper Go period.

Then trying to coast through the day on a combination of solar and batteries.

Note: had only 12.8 kWh last winter

Here’s November to April 6 month totals from last winter.

  • Total Import (5660 kWh)
  • EV (1000 kWh) - 17%
  • ASHP (2481 kWh) - 43%
  • Rest of house (2179 kWh) - 40%

Over those 6 months, 91% of my usage came from the cheaper Go period.

Now I know Cosy only has 3 hour windows, so I could only lob 10.8 kWh in them during the dip, but there two windows of opportunity.

04:00 to 07:00
13:00 to 16:00

This would also suit my two hot water runs a day regime.
One overnight read for the morning and one in the afternoon ready for the evening.

The Cosy dips are currently 15.9p in Yorkshire.
Go is 9p / 29.6p
Intelligent Go is 7.5p / 29.6p

We don’t do a great amount of miles with the EV. So perhaps I won’t miss 9p versus 16p?

In the depths of winter the maximum the heat pump used almost 40kWh in 24 hours and around 6 kWh of that was for hot water.

If you put 6 hours of heating into into the Cosy dips along with the 6kWh from water, that’s around 28kWh remaining. And I can put 21.6kWh into the batteries.

It’s only when it’s sub zero does ASHP usage go beyond 30kWh. It’s less than 20kWh for the whole day through normal winter including hot water.

So i’m thinking I could get almost everything from either Solar or the batteries (filled in the dips/from solar)?

It would also mean I perhaps wouldn’t have to run my complex overnight self written battery management charging code. :rofl:

Or perhaps I could just adapt it to work with Cosy. ie, no point charging the batteries fully if it turns out to be a good winter solar day (and warm).

There you go, thoughts on the table.

Try and beat 14p from last winter on Go or move to Cosy and try to get everything at 16p? :man_shrugging:

Other option now that Zappi is getting Intelligent… move to that, 5 hours at 7.5p?


Ah, but you can’t do 6 hours of heating and reheating the cylinder, as the heat pump is only doing one or the other. I can get away with just 1 DHW cycle per day (though sometimes I skip it), which takes about an hour. That means just 5 hours of heating. If you’re doing two cycles, then that’s another hour of heating you can’t do during the cheap period.

I have some code that figures out the optimal charging for each of the 6 hours, taking into account current levels and predicted demand + solar. Not too complicated. Happy to share that.

The double-dip charging works well for me because I can fill the battery twice, effectively doubling the capacity. Your 16 kWh of storage can only grow to 21.6 kWh per day using this technique, just 35% more. Of course, that’s assuming you manage to use up the morning charge by lunch time. If it’s still half full, then you can only put in another 8 kWh, not 10.

Another consideration: how much are you getting paid for export? 8p? With Cosy you’d be able to switch to Outgoing Octopus, paying you 15p per kWh. This may or may not make it a better deal, depending how much you export. It certainly makes it more complicated to work out :rofl:

1 Like

I’m still old skool 50% deemed exports… :+1:


Given the amount of data generated by both installed kit (some of which is not great - looking at you here Daikin) & 3rd party clamps/monitors/etc you’d think it shouldn’t be beyond the wit of a reasonably talented coder to crunch the numbers to see what tariff works the best for a given set of hardware inputs.

Mine is a hotch potch of standard install plus trial V2H charger with no ability to add solar anywhere because my house is tall & no one wants to quote for anything because “scaffold”. But I’m ~ COP 4.0 overall & ~ 12p per kWh

Imagine a time where smart meters actually deliver on the promise of maximising electricity usage at the right times.


Switch eco mode off to do it faster?

I think Cosy would work for me (I’ve edited the rates table for the cost comparison app to check: shall I post them here?) but the signup is erroring for a week now and no help from support yet.

We were on Agile. I had a lot of code running. The car and heat pump were under computer control. To be honest I think it was stretching people’s patience trying to follow the computer guidance about the dishwasher / washing machine.

The prices I was getting on Agile weren’t very compelling.

I switched to Intelligent. It’s been much easier on the occupants to work out when to turn things on. The price is good enough - we’re averaging 17p a unit.

We do about 250 miles a week. I suspect we’re doing better than other because we only have a granny charger so the charging windows leak outside the 6 hours when we do longer journeys which gives us a bit more cheap power.

About this time of year I expect the heat pump draw to start to be interesting enough to maybe tip the balance. The car isn’t here during the second cosy window.

It’ll certainly all go wrong over Christmas when the car stops moving and the heating draw increases.

I’m too scared to move off Intelligent in case I can’t get the onboarding to work again.


Thanks for this, an interesting exercise. I had a bash at my own calcs this evening, and I think Cosy ought to work for us, with somewhere between 5-20% savings compared to our current bog standard Flexible rates, depending on how well we manage to shift our usage.

I reckon we can shift practically all of our DHW and a reasonable chunk of our space heating demand into the low rate periods (of course offset somewhat by inevitable cookery in the peak period).

And once our battery gets installed, it’ll most definitely work for us.

Just switched from Flux to Cosy at the weekend as we are no longer exporting enough to offset our standing charge + usage.

With the Powerwall and the double dip most usage is being mitigated to the 16.5p rate and there is still some solar top up in between. SE England times are 04:00 - 07:00 & 13:00 - 16:00 and 16.5p/kWh

Due to complaints from “management” the heat pump is only allowed to start heating at 06:00 but as we have a setback of 19 and a daytime running of 21 that seems to be enough to get a good chunk of usage out of the way and start the reheat. With the 5kW inverter once the Daikin HP is running and has done its “start up and use a chunk of power” thing it runs happily on the Powerwall leaving enough capacity for daily usage and appliances.

The battery is able to fully recharge in the window meaning we have ~ 25kw (after some losses) to get us through the day (both WFH so easy to do all the housework during the second dip / on battery) and then through the evening and night. It seems to be working but we currently have a gas oven and hob so evening use is limited.

Once we have a little more sun again (March / April and we are exporting around 50% of our usage, provided prices stay similar we will probably switch back to Flux as we seemed to build up a nice bit of credit for the winter months.

Unfortunately with no EV we cannot use Go or intelligent as with the 13kW battery that would have been ideal. I was looking at a cheap leaf to both reduce CO2 further and allow us to take advantage of the cheaper rates but the MOT reminded me I do circa 2,500 miles a year and most of that is visiting family using longer motorway runs (not something a cheaper leaf would be good at).