Every saving on this site comes from running Orza’s planner through a home’s electricity, half hour by half hour, knowing only what was known at the time. The same battery then runs on its own schedule, and the saving is the difference between the two bills.
The scale
Real homes, whole years, planned half hour by half hour.
460,000+days planned by Orza across every backtest, each on only the prices and forecast known at the time
950,000+days replayed in all, with the battery left alone, set by hand and the best possible run beside Orza
316whole years behind the savings calculator
25real homes’ electricity use in those years, from 9 to 63 kWh a day, with 5 to 45 kWh of battery
Days counted on 29 October 2026 from every result we kept, re-runs included, and the runs that built and tuned the planner as well as the tests below. Runs whose results were not kept aren’t counted, so these are the least it could be.
What the words mean
Homes, runs and days.
Different pages quote different tests. These are the words they use.
A real homeOne home’s own battery and meter: 27 kWh of battery and 13 kWp of panels. Its figures come from its own half hours of sun and use.
A modelled homeA real household’s smart meter readings (Low Carbon London, UK Power Networks), moved to 2025 to 2026 a whole number of weeks so a Thursday stays a Thursday, with modelled panels and a modelled battery on a south facing roof.
A whole year, or a runOne home, one battery and one tariff, planned through a whole year. The same home on another tariff is another whole year, which is how 25 homes make 316.
A day plannedOne day of one run. The same date comes round again for every home, battery, tariff and planner setting, so 460,000+ counts days of planning, not dates on a calendar.
A day replayedA day planned, or the same day run another way: 463,870 planned by Orza, 355,471 run for comparison (the battery left alone, on a timer, set by hand or with no battery), and 136,682 with the best possible plan in hindsight. 956,023 in all.
The savingWhat the same home, battery and tariff cost on the battery’s own schedule, less what they cost following Orza’s plan, with any charge left at the end counted. Where a guide compares with no battery at all, it says so.
What was run
The tests behind the figures.
Each figure on the site says which of these it comes from.
20 households, 40 modelled homes
The year test
Twenty households, each with a 9.5 kWh battery and 4 kWp of panels and again with 13.5 kWh and 6 kWp, through 1 October 2025 to 30 September 2026 on the Octopus tariffs a battery owner might choose. More than 280,000 days planned. Most guides use it.
Six households using 21 to 63 kWh a day, one of them already in the twenty, at 20 and 30 kWh and at a size matched to each home, 25 to 45 kWh, on all four export tariffs. That makes 25 homes and 316 whole years, the runs behind the calculator.
Before those six homes’ matched runs started, the calculator’s curve predicted what each would save. Then we ran them. The typical miss was 6.1%, and 22 of 24 landed within 15%.
The year test’s homes on Go with Outgoing Octopus with a simple routine: fill in the cheap window, then let the battery run the house. It saved £263 a year on average against the battery left alone, where Orza’s plan saved £404.
A backtest on the real home’s own readings, 6 October 2025 to 1 September 2026, with its 27 kWh battery, on four tariffs: £500 to £970 ahead of the same battery left alone. The home page’s large figure.
One real home, 34 days
A real home’s autumn
Its real half hours of sun and use, 3 September to 7 October 2026, planned on each tariff and export limit. The guides that say “one real home” use it.
£500 to £970Large stack, and the hero’s “up to £970”
48 whole-year runs at 25 to 45 kWh in six high-use homes; shaded, the middle half. Rings: the real home on four tariffs, £500 to £970 between the lines.
£360Typical solar home, about 13.5 kWh
100 whole-year runs at 13.5 kWh, one dot a home on a tariff; shaded, the middle half.
~£240Compact, about 9.5 kWh
80 whole-year runs at 9.5 kWh, one dot a home on a tariff; shaded, the middle half.
The rules
The same rules every time.
Only what was known thenEach half hour is planned from the prices published by then and the sun forecast as it was issued.
The weather that cameWhat the panels make comes from the sun that really shone: ERA5 for the modelled homes, its own panels for the real home.
Published pricesOctopus’s own rates for North West England, half hour by half hour.
The same battery both waysOrza’s plan and the comparison run on the same home, battery and tariff.
What is left outStanding charges are left out, and no limit is put on export unless a page says so.
Their electricity use is: real households’ smart meter readings, half hour by half hour. The panels and batteries are modelled, except for the one real home, whose figures come from its own readings. The prices are the ones Octopus published, and the weather is the weather that came.
Why do pages say 20, 25 or 40 homes?
They are quoting different tests. The year test is 20 households, and on two sizes of battery and panels those make 40 modelled homes. Six bigger households, one of them already in the 20, bring the calculator to 25.
How can 25 homes be 460,000 days?
Each home is run on several battery sizes and tariffs, so the same date is planned many times. The count also includes every other backtest we have kept, re-runs and the planner’s development runs among them. It counts days of planning, not different dates.
Does this tell me what my home will save?
Not exactly, since your home, tariff, battery and roof are your own. The calculator gives a figure and a range for your use and battery. Once you connect, Savings compares your battery with a copy of it following Orza’s plan, from your own readings.
See what Orza would do with your own battery.
Free, with no card. Look round an example home first, and connect your battery when you’re ready.