Why I show real numbers, not a claim

Many smart energy control products advertise a generic saving - "up to 30%", often without explaining how the number was calculated, which house it applies to, or over what period.

The numbers on this page are instead taken directly from a real house's own operational log, using a formula simple enough to recompute by hand: for each day, the power-weighted average price the system actually paid for purchased electricity is compared to that day's unweighted average price - the difference, times the kWh bought, is that day's saving. Summed over every day since the system started, that gives the total.

That also means the number can be small in some periods - for example early in the heating season, before heat control has been running long enough for its effect to show clearly in the statistics. I'd rather report that honestly (and label it "early season" when relevant) than show a promise that can't be verified.

The export script that computes the numbers, and the tables they come from, are in the source code - anyone can read exactly how the figures were produced.