Control factor
Most price automations are on-off switches: run the water heater, or don't. The control factor is for the other kind: devices with a dial. Heating thermostat setpoints, EV charging current, fan speeds, or the color tone or brightness of a smart lamp that reflects the spot price. Instead of asking "is it cheap enough", you get a number that says relatively how cheap right now is, and you scale the dial by it.
The idea
Kilowahti ranks every slot of the day by price. The control factor turns that rank into a number between 0 and 1:
- 1.0 — the cheapest slot of the day
- 0.0 — the most expensive
- anything between — proportionally in between
"Cheaper is higher" is deliberate: it makes the factor usable directly as a multiplier. Charging current becomes max_current × factor, no inversion needed.
Each factor also has a bipolar twin running from −1 to +1, for when you want a signed number: positive means cheaper than the middle of the day, negative means more expensive.
Which control factor to use
Three factors, each ranking by a different price:
| Sensor | Ranks by | Use when |
|---|---|---|
control_factor_price |
Energy price only | Your grid fee is flat, or you only care about the market |
control_factor_total |
Energy plus transfer | You want the real cost per kWh — usually the right one |
control_factor_transfer |
Transfer tier only | You want to dodge grid peak tariffs specifically |
They all share one scale, and the configured curve and scaling apply to all three.
On a fixed-price contract
During a fixed-price period every hour costs the same, so control_factor_price sits at 1.0 all day — correct, since no hour is more expensive than another. The total price factor control_factor_total still varies if your transfer tariff varies during the day.
Curve and scaling
Two settings change how the rank is mapped onto the 0–1 range. Both live under Thresholds & control in the setup wizard and in Configure.
Curve — linear spreads the factor evenly across the day. sinusoidal is flat near the extremes and steep in the middle, so the cheapest few slots all score nearly 1.0, the most expensive few nearly 0.0, and the change happens across mid-range prices.
Scaling — an exponent from 1 to 3 applied to the result. Higher values push everything except the cheapest slots downward, making the factor behave more like a switch.
Values for a 24-slot day:
| Rank | Linear | Sinusoidal | Linear, scaling 2 | Sinusoidal, scaling 2 |
|---|---|---|---|---|
| 1 | 1.00 | 1.00 | 1.00 | 1.00 |
| 3 | 0.91 | 0.98 | 0.83 | 0.96 |
| 6 | 0.78 | 0.89 | 0.61 | 0.79 |
| 9 | 0.65 | 0.73 | 0.43 | 0.53 |
| 12 | 0.52 | 0.53 | 0.27 | 0.29 |
| 15 | 0.39 | 0.33 | 0.15 | 0.11 |
| 18 | 0.26 | 0.16 | 0.07 | 0.03 |
| 21 | 0.13 | 0.04 | 0.02 | 0.00 |
| 24 | 0.00 | 0.00 | 0.00 | 0.00 |
Reading across: at rank 6 — a fairly cheap hour — linear says 0.78 while sinusoidal says 0.89, treating it as nearly as good as the best. At rank 18 they diverge the other way, 0.26 against 0.16. Add scaling 2 and rank 12, the middle of the day, drops from 0.52 to 0.27.
Start with linear and scaling 1. Move to sinusoidal if your automations fidget between adjacent ranks; raise the scaling if you want the response concentrated on genuinely cheap hours.
Using the control factor
The pattern is always the same: pick a desired range for your device, and place the factor inside it.
Example: Heating setpoint, 19 °C when expensive, 22 °C when cheapest:
Rank 1 gives 22.0 °C, rank 12 gives 20.6 °C, rank 24 gives 19.0 °C.
Example: EV charging current, between 6 A and 16 A:
16 A at the cheapest hour, 11.2 A mid-day, 6 A at the worst.
Both go in a number.set_value or climate.set_temperature action, triggered however you like — a time pattern every 15 minutes, or Kilowahti's own triggers. Automation examples has these written out in full.
Guard templates against missing values by setting a default
Always guard by adding | float(0) to the template, which converts the template result to a decimal number, or sets it to zero in a problem case. The control factor sensor is Unavailable until the first prices arrive, and an unguarded template throws an error on restart.
When not to use the control factor
The control factor answers "how cheap is now, relative to the rest of today". It says nothing about absolute price. On a day where every hour is expensive, the cheapest hour still scores 1.0. If your device absolutely must stay off when electricity is simply costly, gate it on the price threshold — binary_sensor.kilowahti_{name}_price_acceptable or the "price is acceptable" condition — and use the control factor only to modulate what happens after that gate opens.