How we work out the savings

Step 1 — How much sun

We use NREL's PVWatts data. For Orange County, a kilowatt of panel at optimal tilt produces roughly 1,586 kilowatt-hours a year. Coastal California varies by about ten percent either side of that; the Central Valley is higher, the far north lower.

That's the ceiling. Everything after this reduces it.

Step 2 — How well it's pointed

The vertical figure surprises people, and it's why we push tilt brackets so hard. Hanging a panel flat against a railing throws away nearly 40 percent of what the same panel would make angled up. A $60 bracket recovers most of it.

Step 3 — Shade

These are deliberately coarse. Real shade analysis needs a sun path study, and the honest position is that a three-way question gets you close enough to know whether to buy.

Step 4 — How much of it you actually use

This is the number that most calculators skip and it's worth more than the panel choice.

We nudge these up slightly for households with large bills, since a bigger baseline load absorbs more of a small panel's output, and down slightly for very small bills. The nudge is capped at 92 percent — a big bill can never imply a perfect match.

Step 5 — What a kilowatt-hour is worth

We use a blended average rather than a time-of-use figure by default, because most people don't know their rate plan and a blended average is the honest middle. On the utility pages we show what changes if you're on a time-of-use plan, where offsetting evening consumption is worth substantially more.

Exported power is valued at zero. SB 868 provides no export compensation. If you already have rooftop solar on a net billing tariff with a bidirectional meter, your situation is different — see this guide.

The whole thing in one line

annual saving = 1,586 × kW × direction × mounting × shade × self-consumption × rate

A 900-watt kit, west-facing, on a railing, wide open sky, someone home most days, on SCE:

1,586 × 0.9 × 0.83 × 0.62 × 1.00 × 0.75 × $0.332 ≈ $183 a year

Add a tilt bracket and the mounting factor goes from 0.62 to 1.00, and the same kit makes about $295. That single change is worth more than a second panel.

What we don't model

• Weather in a particular year. PVWatts uses typical meteorological data. A wet winter costs you.
• Dust. A dirty panel loses a few percent. Rinse it occasionally.
• Inverter clipping. On the brightest days a panel can briefly out-produce its inverter. Small effect, real.
• Degradation. Panels lose roughly half a percent of output a year. We ignore it in year one.
• Rate changes. California electricity prices have risen steadily. We hold the rate flat, which makes our figures conservative.
• Any tax credit. The federal 30 percent credit ended for expenditures after 31 December 2025 and there's no state equivalent. If a figure elsewhere looks better than ours, check whether it's still subtracting it.

Measured versus estimated

Everything above is a model. Models are wrong in interesting ways.

We're running a kit on a west-facing balcony in Irvine and logging what it makes.

When there's enough data we'll publish the comparison, including if it makes our model look bad. That's the point of measuring.