Is it worth buying cheap power overnight and selling in the day?

Sometimes — but the honest arithmetic is stricter than the forum posts suggest. Grid arbitrage pays only when the price spread beats round-trip losses (10–15%) plus battery wear, and export prices in most markets sit well below import prices most of the time. Treat it as a seasoning on top of self-consumption, not the main course.

Why the maths is tighter than it looks

Every stored kWh loses 10–15% to charge-discharge efficiency, so a cheap-window purchase must be resold (or self-consumed) at a price at least that much higher just to break even — before counting cycle wear on a battery with a finite cycle life. The classic profitable pattern is not “buy at night, export at noon” (midday export is usually the day’s worst price, thanks to everyone’s solar) but “buy at night, skip the evening peak”: self-consuming stored cheap power against a high import price is a far wider spread than any export deal.

Genuine export arbitrage exists at the edges: dynamic export products occasionally spike during evening scarcity (UK Agile Outgoing-style, Amber in Australia), and a battery that exports into those hours can earn well. But those spikes are episodic; a strategy needs them, plus automation fast enough to catch them.

What to do

1. Compute your break-even spread: cheap import price ÷ 0.85 (round trip) plus a few pence/cents for wear. Only spreads above that line earn anything.
2. Aim the discharge at your own evening peak, not at export — avoided import is almost always the better “sale”.
3. Coordinate with the solar forecast — overnight arbitrage and next-day sun compete for the same battery space; the overnight-vs-solar answer gives the decision rule.
4. Let automation do the catching — spike-riding only works when Home Assistant or the inverter’s own scheduler reacts to prices; if the plan is manual, halve the expected profit before deciding.