Timeframe

The timeframe is the resolution of the chart: a 1-hour chart draws one candle per hour. The same market looks like chop on 1-minute, a clean trend on 4-hour, and a single bar on 1-week. In that sense the timeframe is part of the strategy definition, not a display setting.

What the choice trades off

Lower timeframes (1m to 15m) give more signals and faster feedback, but each trade's edge is smaller while spread, fees, and slippage stay constant, so costs eat a larger share. Noise dominates structure down there, too. Higher timeframes (4h to 1d) offer fewer, larger moves and costs that amortize over bigger wins, but sample sizes build slowly and stops are wide in dollar terms.

Retail systematic trading mostly lives between the extremes (5m to 1d), where edges survive costs and a backtest can still accumulate hundreds of trades.

Multi-timeframe strategies

Many systems read several resolutions at once, such as a 4-hour trend filter gating 15-minute entries. The subtlety is alignment: at any moment, only the last closed higher-timeframe candle is known, and using the currently-forming one leaks the future. AlphaProve strategies declare additional timeframes explicitly (additional_timeframes=("4h",)) and the engine serves each view with only its completed bars, so the leak can't happen by accident.

On AlphaProve

Twelve timeframes are available — 1m, 3m, 5m, 15m, 30m, 1h, 2h, 4h, 6h, 8h, 12h, 1d — and every higher one is aggregated on the fly from finer bars, so you're never limited to whatever base resolution was stored. In the no-code builder the multi-timeframe access is a per-indicator higher-timeframe label rather than a code declaration, which lets a 4h trend filter gate a 15m entry without any hand-written alignment logic.