Timeframes
Timeframes are picked when creating a backtest or bot; multi-timeframe logic can be used inside buy/sell conditions (e.g. higher-timeframe filters) or as analysis reference. See in-app for what is actually supported.
Pluto distinguishes two layers of K-line timeframe in a strategy: the indicator timeframe declared inside each condition (each buy / sell condition can pin its own K-line, e.g. MA(20) @ 1H, MA(50) @ 4H), and the execution main timeframe picked per pair when you create a backtest task or trading bot (the engine "ticks" at this period to evaluate the strategy). The two can differ—and that's exactly where multi-timeframe capability comes from. This page covers the difference between these two layers, common timeframes, how to pick one by trading style, multi-timeframe ideas, and usage tips.
Two timeframes: indicator timeframe vs. execution main timeframe
| Dimension | Indicator timeframe (condition-level) | Execution main timeframe (run-level) |
|---|---|---|
| Definition | The K-line period each indicator condition declares | The "tick" at which the engine evaluates the strategy and emits signals |
| Effect | Decides which K-line bucket the indicator is computed on | Decides how often buy / sell conditions are evaluated |
| Where to set | Inside the strategy editor, per condition | When creating a backtest task or trading bot, per pair |
| How many | A single strategy can reference multiple periods (this is how multi-timeframe works) | A single backtest / bot picks only one |
| Examples | Condition 1: MA(20)@1H crosses above MA(50)@1H; Condition 2: 1D bullish alignment as a filter | Main period = 1H: conditions evaluated on each 1H bar close |
In short: indicator timeframe = "which K-line do we judge on", execution main timeframe = "how often do we judge". They can match or differ—e.g. main period 1H, with conditions referencing 4H / 1D for higher-period filters, is a typical multi-timeframe confluence pattern.
Common timeframes and use cases
The system usually supports K-line periods from minute to daily (the actual list is whatever the system ships); they can serve both as the execution main timeframe and as indicator timeframes referenced inside conditions:
| Period | Typical use |
|---|---|
| 1 minute (1m) | Scalping, ultra short-term; many signals, more noise, sensitive to execution and cost |
| 5 minutes (5m) | Short-term, intraday; higher trade frequency |
| 15 minutes (15m) | Short-term, intraday to overnight |
| 1 hour (1H) | Medium / short-term; balance of signal frequency and stability; very common |
| 4 hours (4H) | Medium-term, swing; fewer signals, longer holds |
| 1 day (1D) | Long-term, trend; few signals, suited to trend following |
The same strategy often behaves differently on different main periods: shorter main = more signals, more false signals, higher cost impact; longer main = fewer signals, longer holds.
Choosing the execution main timeframe by style
Different trading styles map to different main-period ranges; consider instrument volatility and validate with backtest.
| Style | Common main period | Notes |
|---|---|---|
| Intraday | 1m, 5m, 15m | Open and close within the day; short period, high frequency; mind fees and slippage |
| Short-term | 15m, 1H | Hold hours to a few days; 1H is common |
| Swing | 1H, 4H, 1D | Hold days to weeks; 4H is a common middle ground |
| Trend | 4H, 1D | Long holds, few signals; focus on trend filtering and risk |
Backtest 1–2 candidate main periods and compare return and drawdown before fixing one; keep live / paper the same as backtest for clean attribution.
Multi-timeframe approach
When designing strategies, a "higher → medium → lower" hierarchy is common. In Pluto there are two ways to apply it:
- A. Reference multiple timeframes inside conditions (recommended): bring different-period indicators into conditions and AND them together. Example: with main period 1H, add a buy condition like
1D MA bullish alignmentas a higher-period filter. - B. Single main period + manual analysis: keep the strategy itself on a single timeframe and do multi-period reasoning in your analysis or sizing decisions, validated either subjectively or via backtest.
Common patterns:
- Trend confirmation: higher-period trend up + long signal on the lower period (golden cross, RSI oversold bounce, etc.) → go long; mirror for short.
- Confluence: same-direction signals across multiple periods (e.g. MACD golden cross on 1D / 4H / 1H) raise signal weight.
- Higher-period filter: only allow lower-period conditions to fire when the higher period meets some criterion (e.g. 1D MA bullish alignment), to cut counter-trend entries.
Which periods can be referenced and how to combine them depends on what's supported in the system. See Conditions and Indicators.
Usage tips
- Main period matches style: pick a main period that fits the strategy type and holding style (intraday → short period, trend → long period); avoid mismatches like a trend strategy on 1m.
- Validate the main period with backtest: performance differs a lot across main periods—backtest 1–2 candidates and compare before fixing one; keep live / paper the same as backtest for attribution.
- Don't overdo multi-timeframe: referencing 4+ periods inside conditions tends to overfit and is hard to maintain; 2–3 periods (e.g. 1D + 1H or 4H + 1H) is usually enough for typical higher-period filtering.
- Be wary of very short main periods: very short main periods (e.g. 1m) means many signals, lots of noise, and high cost ratios—use them only if the strategy and execution are specifically tuned for it.
Common questions
Can the indicator timeframe differ from the execution main timeframe? Yes—and that's exactly the multi-timeframe capability. With main period 1H, conditions can still reference MA(20)@4H or 1D bullish alignment; the engine ticks on the main period and reads each condition's declared K-line for the indicator computation.
What happens if I pick the wrong main period? Too short → too many signals, more false signals and higher cost; too long → too few signals, missed timing. Compare return, drawdown, and trade count across periods in backtest, then decide.
Should backtest and live use the same main period? Yes. Backtest on 1H, live on 15m makes the logic and statistics inconsistent and the results hard to compare. If you want to switch, re-backtest on the new main period first.
What if timeframes conflict? If the higher period is bullish and the lower bearish (or vice versa), stick with the higher period and only trade with the trend; you can also AND the higher-period condition into the buy / sell conditions so it acts as a filter, only firing when both align.
Next steps
- Conditions — Use multiple periods inside conditions
- Indicators — Indicators and available periods
- Run Backtest — Pick the execution main timeframe and backtest
- Live Trading — Pick the execution main timeframe when creating a bot