Supertrend Strategy: Trend-Following Trading with ATR Signals
Supertrend combines an ATR-based band with a persistent direction state. A color change can define a strategy event, but it does not by itself establish a profitable entry or place an order. The calculation, confirmation time and exit mechanism must all match the test.
How Supertrend Is Calculated
The basic bands start at the bar midpoint, (high + low)/2, plus or minus ATR times a multiplier. The active bands also depend on their previous values and the previous close. Supertrend selects the lower band for its up state and the upper band for its down state. TradingView's formula documents the persistence and switching rules.
This is not simply a fresh ATR envelope around every close. During an established state, the retained band may stay fixed even as the newly calculated basic band moves away. An increase in ATR does not guarantee that an existing protective stop should widen.
ATR measures range in the chart's price units over bars. An ATR on a five-minute chart is not a daily range. The multiplier determines the basic band's distance; the ATR length determines its smoothing history.
Define the Signal Event
A bullish event is a change from the down state to the up state. A bearish event reverses it. Remaining in an up state for ten bars should not be counted as ten new bullish flips unless that is an intentional entry rule.
Decide whether to act while the current candle is developing or only after it closes. Intrabar high, low and close can change the calculation and reverse a provisional signal. A confirmed-bar policy delays action but makes its timing easier to reproduce.
Waiting for two consecutive up-state closes is another rule. It can remove some short-lived flips and miss part of other moves. No documented 30% reduction in false signals is established here.
A Baseline Strategy Specification
The following is an untested long-only research example:
| Decision | Rule to record |
|---|---|
| Inputs | ATR length 10 and multiplier 3 as trial values |
| Entry | First completed bullish state change while flat |
| Fill | Next available market fill after confirmation |
| Trend exit | First completed bearish state change |
| Risk | Explicit initial stop and position sizing from its distance |
| Re-entry | Next qualifying flip after the previous position closes |
Do not combine a close-based exit with the assumption of an intrabar fill at the drawn line. A separate stop order can trigger before the close; a close-confirmed flip is only known later. Specify the priority if both conditions occur in one bar. TradingView explains the simulator's assumptions in its strategy documentation.
Settings and Additional Filters
Compare a small prespecified set of lengths and multipliers, changing one dimension at a time. A larger multiplier places basic bands farther from the midpoint; it does not guarantee fewer losses or better net return. A longer ATR length changes smoothing, not the price series' underlying trend quality.
A 50-EMA filter can require price above the EMA on a bullish flip. An RSI filter needs a precise threshold, and a MACD rule must name the two crossing series and whether their event must coincide with the flip. MACD is calculated from prices; it does not observe momentum before price changes occur.
Record signals rejected by each filter. Compare avoided losses, missed winners, exposure and costs against the baseline. A filter that leaves only a few trades can make its apparent win rate unstable.
Risk and Trailing Stops
Size from the actual entry-to-stop distance, including contract point value where relevant. If the initial stop is below a swing low, define when that swing becomes known; a pivot confirmed by later bars cannot be used earlier.
When using Supertrend as a trailing reference, update the order only when the intended value becomes available. Decide whether the stop may loosen, and retain the previous stop if the rule permits tightening only. A drawn line does not ensure an executable price through gaps.
Partial exits change both the average win and the remaining exposure. Specify their quantity and trigger before testing; do not retrospectively select a profitable resistance level.
What a Performance Report Must Show
No verified BTC 60% capture rate, seven-year profit factor of 1.07, or client result of 1.44 and 11.61% is provided here. Those numbers require a complete strategy, underlying data and trade records.
Profit factor is gross profit divided by gross loss. It is not dollars returned per dollar of initial risk, and a single aggregate value cannot demonstrate steady profitability through every year. Review year-by-year results, drawdown, trade count, average outcomes and costs.
Use a later evaluation period excluded from parameter selection. Check the result on standard candles and inspect individual fills. Then forward-test the fixed rules without editing historical signals. Pineify can help construct a strategy from these conditions, but generated code and syntax validation do not verify a trading edge.
