Does the Supertrend indicator actually work?
We traded every Supertrend colour flip on the Nasdaq 100 and gold — 10,462 trades on three timeframes — and compared the result with random entries, before and after 2024.
01What Supertrend is
Supertrend is a line that trails price at a distance of three times its average range. When price closes through it, it flips colour.
In an uptrend the line sits below price and is green. It ratchets up with price but never down.
When a candle closes below the green line, the trend is called over: the line jumps above price and turns red. The same happens in reverse for a downtrend.
The usual advice: buy when it turns green, sell when it turns red. That is exactly what we tested — with the default settings (ATR 10, factor 3) that most charting platforms use.
02What we tested
One fixed rule: trade every colour flip, always in the market. No settings were tuned.
- Buy when the Supertrend turns green, sell short when it turns red
- Exit at the next flip — and enter the other way at the same moment
- No extra stop, no profit target, no filter
We ran this on two markets and three timeframes:
Markets: Nasdaq 100 and gold (XAU/USD). Timeframes: 15 minutes, 1 hour, 4 hours. Results are shown per timeframe with both markets pooled — 3 variants in total. The 1-hour chart was set as the main test before any result was known.
03How a trade is counted
A flip is only known once the candle has closed. So every trade enters at the open of the next candle.
| Entry | Flip on a closed candle, entry at the open of the next candle |
|---|---|
| Exit | Open of the candle after the opposite flip |
| Risk unit | 1R = 3 × ATR(10) at the flip — roughly the distance from price to the Supertrend line |
| Stop | No separate stop, only the opposite flip. A trade can lose more than 1R (worst: -3.1R) |
| Holding time | On average 39.1 candles; nights and weekends included |
We measure in R, not dollars. +0.1R per trade means: on average you win a tenth of your risk unit.
A Supertrend system is a classic trend follower: many small losses, a few large wins.
04Costs
Costs are small on the slow charts. On the 15-minute chart they decide the result.
Every trade pays the real spread at entry plus 0.5 price units slippage (0.5 index points on the Nasdaq, $0.50 on gold). On the 1-hour chart that is a median of 1.2 % of the risk on the Nasdaq and 3.2 % on gold.
The 15-minute chart flips about 4 times as often as the 1-hour chart. Before costs it made +296R; after costs it ended at -125R.
05Build first, judge later
2020–2023 is the period anyone could have looked at. Only 2024–2026 shows whether it lasts.
We split the data in two. 2020–2023 is the "looks good on the chart" period. 2024 – June 2026 is the test: what would have happened if you had started trading it after seeing those first four years.
Before 2024, the 1-hour Supertrend was the best of the three timeframes: +0.087R per trade after costs, +103R in total. Then it stopped.
From 2024 on, the 1-hour version ended at -7.9R on 772 trades. Its win rate fell from 39.3 % to 35.9 %.
06The random test
A rule only counts if it beats random trades with the same direction, the same hour and the same holding time.
The Nasdaq and gold both rose strongly in this period. Long trades had the wind at their back. So "it made money" proves nothing.
That's why every real trade got a random twin: same market, same direction, same entry hour, held for the same number of candles — but on a random weekday of the same period. We built 1,000 such random versions of the whole 1-hour system.
"Better than 95 % of random versions" means only 1 in 20 random versions did as well. The p-value is the other side of the same coin: p = 0.05 means 5 % did as well. 50 % means the rule is indistinguishable from chance.
| 1-hour Supertrend | Trades | Per trade | Random median | Beats random | p-value |
|---|---|---|---|---|---|
| 2020–2023 | 1,184 | +0.087R | -0.027R | 98.6 % | 0.015 |
| 2024 – June 2026 | 772 | -0.010R | -0.005R | 46.9 % | 0.531 |
| Whole period | 1,956 | +0.049R | -0.018R | 96.2 % | 0.039 |
Over the whole period the result still looks "significant" (p = 0.039). That is the trap: the number is carried entirely by the years before 2024.
07Which timeframe?
The ranking of the timeframes turned upside down after 2024. That is what chance looks like.
| Timeframe | 2020–2023 | Beats random | 2024–2026 | Beats random |
|---|---|---|---|---|
| 15-minute | -0.051R | 69.5 % | +0.040R | 98.5 % |
| 1-hour | +0.087R | 98.6 % | -0.010R | 46.9 % |
| 4-hour | -0.011R | 48.0 % | +0.138R | 76.0 % |
Yes, the 15-minute chart beat random entries after 2024 (p = 0.020). But before 2024 it was the worst of the three, at -248R. Nobody would have picked it in January 2024.
With three timeframes and two periods, one result like that is not surprising. A real edge would show up in the same place twice.
The 4-hour chart has the best number after 2024 (+0.138R), but only 200 trades and no clear difference from random (76.0 %).
08Long, short and markets
The long trades made the money. In two markets that mostly went up, that is not a signal.
Buying the green flip made about +0.17R per trade in both periods. Shorting the red flip broke even before 2024 and averaged -0.20R per trade after.
The random twins trade in the same direction, so the random test already accounts for this rising market. The question is whether a flip picks better moments than chance — and after 2024 it didn't.
| 1-hour, per market | 2020–2023 | 2024–2026 | Beats random after 2024 |
|---|---|---|---|
| Nasdaq 100 | +0.111R | -0.056R | 24.7 % |
| Gold | +0.058R | +0.044R | 68.6 % |
Neither market on its own beat random entries after 2024.
09Verdict
The standard Supertrend beat random entries on old data, then did no better than chance on 772 new trades. The timeframe that "worked" changed from one period to the next.
The Supertrend does what it promises: it keeps you in a trend until the trend breaks. What we could not find is a reliable advantage over entering at random times in the same direction.
If you trade the Supertrend
- Expect to be wrong most of the time. 62 % of trades lost. The system lives on a few long trends.
- Don't judge it on one good stretch. Four strong years were followed by two and a half flat ones.
- Avoid fast charts. On 15 minutes, costs took more than everything the signal earned.
10Limitations
- Only the default setting (ATR 10, factor 3) was tested. Other settings may behave differently — and tuning them adds its own risk of luck.
- No extra stop: a gap through the line can cost more than 1R. Nights and weekends are included.
- Two markets, both mostly rising from 2020 to 2026. Falling or sideways markets are under-represented.
- The broker's spread was partly recorded as zero in early years; then only the fixed 0.5 slippage was charged.
- Pooled results count each trade as 1R, even when both markets had a trade open at the same time.
- p-values are not corrected for testing several timeframes; the 4-hour chart has few trades.
The 60-second version
This study started as a Reel. Share it with someone who still buys every green flip.
Next in the research library
- Do fair value gaps actually work? We tested almost 100,000 No usable edge
- Can one 5-minute candle predict the next six hours? Yes — the direction carries
- Does the double top pattern actually work? Worse than random
- Do Bollinger Bands actually predict reversals? in the lab
- Does Fibonacci actually work in trading? in the lab
- Does the weather in New York actually affect the Nasdaq? in the lab