Supertrend

Follow the ATR trailing band. Few trades, long holds, brutal in chop.

What it is

Reference — Glowwiki: Supertrend
The definitions and formulas below follow that article. It states its own limit plainly:
“Whether this produces profit is not addressed in that article.”
That is the question this page answers.

Note from the reference: The name has no encyclopaedia or regulator definition — the reference establishes that explicitly, as it does for grid trading. What does have one is ATR, which the source defines as a smoothed moving average of true range. Because that smoothing has no fixed window, the band inherits the same history-length sensitivity as RSI. With no canonical definition, the basis, the multiplier and the flip rule all vary between implementations and the reference cannot say which is correct.

Industry usage. The name has no encyclopaedia or regulator definition; its component, ATR, does. Also known as ATR trailing band.

Formulas

Bands

basis = (H+L)/2 ; band = basis ± k·ATRₙ

This page uses the median price as the basis. With no canonical definition, other implementations differ.

ATR

ATRₙ = SMMA of true range over n

A smoothed moving average, not a fixed window — the same family as RSI

Established

Average true range is an n-period smoothed moving average of the true range values. [3]
ATR gives no indication of price direction, only the degree of price volatility. [3]
Because ATR is built from differences, adding or subtracting a constant from every price leaves it unchanged. Standard-deviation based volatility measures do not have that property. [3]

These rest on a single line of sourcing. The Glowwiki reference article draws on the same encyclopaedia cited here, so two domains do not make two independent accounts. Nothing below has been checked against a second, separately grounded source. It is listed as established because the sources state it as definition rather than as finding — not because it has been cross-verified.

Conventional reading

A close through the band is conventionally read as a trend flip. [1, 2]
The band width measures volatility only. Direction comes from the basis and the flip rule, not from ATR. [1, 3]
With no canonical definition, the basis, the multiplier and the flip rule all vary between implementations; the reference cannot say which is correct. What is measured here is this implementation. [1]
The usefulness of technical analysis as a whole is contested by the efficient-market hypothesis. The reference article states the dispute without ruling on it — and neither does this page. What is below is one measurement, not a verdict. [1, 2]

Listed as convention, not as established fact.

Why your chart may not match ATR is a smoothed moving average with no fixed window, so Supertrend carries the same history-length sensitivity as RSI — and here it is larger. Measured on BTC daily across the 365-bar rolling windows used on this site: mean relative difference 0.47% between computing inside the window and carrying full history, and the close-above-band test flips on 2.43% of bars, against 0.22% for RSI. The indicators on this site split cleanly: SMA, Bollinger, Stochastic and Donchian use fixed windows and are unaffected; EMA, MACD, RSI, Supertrend, SAR and ADX all use smoothing and are.

Sources

Supertrend — Glowwiki (en)
Technical analysis — Wikipedia (en)
Average true range — Wikipedia (en)

The short answer

Beat buy & hold in 1 of 3 markets tested.

ETH/USDT +1516% · BTC/USDT -246%. A rule that wins in one market and loses in another did not find an edge; it found a rising asset. Each figure is measured only against buying and holding that same market over that same period.

Market	Return	Buy & hold	vs B&H	Beat it	Max DD	Underwater	Grid beat
BTC/USDT	284%	530%	-246%	no	68%	840d	63%
ETH/USDT	1792%	276%	+1516%	yes	53%	926d	100%
Gold (PAXG)	17%	119%	-101%	no	23%	1189d	0%

All figures on this page are from daily candles unless a row says otherwise.

Evidence

Everything below is the same rules run over real price history. Pick a market to see the fills, the equity path and the parameter grid for it.

Daily candlesBTC/USDT · 3188 bars · 2018-01-01 → 2026-09-23

Total return

283.6%

10,000 → 38,365

vs buy & hold

-246.3%

over the same period 529.9%

Max drawdown

67.6%

10,000 → 3,236

Longest underwater

840

840 days · 98% of the period

Settings that beat it

10 / 16

of 16 parameter settings tested

Win rate

39%

95% CI 26%–55% · n=38

The win rate on this page is not a readable number. With 38 closed trades the 95% confidence interval runs 26%–55% — it spans a coin flip, so the headline 39% cannot be distinguished from chance at this sample size. It is shown because leaving it out would be worse, not because it supports a conclusion.

A 0% risk of ruin here is a property of the test, not a safety rating. These runs are long-only spot with no borrowed money, so there is no margin call and nothing can force the position closed. The account cannot be wiped out — it gets stuck holding the asset instead. Capital deployed (5×) and max drawdown (67.6%) are the numbers that carry the risk on this page. The same rules run on margin would be a different measurement, and this figure would not stay at zero.

Daily candles

account equitybuy & holddashed = starting capital

77 fills. Signals are taken on the close and filled at the next open.

Timeframe

Everything above uses the daily row. These are the same rules and the same parameters on a different candle size.

Timeframe	Return	Buy & hold	vs B&H	Beat it	Trades
4H	1149%	524%	+625%	yes	212
1D ←	284%	530%	-246%	no	38
1W	579%	524%	+55%	yes	5

Outcome distribution

A single return is one chosen start date. These are 95 runs of the same rules with the start shifted forward 30 days at a time.

Worst	p10	p25	Median	p75	p90	Best	Losing starts
-63.0%	-24.7%	-8.1%	26.5%	71.9%	149.5%	310.0%	32%

Does it survive its own parameters?

The headline above uses one parameter setting. Here is every setting tested: 16 combinations, median 599.2%, best 1205.3%, worst 281.9%. 10 of them beat buy & hold.

Parameter grid

Supertrend run across its own parameter range over the full period. The same rules flip from profit to loss depending only on the numbers you picked — which is why a single headline return is meaningless on its own.

	length 7	length 10	length 14	length 21
mult 2	1205%	650%	503%	607%
mult 2.5	808%	599%	757%	393%
mult 3	890%	284%	282%	476%
mult 4	870%	515%	983%	598%

Hover a cell for trade count and drawdown. “·” = invalid combination (fast ≥ slow).

Year by year

Year	2018	2019	2020	2021	2022	2023	2024	2025	2026
Return	-63%	104%	262%	48%	-51%	98%	27%	-18%	16%
Max DD	64%	44%	19%	33%	53%	20%	41%	28%	21%
Trades	5	3	2	4	6	3	6	5	3

Recent closed trades 38 closed trades — showing the last 38

Entered	Entry	Exited	Exit	Held	Size	P&L (USDT)	Return
2026-04-14	74,455.21	2026-05-23	75,501.73	39 bars	29,203.84	+351.31	+1.2%
2026-03-05	72,703.1	2026-03-30	65,977.92	25 bars	32,245.03	-3,041.19	-9.4%
2026-01-14	95,461.7	2026-01-21	88,383.45	7 bars	34,897.12	-2,652.09	-7.6%
2025-10-02	118,654.29	2025-10-11	112,718.1	9 bars	36,808.49	-1,911.37	-5.2%
2025-07-10	111,289.62	2025-08-22	112,443.75	43 bars	36,503.62	+304.87	+0.8%
2025-04-23	93,489.71	2025-06-23	100,913.39	61 bars	33,885.94	+2,617.68	+7.7%
2025-03-03	94,317.12	2025-03-11	78,556.56	8 bars	40,765.82	-6,879.88	-16.9%
2025-01-18	104,129.51	2025-02-25	91,507.1	38 bars	46,481.89	-5,716.07	-12.3%
2024-09-20	62,979.47	2024-12-24	94,834.03	95 bars	30,930.52	+15,551.37	+50.3%
2024-08-24	64,069.26	2024-09-06	56,151.91	13 bars	35,362.34	-4,431.82	-12.5%
2024-07-16	64,756.42	2024-08-04	60,667.64	19 bars	37,821.2	-2,458.86	-6.5%
2024-05-21	71,482.34	2024-06-22	64,111.49	32 bars	42,253.9	-4,432.7	-10.5%
2024-02-09	45,311.3	2024-03-20	61,906.44	40 bars	30,988.9	+11,265	+36.4%
2023-10-02	28,006.58	2024-01-23	39,548.24	113 bars	21,989.11	+8,999.79	+40.9%
2023-06-21	28,322.15	2023-08-18	26,610.1	58 bars	23,450.71	-1,461.6	-6.2%
2023-03-14	24,124.33	2023-04-22	27,249.21	39 bars	20,803	+2,647.71	+12.7%
2023-01-12	17,952.23	2023-03-09	21,693.52	56 bars	17,249.76	+3,553.24	+20.6%
2022-10-27	20,782	2022-11-09	18,536.11	13 bars	19,378.5	-2,128.75	-11.0%
2022-09-11	21,658.03	2022-09-21	18,864.87	10 bars	22,292.25	-2,913.75	-13.1%
2022-07-20	23,410.18	2022-08-20	20,823.97	31 bars	25,110.98	-2,818.73	-11.2%
2022-03-02	44,443.41	2022-04-09	42,230.89	38 bars	26,479.48	-1,368.49	-5.2%
2022-02-08	43,861.91	2022-02-22	36,989.66	14 bars	31,461.92	-4,982.44	-15.8%
2021-10-06	51,497.73	2021-11-19	56,863.17	44 bars	28,550.31	+2,911.61	+10.2%
2021-07-26	35,398.71	2021-09-11	44,819.78	47 bars	22,594.21	+5,956.1	+26.4%
2021-03-14	61,218.97	2021-04-22	53,760.73	39 bars	25,780.22	-3,186.01	-12.4%
2021-02-09	46,398.05	2021-02-27	46,253.74	18 bars	25,912.42	-132.2	-0.5%
2020-10-13	11,534	2021-01-22	30,836.56	101 bars	9,711.6	+16,200.82	+166.8%
2020-04-24	7,487.7	2020-09-04	10,133.22	133 bars	7,190.52	+2,521.08	+35.1%
2020-01-07	7,762.78	2020-02-27	8,781.61	51 bars	6,369.01	+821.51	+12.9%
2019-10-26	8,660.21	2019-11-22	7,623.98	27 bars	7,249.16	-880.14	-12.1%
2019-08-06	11,835.91	2019-08-29	9,716.14	23 bars	8,848.39	-1,599.23	-18.1%
2019-02-19	3,899.3	2019-07-15	10,169.09	146 bars	3,399.67	+5,448.72	+160.3%
2018-12-21	4,053.89	2019-01-30	3,408.33	40 bars	4,051.68	-652.01	-16.1%
2018-09-03	7,305.65	2018-09-07	6,513.58	4 bars	4,553.47	-501.79	-11.0%
2018-07-18	7,321.1	2018-08-05	7,006.34	18 bars	4,767.56	-214.09	-4.5%
2018-04-16	8,359.25	2018-05-12	8,401.74	26 bars	4,752.95	+14.61	+0.3%
2018-03-04	11,470.2	2018-03-10	9,225.39	6 bars	5,921.32	-1,168.37	-19.7%
2018-01-11	14,947.47	2018-02-03	8,868.59	23 bars	10,000	-4,078.68	-40.8%

Prices include slippage; size is what was actually committed to that trade including fees.

Assumptions

Period	2018-01-01 → 2026-09-23
Parameters	{"length":10,"mult":3}
Rolling test	365 bars × 95 start dates
Input hash	130e9786afb42c8c

Daily candlesETH/USDT · 3188 bars · 2018-01-01 → 2026-09-23

Total return

1792.0%

10,000 → 189,196

vs buy & hold

+1516.2%

over the same period 275.8%

Max drawdown

53.3%

10,000 → 5,948

Longest underwater

926

926 days · 97% of the period

Settings that beat it

16 / 16

of 16 parameter settings tested

Win rate

53%

95% CI 37%–69% · n=34

The win rate on this page is not a readable number. With 34 closed trades the 95% confidence interval runs 37%–69% — it spans a coin flip, so the headline 53% cannot be distinguished from chance at this sample size. It is shown because leaving it out would be worse, not because it supports a conclusion.

A 0% risk of ruin here is a property of the test, not a safety rating. These runs are long-only spot with no borrowed money, so there is no margin call and nothing can force the position closed. The account cannot be wiped out — it gets stuck holding the asset instead. Capital deployed (18×) and max drawdown (53.3%) are the numbers that carry the risk on this page. The same rules run on margin would be a different measurement, and this figure would not stay at zero.

Daily candles

account equitybuy & holddashed = starting capital

69 fills. Signals are taken on the close and filled at the next open.

Timeframe

Everything above uses the daily row. These are the same rules and the same parameters on a different candle size.

Timeframe	Return	Buy & hold	vs B&H	Beat it	Trades
4H	1363%	271%	+1092%	yes	212
1D ←	1792%	276%	+1516%	yes	34
1W	-90%	271%	-362%	no	1

Outcome distribution

A single return is one chosen start date. These are 95 runs of the same rules with the start shifted forward 30 days at a time.

Worst	p10	p25	Median	p75	p90	Best	Losing starts
-47.2%	-23.0%	-10.0%	16.0%	74.3%	405.0%	798.1%	31%

Does it survive its own parameters?

The headline above uses one parameter setting. Here is every setting tested: 16 combinations, median 1293.1%, best 4833.8%, worst 407.7%. 16 of them beat buy & hold.

Parameter grid

	length 7	length 10	length 14	length 21
mult 2	2920%	1108%	1116%	408%
mult 2.5	3339%	4834%	2210%	1669%
mult 3	2101%	1792%	1152%	1409%
mult 4	1293%	860%	1235%	1189%

Hover a cell for trade count and drawdown. “·” = invalid combination (fast ≥ slow).

Year by year

Year	2018	2019	2020	2021	2022	2023	2024	2025	2026
Return	-27%	16%	418%	125%	-31%	22%	-27%	-2%	41%
Max DD	46%	44%	31%	42%	42%	30%	53%	35%	20%
Trades	3	4	2	4	5	4	7	4	2

Recent closed trades 34 closed trades — showing the last 34

Entered	Entry	Exited	Exit	Held	Size	P&L (USDT)	Return
2026-03-16	2,179.38	2026-05-18	2,129.93	63 bars	127,243.58	-3,135.3	-2.5%
2026-01-14	3,327.49	2026-01-21	2,938.41	7 bars	144,380.72	-17,137.14	-11.9%
2025-10-07	4,686.36	2025-10-11	3,827.81	4 bars	177,118.45	-32,737.73	-18.5%
2025-07-10	2,770.12	2025-09-26	3,872.41	78 bars	126,955.06	+50,163.39	+39.5%
2025-05-09	2,208.49	2025-06-22	2,294.57	44 bars	122,437.11	+4,517.96	+3.7%
2024-11-08	2,896.92	2024-12-20	3,415.3	42 bars	104,061.18	+18,375.92	+17.7%
2024-10-21	2,748.28	2024-11-05	2,397.01	15 bars	119,549.8	-15,488.62	-13.0%
2024-09-24	2,648.3	2024-10-04	2,348.63	10 bars	135,073.92	-15,524.12	-11.5%
2024-07-16	3,484.94	2024-08-03	2,988.12	18 bars	157,847.7	-22,773.78	-14.4%
2024-05-21	3,663.62	2024-06-25	3,351.06	35 bars	172,915.81	-15,068.11	-8.7%
2024-02-13	2,661.32	2024-03-20	3,157.07	36 bars	146,054.88	+26,860.93	+18.4%
2023-10-24	1,766.35	2024-01-25	2,233.9	93 bars	115,717.2	+30,337.69	+26.2%
2023-10-02	1,734.67	2023-10-11	1,566.84	9 bars	128,368.65	-12,651.46	-9.9%
2023-06-22	1,890.04	2023-08-18	1,680.65	57 bars	144,651.36	-16,282.71	-11.3%
2023-03-14	1,674.28	2023-04-22	1,847.85	39 bars	131,326.59	+13,324.77	+10.2%
2023-01-10	1,321.06	2023-03-10	1,436.6	59 bars	121,006.19	+10,320.4	+8.5%
2022-10-26	1,460.47	2022-11-09	1,334.11	14 bars	132,732.2	-11,726.01	-8.8%
2022-07-17	1,356.52	2022-08-20	1,608.2	34 bars	112,184.17	+20,548.03	+18.3%
2022-03-25	3,112.31	2022-04-12	2,978.26	18 bars	117,468.05	-5,283.88	-4.5%
2022-02-08	3,141.34	2022-03-07	2,550.02	27 bars	144,996.98	-27,528.93	-19.0%
2021-10-15	3,792.72	2021-11-19	3,994.9	35 bars	137,934.12	+7,062.86	+5.1%
2021-07-27	2,228.92	2021-09-11	3,207.25	46 bars	96,051.32	+41,882.8	+43.6%
2021-03-14	1,923.31	2021-05-20	2,437.7	67 bars	75,934.85	+20,116.47	+26.5%
2020-10-22	390.51	2021-02-26	1,481.34	127 bars	20,057.74	+55,877.1	+278.6%
2020-04-19	187.49	2020-09-06	334.87	140 bars	11,252.73	+8,805.01	+78.3%
2020-01-11	144.9	2020-02-27	223.87	47 bars	7,298.1	+3,954.63	+54.2%
2019-10-30	190.55	2019-11-22	160.94	23 bars	8,657.93	-1,359.83	-15.7%
2019-09-18	207.95	2019-09-25	165.64	7 bars	10,891.61	-2,233.68	-20.5%
2019-04-03	163.34	2019-07-05	282.72	93 bars	6,305.26	+4,586.35	+72.7%
2019-02-11	125.18	2019-03-05	125.62	22 bars	6,296.03	+9.23	+0.1%
2018-12-21	113.77	2019-01-14	114.11	24 bars	6,289.5	+6.53	+0.1%
2018-09-22	247.79	2018-11-15	187.34	54 bars	8,335.92	-2,046.42	-24.6%
2018-04-15	502.93	2018-05-23	640.54	38 bars	6,558.2	+1,777.72	+27.1%
2018-01-11	1,249.65	2018-02-05	821.19	25 bars	10,000	-3,441.8	-34.4%

Prices include slippage; size is what was actually committed to that trade including fees.

Assumptions

Period	2018-01-01 → 2026-09-23
Parameters	{"length":10,"mult":3}
Rolling test	365 bars × 95 start dates
Input hash	d16cfa5428385e2a

Daily candlesGold (PAXG) · 2214 bars · 2020-09-01 → 2026-09-23

Total return

17.5%

10,000 → 11,749

vs buy & hold

-101.1%

over the same period 118.6%

Max drawdown

23.4%

10,000 → 8,366

Longest underwater

1189

1189 days · 98% of the period

Settings that beat it

0 / 16

of 16 parameter settings tested

Win rate

39%

95% CI 24%–58% · n=28

The win rate on this page is not a readable number. With 28 closed trades the 95% confidence interval runs 24%–58% — it spans a coin flip, so the headline 39% cannot be distinguished from chance at this sample size. It is shown because leaving it out would be worse, not because it supports a conclusion.

A 0% risk of ruin here is a property of the test, not a safety rating. These runs are long-only spot with no borrowed money, so there is no margin call and nothing can force the position closed. The account cannot be wiped out — it gets stuck holding the asset instead. Capital deployed (1×) and max drawdown (23.4%) are the numbers that carry the risk on this page. The same rules run on margin would be a different measurement, and this figure would not stay at zero.

Daily candles

account equitybuy & holddashed = starting capital

56 fills. Signals are taken on the close and filled at the next open.

Timeframe

Everything above uses the daily row. These are the same rules and the same parameters on a different candle size.

Timeframe	Return	Buy & hold	vs B&H	Beat it	Trades
4H	23%	119%	-95%	no	120
1D ←	17%	119%	-101%	no	28
1W	61%	123%	-62%	no	4

Outcome distribution

A single return is one chosen start date. These are 62 runs of the same rules with the start shifted forward 30 days at a time.

Worst	p10	p25	Median	p75	p90	Best	Losing starts
-12.7%	-6.5%	-2.2%	4.3%	10.6%	16.7%	33.8%	29%

Does it survive its own parameters?

The headline above uses one parameter setting. Here is every setting tested: 16 combinations, median 27.8%, best 93.0%, worst 13.5%. 0 of them beat buy & hold.

A strategy that only works at one setting did not work — it was fitted.

Parameter grid

	length 7	length 10	length 14	length 21
mult 2	27%	27%	28%	46%
mult 2.5	14%	16%	34%	47%
mult 3	14%	17%	22%	36%
mult 4	93%	81%	77%	69%

Hover a cell for trade count and drawdown. “·” = invalid combination (fast ≥ slow).

Year by year

Year	2021	2022	2023	2024	2025	2026
Return	-11%	5%	0%	25%	16%	-4%
Max DD	13%	14%	9%	9%	16%	23%
Trades	4	3	5	3	7	4

Recent closed trades 28 closed trades — showing the last 28

Entered	Entry	Exited	Exit	Held	Size	P&L (USDT)	Return
2026-08-06	4,274.19	2026-09-02	4,335.31	27 bars	11,606.3	+142.46	+1.2%
2026-04-08	4,812.15	2026-05-20	4,489.5	42 bars	12,465.32	-859.02	-6.9%
2026-03-01	5,372.47	2026-03-14	5,026.32	13 bars	13,350.47	-885.15	-6.6%
2026-01-13	4,617.61	2026-01-31	4,921.99	18 bars	12,549.94	+800.53	+6.4%
2025-11-29	4,234.95	2025-12-30	4,362.16	31 bars	12,208.35	+341.6	+2.8%
2025-08-29	3,417.52	2025-10-27	4,057.92	59 bars	10,302.27	+1,906.08	+18.5%
2025-07-23	3,426.5	2025-07-31	3,297.36	8 bars	10,727.19	-424.92	-4.0%
2025-06-14	3,463.71	2025-06-28	3,294.55	14 bars	11,300.56	-573.37	-5.1%
2025-04-11	3,216.61	2025-05-15	3,195.4	34 bars	11,398.33	-97.77	-0.9%
2025-03-14	3,004.5	2025-04-07	3,001.5	24 bars	11,432.57	-34.25	-0.3%
2025-01-17	2,718.36	2025-03-02	2,853.57	44 bars	10,912.65	+519.92	+4.8%
2024-11-23	2,723.36	2024-12-19	2,600.7	26 bars	11,450.23	-537.57	-4.7%
2024-07-12	2,393.2	2024-11-07	2,653.67	118 bars	10,346.98	+1,103.25	+10.7%
2024-03-05	2,066.03	2024-04-22	2,375.81	48 bars	9,015.86	+1,331.12	+14.8%
2023-12-28	2,053.03	2024-02-14	1,987.01	48 bars	9,334.07	-318.21	-3.4%
2023-11-23	1,982.99	2023-12-11	1,990	18 bars	9,319.8	+14.28	+0.1%
2023-10-14	1,917.96	2023-11-12	1,919.04	29 bars	9,333.19	-13.4	-0.1%
2023-07-27	1,956.98	2023-08-17	1,891.05	21 bars	9,677.89	-344.7	-3.6%
2023-03-11	1,860.93	2023-05-25	1,971.01	75 bars	9,155.66	+522.23	+5.7%
2022-11-09	1,701.85	2023-02-05	1,859.07	88 bars	8,398.16	+757.5	+9.0%
2022-10-04	1,696.85	2022-10-15	1,636.18	11 bars	8,726.98	-328.82	-3.8%
2022-08-05	1,789.89	2022-08-23	1,730.13	18 bars	9,046.49	-319.51	-3.5%
2022-01-20	1,847.92	2022-04-26	1,907.05	96 bars	8,783.58	+262.91	+3.0%
2021-10-26	1,813.91	2021-11-24	1,791.1	29 bars	8,913.21	-129.63	-1.5%
2021-09-04	1,830.91	2021-09-17	1,763.12	13 bars	9,274.48	-361.27	-3.9%
2021-04-19	1,801.3	2021-06-18	1,789.96	60 bars	9,351.9	-77.42	-0.8%
2020-12-18	1,905.15	2021-01-09	1,870.4	22 bars	9,544.71	-192.81	-2.0%
2020-09-11	1,954.51	2020-11-20	1,869.25	70 bars	10,000	-455.29	-4.5%

Prices include slippage; size is what was actually committed to that trade including fees.

Assumptions

Period	2020-09-01 → 2026-09-23
Parameters	{"length":10,"mult":3}
Rolling test	365 bars × 62 start dates
Input hash	30f20df67fe43a7b

Other strategies

Buy & Hold Martingale Grid DCA MA Cross EMA Cross RSI Reversion MACD Cross Bollinger Reversion Donchian Breakout Stochastic Cross Parabolic SAR