A Keltner Channel strategy can be traded two opposite ways: buy when price breaks above the upper channel (trend), or buy when it drops below the lower channel (mean reversion). We tested both, plus a short breakout, on Nasdaq-100 futures (NQ) from June 2010 to April 2026 using the modern 20 EMA ± 2 ATR channel. The answer to "Keltner breakout vs fade" on NQ was clear: fade the lower band, don't chase the upper one.
From Chester Keltner to the ATR channel
Keltner Channels are commonly credited to Chester W. Keltner, who described a "ten-day moving average trading rule" in his 1960 book How to Make Money in Commodities. His original version used a 10-day average of the typical price with bands based on the average daily range. The version most charting platforms use today swaps in an exponential moving average for the middle line and average true range (ATR) for the band width, a change usually credited to Linda Raschke in the 1980s.
The result is an ATR channel that works a lot like Bollinger Bands, with one difference: Bollinger Bands use standard deviation of closes, while Keltner Channels use ATR, which includes gaps and intraday ranges. Keltner bands tend to be smoother and react less to a single big close.
The exact Keltner Channel rules we tested
- Market: NQ continuous front-month futures, daily bars, June 7, 2010 to April 16, 2026.
- Channel: 20-day EMA of the close ± 2 × ATR(20). ATR uses Wilder-style smoothing.
- Breakout long: buy at the close when it's above the upper channel. Exit at the close when it falls below the 20 EMA.
- Breakout short: short at the close when it's below the lower channel. Cover at the close when it rises above the 20 EMA.
- Fade long: buy at the close when it's below the lower channel. Exit at the close when it rises above the 20 EMA.
- Stops: none.
- Costs: 1.0 NQ point per round trip. Results are NQ points per 1 contract; MNQ = $2/point, NQ = $20/point.
Keltner breakout vs fade: results on NQ
| Variant | Trades | Win rate | Avg pts | Total pts | Profit factor | Closed-trade DD (pts) | Pts 2010–19 | Pts 2020–26 | PF 2020–26 |
|---|---|---|---|---|---|---|---|---|---|
| Breakout, long | 76 | 51.3% | 24.88 | 1,891.0 | 1.27 | -2,234.5 | 1,034.0 | 857.0 | 1.16 |
| Breakout, short | 55 | 18.2% | -177.93 | -9,786.0 | 0.28 | -9,786.0 | -3,459.2 | -6,326.8 | 0.35 |
| Lower-channel fade, long | 55 | 81.8% | 175.93 | 9,676.0 | 3.59 | -1,263.8 | 3,397.2 | 6,278.8 | 2.81 |
Buy and hold over the same span: 24,677.2 points, worst decline 5,878.2 points.
One signal, two opposite trades
Notice something in the table. The breakout short and the lower-channel fade use the same entry (close below the lower channel) and the same exit (close back above the 20 EMA). They're the exact same trades taken in opposite directions. So their results are mirror images: the fade made 9,676.0 points and the short lost -9,786.0. The 110-point difference in size is simply the 1-point cost, charged once on each version's 55 trades.
That makes the lesson unusually clean. When NQ closed below its lower Keltner Channel, the next move back to the 20 EMA was usually up, not down. The fade won 81.8% of the time with a profit factor of 3.59 and 9,676.0 points ($19,352 on one MNQ contract).
The upside breakout: weak
Buying closes above the upper channel made 1,891.0 points over 76 trades, a profit factor of just 1.27. It had big swings: -1,340.5 points in 2020, 2,290.2 in 2023, -958.2 in 2024. For a trend strategy on a market that rose this much, that's a poor result. Our read: by the time NQ closes two ATRs above its 20 EMA, much of the short-term move may already be done.
The fade's weak spot
The fade's average loss (-373.8 points) was bigger than its average win (298.09). With no stop, a close below the lower channel during a real selloff just keeps going. 2020 was the fade's worst year at -887.8 points.
The robustness check: 2010–2019 vs 2020–2026
NQ climbed from about 1,800 to about 26,000 over the test, so point totals lean toward recent years. The lower-channel fade made 3,397.2 points in 2010–2019 and 6,278.8 in 2020–2026 (profit factor 2.81 in the later period). It worked in both decades. The upside breakout made 1,034.0 and 857.0; notice it earned fewer points in the higher-priced period, with a 2020–2026 profit factor of just 1.16. The short breakout lost in both.
Sample size: 55 fade trades and 76 breakout trades are above the 30-trade line we use as a minimum, but still modest. Treat the profit factors as rough.
What this backtest does not model
- Open-trade drawdown. With no stops, the closed-trade drawdown understates the real pain of holding a losing fade.
- Continuous-contract roll gaps, which inflate ATR and can move the channel on roll days.
- Fills at the exact daily close.
- Other settings (10-period, 1.5 ATR, intraday charts) and Chester Keltner's original typical-price version.
- Overnight gaps, margin and taxes.
This is a historical backtest. Past results don't guarantee future results, and nothing here is financial advice.
Key takeaways
- Keltner Channel (20 EMA ± 2 ATR) on NQ, 2010–2026, after 1 point of costs per trade.
- Fading the lower channel (long): 55 trades, 81.8% winners, PF 3.59, positive in both decades.
- Upside breakout (long): 76 trades, PF 1.27. Weak, and weaker after 2020.
- Downside breakout (short): the mirror image of the fade, and it lost -9,786.0 points.
- Average losses on the fade were larger than average wins, and there's no stop.
Put the research to work on your Desk
Explore the tools behind the research, compare the available access, or open your Desk if you already have an account.
FAQ
Is it better to trade Keltner Channel breakouts or fades?
On NQ daily bars from 2010 to 2026, fading was better. Buying closes below the lower channel had a profit factor of 3.59 over 55 trades; buying closes above the upper channel had a profit factor of 1.27 over 76 trades.
What Keltner Channel settings did you use?
A 20-day exponential moving average with bands at 2 × the 20-day average true range (ATR), on daily bars.
Who invented the Keltner Channel?
It's commonly credited to Chester W. Keltner, who described a ten-day moving average trading rule in his 1960 book How to Make Money in Commodities. The EMA and ATR version used today is usually credited to Linda Raschke.
What's the difference between Keltner Channels and Bollinger Bands?
Bollinger Bands use the standard deviation of closing prices to set band width; Keltner Channels use average true range. On NQ both gave a similar message: buying below the lower band worked better than chasing breakouts.
Loomi.Ai provides educational market research, not financial advice. Trading futures and options involves substantial risk of loss, and backtested results are historical and do not guarantee future returns.
Sources
- Loomi backtest: blog_strategy_tests.py + results JSON, Loomi research
- Keltner Channels, StockCharts ChartSchool
- Keltner Channel – Overview, How it Works, Uses, and Formula, Corporate Finance Institute