What are the best RSI settings? For a buy-the-dip system on Nasdaq futures, our backtest says the answer is a short lookback — RSI(2) or RSI(3) — for a simple reason: the classic RSI(14) almost never reaches oversold while the market is in an uptrend. We tested RSI(2), RSI(3), RSI(6) and RSI(14) at nine oversold thresholds on daily NQ futures from June 2010 to April 2026.
The Relative Strength Index was introduced by J. Welles Wilder in his 1978 book New Concepts in Technical Trading Systems, and 14 periods became the default most charting platforms ship with. Decades later, Larry Connors popularized very short RSI periods, especially RSI(2), for short-term mean reversion; Short Term Trading Strategies That Work (2008, with Cesar Alvarez) is the book most traders associate with it.
The rules: one system, four RSI periods
To compare the RSI period fairly we held everything else constant and only changed the lookback and the buy threshold:
- Market: NQ continuous front-month futures, daily bars, June 7, 2010 to April 16, 2026. Long only.
- Trend filter: the daily close must be above its 200-day simple moving average.
- Entry: buy at the close when RSI(n) closes below the threshold (5, 10, 15 … 45), for n = 2, 3, 6 or 14. RSI uses Wilder-style smoothing.
- Exit: sell at the close on the first day the close is above its 5-day simple moving average.
- Stop: none. One position at a time.
- Costs: 1.0 NQ point per round trip.
All results are in NQ index points for one contract. MNQ is $2 per point and NQ is $20 per point, so MNQ dollars below are simply points × 2. Every trade has 1.0 point deducted for costs (slippage plus commission). The entry and exit rules follow the version of the RSI(2) pullback system most often attributed to Connors; we did not verify every detail against the book.
RSI 2 vs RSI 14: how often does each one even fire?
This is the table that answers most of the “RSI period” debate. It shows the number of trades each setting produced over almost 16 years.
| Setting | <5 | <10 | <15 | <20 | <25 | <30 | <35 | <40 | <45 |
|---|---|---|---|---|---|---|---|---|---|
| RSI(2) | 82 | 166 | 225 | 284 | 328 | 372 | 412 | 455 | 492 |
| RSI(3) | 22 | 57 | 103 | 158 | 204 | 262 | 302 | 351 | 395 |
| RSI(6) | 0 | 3 | 11 | 34 | 62 | 116 | 162 | 217 | 285 |
| RSI(14) | 0 | 0 | 0 | 0 | 2 | 12 | 35 | 85 | 172 |
RSI(14) produced zero trades at thresholds of 5, 10, 15 and 20, only 2 at 25, and 12 at Wilder's classic 30 line. In an uptrend (which is what the 200-day filter demands), a 14-day average of gains and losses is simply too slow to fall that far. RSI(2) below 10 fired 166 times. The shorter the lookback, the more sensitive the oscillator, and the more pullbacks it can actually catch.
Best RSI settings by the numbers
| Setting | Trades | Win rate | Total pts | Profit factor | Closed-trade DD (pts) | 2010–2019 pts | 2020–2026 pts |
|---|---|---|---|---|---|---|---|
| RSI(2) < 10 | 166 | 71.7% | +10,039.2 | 2.6 | 1,993.5 | +1,415 | +8,624.2 |
| RSI(2) < 25 | 328 | 73.2% | +11,374 | 1.82 | 2,495 | +1,777.2 | +9,596.8 |
| RSI(3) < 10 | 57 | 71.9% | +2,078 | 1.52 | 1,993.5 | +45.8 | +2,032.2 |
| RSI(3) < 25 | 204 | 72.1% | +12,002.5 | 2.7 | 1,993.5 | +1,738.2 | +10,264.2 |
| RSI(6) < 25 | 62 | 77.4% | +3,782.5 | 2.2 | 1,718.8 | +727.2 | +3,055.2 |
| RSI(6) < 35 | 162 | 72.8% | +8,473 | 2.19 | 1,993.5 | +1,854 | +6,619 |
| RSI(14) < 30 | 12 | 83.3% | +839.8 | 1.8 | 1,007.2 | +560.2 | +279.5 |
| RSI(14) < 35 | 35 | 85.7% | +3,408.8 | 3.86 | 1,007.2 | +1,494 | +1,914.8 |
| RSI(14) < 45 | 172 | 72.7% | +7,666.8 | 1.87 | 1,718.8 | +1,704 | +5,962.8 |
Some patterns stand out:
- Short periods won. RSI(2) < 10 made 10,039.2 points on 166 trades with a profit factor of 2.6. RSI(3) < 25 made 12,002.5 points on 204 trades with a profit factor of 2.7, the highest of any setting with 30 or more trades.
- Both decades were positive for the short-period settings in the table. RSI(2) < 10 made 1,415 points in 2010–2019 and 8,624.2 in 2020–2026. The 2020s dominate because NQ went from roughly 1,800 to roughly 26,000, so each point move got bigger.
- RSI(14) at 30 is not statistically meaningful. 12 trades over 16 years is far too few to draw a conclusion, whatever its win rate. Even RSI(14) < 35 had only 35 trades.
- RSI(14) works if you move the goalposts. At < 45 it produced 172 trades and 7,666.8 points — but at that point you are no longer buying “oversold,” you are buying any mild dip.
A fairer head-to-head is at similar trade counts. RSI(2) < 5 (82 trades) made 4,094.2 points, profit factor 1.86; RSI(14) < 40 (85 trades) made 3,449.2 points, profit factor 1.63. Closer than the headline numbers suggest, but the short period still came out ahead.
One more caution: a loss of -1,481.5 points in 2022 appears in the year-by-year results of many of the RSI(2), RSI(3) and RSI(6) variants. The identical figure suggests the same trade or trades, which is a reminder that a stop-less mean-reversion system can give back a lot on one bad dip.
Our take: The “RSI 2 vs RSI 14” argument is mostly about signal frequency, not magic. With a 200-day trend filter, RSI(14) rarely gets oversold, so it rarely trades. If you want a dip-buying signal on an index, RSI(2) or RSI(3) with a threshold between 10 and 25 was the most robust zone in this test. We would not chase the single best cell; neighbouring settings did similarly well, and that is what you want to see. The deeper look is in our RSI(2) strategy breakdown; for intraday use, see best RSI settings for day trading.
How this compares with buy and hold
Buying one NQ contract on June 7, 2010 and holding to April 16, 2026 gained 24,677.2 points, with a 5,878.2-point drawdown on the way. The best RSI settings here made roughly half of that, while only holding positions after pullbacks rather than all the time. That is the trade-off: less total profit, much less time exposed to the market.
Limitations
- This is a backtest; past results do not guarantee future results.
- Closed-trade drawdown ignores losses while a trade is open, and there is no stop, so real drawdowns were larger.
- Continuous-contract data contains roll gaps; fills are assumed exactly at the daily close. No taxes are modelled.
- Testing 36 combinations raises the risk that the best one looks good by luck.
Key takeaways
- With a 200-day filter, RSI(14) produced zero trades below 20 and just 12 trades below 30 in almost 16 years.
- RSI(2) < 10: 166 trades, 10,039.2 pts, profit factor 2.6; positive in both 2010–2019 and 2020–2026.
- RSI(3) < 25 had the highest profit factor (2.7) of any setting with 30+ trades.
- Short RSI periods fire more often, which is the main reason they beat RSI(14) here.
- 1.0 point per trade deducted for costs; MNQ = $2/pt, NQ = $20/pt.
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
What is the best RSI setting for swing trading index futures?
In our 2010-2026 NQ daily test with a 200-day trend filter, short periods did best: RSI(2) below 10 made 10,039.2 points (profit factor 2.6) and RSI(3) below 25 made 12,002.5 points (profit factor 2.7), after 1 point of costs per trade.
Why didn't RSI(14) work?
It barely traded. Combined with the rule that price must be above its 200-day average, RSI(14) produced zero trades at thresholds of 20 or lower and only 12 trades below 30 in almost 16 years - too few to be meaningful.
Is RSI 2 better than RSI 14?
For dip-buying in an uptrend, it was in this test, mainly because RSI(2) reaches oversold readings often enough to trade. It is not a universal rule; results come from one market and one set of rules.
Who created RSI?
J. Welles Wilder introduced the Relative Strength Index in his 1978 book New Concepts in Technical Trading Systems. Very short RSI settings like RSI(2) were popularized by Larry Connors.
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
- J. Welles Wilder Jr., Wikipedia
- New Concepts in Technical Trading Systems (J. Welles Wilder, 1978), Internet Archive
- Short Term Trading Strategies That Work (Larry Connors, Cesar Alvarez, TradingMarkets, 2008), Amazon listing