DeployQuantRSI(14) Mean Reversion
Buy when the 14-day RSI drops below 30 (oversold), sell when it recovers above 70 (overbought).
This page tests one rule pair on 59 ETFs over the same window, from 2021-01-04 to 2026-10-02. The rule buys at the open after the 14-day RSI closes under 30 and sells the whole position at the open after RSI closes above 70. Each run starts with $10,000, uses 98% of the sleeve when it buys, takes no margin, and has no fees or slippage in the headline numbers. The two cost runs add 5 and 10 basis points.
Across the 59 funds the median CAGR is 2.98%, the median max drawdown is 24.63%, and the median Sharpe is 0.34. The median fund saw 15 round trips and was invested 37.2% of the time. The rule beat buy-and-hold of the same fund in 25 of 59 cases and had a shallower drawdown in 58. Positive CAGR showed up in 38 funds.
Those two counts say different things. A rule that sits in cash most of the time will nearly always have a shallower drawdown than the fund it trades, so 58 of 59 describes the exposure more than the signal.
For the single best rule-and-fund pairs, the RSI(2) snapback page uses a much shorter lookback, and the compare page for the two puts them side by side. The RSI(2) version has a higher median CAGR of 4.74% on the same universe.
The numbers here are one backtest on one 5.74-year window. The window starts in January 2021 and contains the 2022 bear market, so it has one large down year and several up years. Other windows would give different medians.
The rules
- WHEN the market opens · IF not invested AND RSI(14) < 30 · THEN buy with 98% of the sleeve
- WHEN the market opens · IF invested AND RSI(14) > 70 · THEN sell the whole position
A standard mean-reversion setup. The Relative Strength Index measures how stretched recent price action is. Readings under 30 have historically marked short-term washouts in uptrending assets. This template buys at the next session open and holds until RSI crosses back above 70. It has no profit target and no stop.
Good for: assets that trend up over time but overshoot on the way, such as broad index ETFs.
Watch out: in a persistent downtrend, RSI can stay oversold for weeks while the position keeps losing; there is no stop-loss in this template.
What the two rules do
The entry rule fires only when the strategy is flat. It reads RSI(14) at the close, and if the reading is under 30 it buys at the next open with 98% of the sleeve. The exit rule fires only when the strategy is long. If RSI(14) is above 70 it sells everything at the next open. There is no profit target, no stop and no time limit. A position can sit through a 30% fall as long as RSI stays under 70.
That design has a direct consequence for how results split across funds. The rule needs RSI to travel from under 30 to over 70 inside one holding period. On a fund that trends up with sharp pullbacks, the oversold reading comes near a low and the overbought reading comes after a recovery of several weeks. The exposure column shows how rarely that happens: the large index funds are in the market for roughly 27% to 29% of days for SPY, VOO and QQQ, and the median across all 59 is 37.2%.
On a fund that falls steadily, RSI can print under 30, recover part of the way, fall again and never reach 70. The position then stays open through the whole decline. The inverse and volatility funds show this pattern in the data, with exposure between 62% and 69% for most of them. The 3-month momentum page tests a rule with the opposite logic, and its median CAGR on the same universe is 0%, so neither style dominates across this universe.
A fund with RSI(14) below 30 has fallen on most of the last 14 sessions. Whether the next move is a bounce or a continuation depends on what drove the fall, and the rule has no way to tell. It treats a one-week selloff in a broad index fund the same as the first leg of a 70% decline in a leveraged inverse fund.
Build it from blocks (or type it in English), backtest it on 5.7 years of minute data in seconds, tweak any parameter, then paper trade it on live data. No card, no broker needed to start.
Results on every ETF
| ETF | CAGR | buy & hold | max DD | Sharpe | trades | win rate |
|---|---|---|---|---|---|---|
| TECL | 23.9% | 38.2% | −64.1% | 0.67 | 18 | 72% |
| FAS | 22.3% | 18.3% | −50.1% | 0.71 | 20 | 85% (+1 open) |
| SOXL | 14.4% | 33.3% | −81.0% | 0.54 | 14 | 71% |
| SOXX | 12.8% | 31.1% | −35.0% | 0.64 | 14 | 86% |
| ROM | 12.8% | 30.2% | −52.0% | 0.52 | 16 | 69% |
| TQQQ | 12.1% | 25.4% | −66.6% | 0.48 | 16 | 75% |
| XLF | 11.8% | 12.6% | −18.7% | 0.92 | 20 | 85% (+1 open) |
| EEM | 9.5% | 6.6% | −14.0% | 0.80 | 18 | 83% |
| QLD | 8.8% | 23.6% | −48.1% | 0.42 | 15 | 73% |
| XLK | 8.7% | 22.1% | −26.0% | 0.58 | 16 | 69% |
| VOOV | 8.0% | 12.6% | −13.1% | 0.86 | 17 | 76% (+1 open) |
| SSO | 7.4% | 21.9% | −34.9% | 0.41 | 13 | 77% |
| QQQE | 7.2% | 9.8% | −27.1% | 0.55 | 16 | 81% (+1 open) |
| VOOG | 7.1% | 15.7% | −21.6% | 0.53 | 16 | 75% |
| XLY | 7.0% | 6.3% | −20.6% | 0.49 | 15 | 73% (+1 open) |
| VTV | 6.8% | 13.4% | −13.1% | 0.72 | 16 | 88% (+1 open) |
| XLP | 6.5% | 5.8% | −9.6% | 0.75 | 16 | 81% |
| QQQM | 6.5% | 16.7% | −24.6% | 0.48 | 15 | 73% |
| QQQ | 6.4% | 16.7% | −24.4% | 0.47 | 15 | 73% |
| RINF | 6.3% | 6.6% | −9.8% | 0.93 | 15 | 93% |
| SPUU | 5.4% | 22.5% | −34.6% | 0.34 | 12 | 75% |
| VV | 5.0% | 14.3% | −17.7% | 0.47 | 14 | 79% |
| EEV | 5.0% | −16.1% | −56.8% | 0.31 | 17 | 65% (+1 open) |
| IAU | 5.0% | 13.7% | −15.7% | 0.53 | 17 | 82% (+1 open) |
| IOO | 4.5% | 16.7% | −18.8% | 0.43 | 13 | 85% |
| SPY | 4.4% | 14.6% | −17.1% | 0.42 | 13 | 77% |
| VOO | 4.2% | 14.4% | −17.3% | 0.41 | 13 | 77% |
| CLSE | 3.8% | 19.5% | −10.5% | 0.56 | 8 | 75% |
| USDU | 3.4% | 5.2% | −5.2% | 0.92 | 16 | 88% |
| CTA | 3.0% | 8.9% | −15.0% | 0.41 | 11 | 73% |
| VOX | 2.4% | 9.1% | −35.1% | 0.23 | 14 | 71% |
| QAI | 2.1% | 3.9% | −7.6% | 0.51 | 15 | 73% |
| ALTY | 2.1% | 7.6% | −11.3% | 0.32 | 14 | 64% (+1 open) |
| KMLM | 1.9% | 7.1% | −15.7% | 0.29 | 13 | 69% |
| TBF | 0.9% | 12.2% | −14.2% | 0.15 | 11 | 64% |
| IWM | 0.7% | 7.5% | −25.8% | 0.12 | 13 | 54% (+1 open) |
| UDN | 0.6% | −1.1% | −14.6% | 0.15 | 17 | 65% (+1 open) |
| FXE | 0.3% | −0.9% | −16.2% | 0.07 | 16 | 56% (+1 open) |
| SGOV | 0.0% | 3.2% | −0.0% | 0.00 | 0 | – |
| IEI | −0.2% | −0.4% | −8.7% | -0.05 | 16 | 69% (+1 open) |
| AGG | −0.2% | −0.8% | −11.6% | -0.03 | 15 | 67% (+1 open) |
| BND | −0.2% | −0.8% | −11.6% | -0.03 | 16 | 69% (+1 open) |
| IGIB | −0.5% | 0.2% | −13.6% | -0.09 | 13 | 62% (+1 open) |
| IEF | −0.8% | −2.3% | −13.6% | -0.13 | 15 | 67% (+1 open) |
| SPDN | −4.0% | −9.7% | −28.5% | -0.34 | 14 | 43% (+1 open) |
| SH | −4.3% | −10.0% | −29.5% | -0.37 | 14 | 43% (+1 open) |
| UST | −5.0% | −8.5% | −28.0% | -0.42 | 14 | 57% (+1 open) |
| TLT | −5.4% | −8.2% | −36.0% | -0.42 | 13 | 54% (+1 open) |
| PSQ | −5.7% | −13.8% | −40.9% | -0.33 | 16 | 50% (+1 open) |
| VIXM | −7.4% | −16.1% | −51.6% | -0.29 | 14 | 64% (+1 open) |
| VXZ | −8.1% | −15.1% | −49.9% | -0.35 | 14 | 64% (+1 open) |
| SDS | −11.0% | −21.6% | −56.8% | -0.47 | 14 | 43% (+1 open) |
| QID | −14.2% | −29.5% | −68.0% | -0.37 | 16 | 44% (+1 open) |
| REW | −18.5% | −36.1% | −76.5% | -0.40 | 17 | 47% (+1 open) |
| SQQQ | −23.6% | −42.3% | −85.2% | -0.38 | 16 | 44% (+1 open) |
| TMF | −25.0% | −31.2% | −81.1% | -0.67 | 13 | 46% (+1 open) |
| TECS | −29.1% | −46.7% | −89.5% | -0.41 | 18 | 50% (+1 open) |
| UVXY | −48.5% | −48.7% | −98.0% | -0.55 | 12 | 50% (+1 open) |
| SOXS | −53.7% | −48.3% | −99.5% | -0.52 | 15 | 40% (+1 open) |
Which funds the rule worked on
The top of the table is led by leveraged funds and a few sector funds. TECL returned 23.89% a year with a 64.11% max drawdown, against 38.16% for buy-and-hold. FAS returned 22.26%, ahead of its own buy-and-hold at 18.34%, with a 50.07% drawdown and an 85% win rate over 20 round trips. SOXL returned 14.43% against 33.32% for holding, with an 81.01% drawdown.
The pattern in those three is that the rule captured part of the return and gave up the rest while sitting in cash. TECL, SOXL and the other leveraged funds that top the table also had the deepest drawdowns, so the CAGR figure alone overstates how comfortable the path was. FAS is the only one of the top three where the rule beat holding on return.
The cleaner results sit lower. XLF returned 11.82% with an 18.73% drawdown and a 0.92 Sharpe, and buy-and-hold made 12.58%, so the rule gave up little return and kept the path shallow. EEM returned 9.51% against 6.58% for holding, with a 14.01% drawdown and 18 round trips. VOOV made 8.01% with a 13.06% drawdown, while holding made 12.6%. RINF won 93% of its 15 round trips, the highest win rate in the table, and returned 6.34% against 6.63%.
The broad index funds that most readers start with come out in the middle. QQQ returned 6.4% with a 24.35% drawdown against 16.7% for holding, and SPY returned 4.37% against 14.56%. Both had 13 to 15 round trips and were invested about 27% to 29% of the time. On those two funds, the rule gave up a large share of the buy-and-hold return in exchange for a shallower path.
The bottom of the table is dominated by inverse, volatility and long-bond leveraged funds. SOXS lost 53.71% a year with a 99.45% drawdown, and UVXY lost 48.55% with a 97.97% drawdown. TECS lost 29.14% a year and TMF lost 25.04%. SOXS is the one case where the rule did worse than holding, which lost 48.27% a year. On UVXY the rule and holding are almost level, at 48.55% and 48.74% lost a year.
One row has no trades. SGOV is a Treasury bill fund whose price barely moves, so RSI never fell under 30. The rule stayed in cash for the whole window, returned 0% with a 0% drawdown, and buy-and-hold made 3.22%. That row is counted among the 59 and pulls the median down slightly.
Results by fund type
| Fund type | ETFs | Median CAGR | Median buy & hold | Median max DD | Beat holding |
|---|---|---|---|---|---|
| Broad index ETFs | 12 | 6.5% | 14.4% | −18.8% | 1 of 12 |
| Sector ETFs | 6 | 8.7% | 12.6% | −26.0% | 2 of 6 |
| Leveraged ETFs | 10 | 12.1% | 23.6% | −52.0% | 3 of 10 |
| Inverse ETFs | 11 | −11.0% | −21.6% | −56.8% | 9 of 11 |
| Bond ETFs | 7 | −0.2% | −0.8% | −11.6% | 5 of 7 |
| Commodity ETFs | 1 | 5.0% | 13.7% | −15.7% | 0 of 1 |
| Currency ETFs | 3 | 0.6% | −0.9% | −14.6% | 2 of 3 |
| Volatility products | 3 | −8.1% | −16.1% | −51.6% | 3 of 3 |
| Alternative-strategy ETFs | 6 | 3.0% | 7.6% | −11.3% | 0 of 6 |
Results by fund category
Grouping the 59 funds by type shows where the median comes from. Leveraged ETFs have the highest median CAGR at 12.08%, against a median buy-and-hold of 23.55% for the same group. Their median max drawdown is 52%, and the rule beat holding on only 3 of the 10 funds. Sector ETFs have a median CAGR of 8.74% against 12.58% for holding, a 25.98% median drawdown, and 2 of 6 beats.
Broad index ETFs come next at 6.46% against 14.42% for holding, with an 18.82% median drawdown. Only 1 of 12 beat holding. This is the group the rule's description says it suits, and the data in this window does not show the rule keeping pace with the index. The 3-month momentum page, which is trend-following, uses the same broad funds and makes a useful second view of the same universe.
The alternative-strategy group has a median CAGR of 2.98% against 7.61% for holding, and the lowest drawdown among the equity-like groups at 11.28%. None of its 6 funds beat holding. The commodity group is one fund, IAU, at 4.99% against 13.66% and a 15.68% drawdown.
Currency ETFs show 0.62% against a buy-and-hold of negative 0.9%, with 2 of 3 beating holding. Bond ETFs have a median of negative 0.23% against negative 0.78%, with 5 of 7 ahead. Both groups moved little over the window, so the comparison is between two small numbers. The bond median drawdown is 11.61% and the currency median is 14.61%.
Inverse ETFs have a median CAGR of negative 11.04% and a median drawdown of 56.79%. The rule beat holding on 9 of 11, which sounds good until the level is read: buy-and-hold in that group had a median of negative 21.57%. Volatility products show the same shape, with a median of negative 8.13% against negative 16.06% and a 51.57% median drawdown. In both groups the rule lost less than holding because it was out of the market part of the time, and it still lost.
The category table makes one structural point. Mean reversion needs an asset with an upward drift, because the oversold bounce has to land above where the position was bought. Funds with a negative drift, such as inverse funds that decay daily, turn every oversold reading into a falling knife.
Year by year, median across all ETFs
| Year | RSI mean reversion | Buy & hold | ETFs with a gain |
|---|---|---|---|
| 2021 | 2.5% | 4.1% | 33 of 59 |
| 2022 | −4.0% | −12.7% | 27 of 59 |
| 2023 | 2.9% | 8.9% | 36 of 59 |
| 2024 | 5.3% | 9.7% | 48 of 59 |
| 2025 | 0.0% | 11.1% | 29 of 59 |
| 2026 | 2.1% | 3.7% | 34 of 59 |
Year by year across the 59 funds
The median calendar-year figures show how the rule behaved against holding in each regime. In 2021 the median fund returned 2.5% under the rule against 4.1% for holding, and 33 funds were positive. In 2022 the median under the rule was negative 4%, against negative 12.7% for holding, with 27 funds positive. This is the year where the rule paid for itself in relative terms, because it spent much of the decline in cash.
The recovery years show the cost of that cash. In 2023 the median was 2.9% against 8.9% for holding, with 36 funds positive. In 2024 the figures were 5.3% against 9.7%, with 48 funds positive, the highest count of any year. In 2025 the median under the rule was 0% against 11.1% for holding, and 29 funds were positive. The 2025 result is the widest gap in the table. The median fund was in the market for a short part of a year in which holding paid well.
The partial year 2026 shows a median of 2.1% against 3.7% for holding, with 34 funds positive. The year runs only to 2026-10-02, so it is not directly comparable with the full years.
The pattern across six years is consistent. The rule beat holding at the median only in 2022. In every other year the median fund did better by holding, and the margin was widest in 2025. That matches how the rule is built: it caps downside by staying in cash and gives up upside by waiting for an oversold signal before it is invested.
The 2022 result should be read with the exposure in mind. The median fund was invested 37.2% of days across the whole window. In a falling market with repeated RSI dips, the rule bought and held through several declines, so the smaller loss comes from the cash days and from the exits at RSI over 70, not from avoiding every drop. Individual funds can sit far from the median in any year. The ETF pages linked from the results table list year-by-year numbers for each fund.
Changing the parameters
| Version | Median CAGR | Median max DD | Median round trips |
|---|---|---|---|
| Published rules | 3.0% | −24.6% | 15 |
| RSI < 25 / > 70 | 1.3% | −21.5% | 11 |
| RSI < 35 / > 70 | 3.8% | −26.1% | 18 |
| RSI < 30 / > 65 | 2.6% | −20.4% | 17 |
| RSI < 30 / > 75 | 3.7% | −23.3% | 12 |
What changing the RSI thresholds does
The template uses 30 and 70. Four variants were run on every fund, each moving one threshold. The table above reports the median across the 59 funds for each.
Lowering the entry threshold to 25 gives a median CAGR of 1.32%, a median drawdown of 21.51%, and 11 round trips. The stricter entry means fewer trades and a smaller average exposure, and the median return falls from 2.98% to a lower figure. Raising the entry threshold to 35 does the opposite: a median CAGR of 3.77%, a median drawdown of 26.1%, and 18 round trips. More signals led to more return and a deeper path at the median.
Moving the exit threshold changes how long winners are held. Exiting at 65 gives a median CAGR of 2.59%, a median drawdown of 20.37%, and 17 round trips. Exiting at 75 gives 3.72%, a 23.33% drawdown, and 12 round trips. The later exit held positions through more of the recovery, so the median return rose and the trade count fell. The earlier exit cut trades short, which raised the count and lowered the return, while the drawdown was the shallowest of the five settings.
Two of the four variants beat the base rule on median CAGR: entry at 35 and exit at 75. Both let the position run for more of the move. Entry at 25 and exit at 65 both lowered the median, and both reduce time in the market. The data supports a direction, which is that more exposure helped at the median over this window, and it does not support a single best pair. The variant medians sit within a narrow band on CAGR, and each fund has its own best setting.
The RSI 30/70 pair comes from Welles Wilder's original description of the indicator. It was not tuned on this data. Changing it after seeing results on these 59 funds would fit the rule to one window, and the medians above are the better guide to how sensitive the rule is. In this test, moving any threshold by 5 points changed the median CAGR by less than two points either way. Funds with few trades, such as CLSE with 8 round trips, can change a lot from a single shifted signal. CLSE also has a shorter history than most of the funds.
The dip buyer is a different mean-reversion design with a price-based entry and a fixed profit target. Its median CAGR on the same universe is 1.33%. The weekly 7% target rule has the highest median of the other templates at 6.61%, and the monthly cycle rule follows at 5.46%. Both are linked from the facts as sibling strategies on the same 59 funds.
Frequently asked questions
What is the RSI mean reversion strategy?
Buy when the 14-day RSI drops below 30 (oversold), sell when it recovers above 70 (overbought). A standard mean-reversion setup. The Relative Strength Index measures how stretched recent price action is. Readings under 30 have historically marked short-term washouts in uptrending assets. This template buys at the next session open and holds until RSI crosses back above 70. It has no profit target and no stop.
Does RSI mean reversion beat buy-and-hold?
Across 59 ETFs backtested 2021-01-04 to 2026-10-02, it beat same-ETF buy-and-hold on 25 of 59 (42%). Median CAGR was 3.0% with a median max drawdown of 24.6%. Per-ETF results vary widely; the table lists every one.
Why RSI 30/70?
They are the conventional oversold and overbought bands from Welles Wilder's original formulation. Treat them as a starting point. In DeployQuant you can change them to 25/65 or anything else and re-backtest in seconds.
Does this strategy use a stop-loss?
No. The only exit is RSI recovering above 70. Adding a stop or a take-profit block is a one-block edit in the Lab.
Does RSI(14) mean reversion beat buy-and-hold?
In this test it beat buy-and-hold of the same fund in 25 of 59 cases. The median CAGR was 2.98%, and the median fund was invested 37.2% of the time. It had a shallower drawdown than holding in 58 of 59 cases, mostly because it sat in cash.
Which ETFs did the RSI 30/70 rule work best on?
TECL had the highest CAGR at 23.89%, followed by FAS at 22.26% and SOXL at 14.43%. FAS also beat its own buy-and-hold of 18.34%. Among funds with a drawdown under 20%, XLF at 11.82% and EEM at 9.51% stand out.
How many trades does the strategy make?
The median fund saw 15 round trips over the 5.74-year window. Counts ranged from none on SGOV to 20 on FAS and XLF. CLSE had 8, on a shorter history than most of the funds, so one trade moves its result more.
Does the strategy have a stop-loss?
No. The only exit is RSI(14) closing above 70. On funds that fall steadily this can mean holding through large losses, which is why UVXY and SOXS show drawdowns of 97.97% and 99.45%.
What happens if I use RSI 25/70 or 35/70 instead of 30/70?
Entry at 25 gave a median CAGR of 1.32% with 11 round trips. Entry at 35 gave 3.77% with 18 round trips and a 26.1% median drawdown. The base 30/70 median is 2.98%, so the looser entry made more at the median and drew down more.
Does RSI mean reversion work on leveraged and inverse ETFs?
On leveraged funds the median CAGR was 12.08% with a 52% median drawdown, and the rule beat holding on 3 of 10. On inverse funds the median was negative 11.04%. Inverse funds decay over time, so oversold readings are not followed by reliable bounces.
Compare with other strategies
Backtests are hypothetical, computed by DeployQuant's engine on minute-resolution consolidated US market data (2021-01-04 to 2026-10-02, $10,000 starting capital, no margin, no fees or slippage in the headline run; buy-and-hold puts 98% of the account in at the first open, as the templates do) and do not guarantee future results. Nothing on this page is investment advice. Live trading involves risk of loss.