DeployQuantRSI(2) Dip Snapback
Buy extreme 2-day RSI washouts under 10, exit as soon as RSI(2) recovers above 70.
RSI(2) Dip Snapback is the fastest template in the library. It holds a fund only after a sharp multi-day selloff and leaves as soon as the 2-day RSI is back above 70. Over 2021-01-04 to 2026-10-02 it made a median of 4.7% a year across the 59 funds tested, with a median maximum drawdown of 17.5%. It beat buy-and-hold on 34 of the 59 funds and had a shallower drawdown than buy-and-hold on 58.
Those two counts say different things. The strategy sits in cash most of the time, with a median exposure of 39%, so a shallower drawdown follows largely from the low exposure. Beating buy-and-hold on return is the harder test, and 34 of 59 is a little better than a coin flip. The median trade count is 166 round trips in under six years, which works out to a new position every few weeks on a typical fund.
The headline run has no fees or slippage. With this many trades that matters more than it does for any other template here, and the cost runs on each fund page show how much of the return survives 5 and 10 basis points per trade. The result is also compared the result with the other eleven templates, such as the monthly cycle and RSI mean reversion.
The rules
- WHEN the market opens · IF not invested AND RSI(2) < 10 · THEN buy with 98% of the sleeve
- WHEN the market opens · IF invested AND RSI(2) > 70 · THEN sell the whole position
A short-horizon mean-reversion template popularized by Larry Connors' RSI-2 research. A 2-period RSI under 10 flags a sharp multi-day selloff. In assets with a persistent upward drift, those selloffs have tended to snap back within days. Trades are frequent and short. This is the highest-turnover template in the library.
Good for: liquid index ETFs with strong long-term drift; turnover is high so per-trade edges are small.
Watch out: high trade counts make results sensitive to execution quality; a crash that keeps crashing will hand this template several losing entries in a row.
What the two rules do
The entry rule fires at the open when the strategy is flat and RSI(2) is below 10. It buys with 98% of the sleeve. The exit rule fires at the open when the strategy is invested and RSI(2) is above 70, and it sells the whole position. There is no stop, no target and no time limit. A trade ends only when the oscillator recovers.
RSI(2) uses a two-day lookback, so it swings between extremes constantly. A fund that closes down two days running with a large second drop will often print a reading under 10. A single strong up day is enough to push it back above 70. That is why exposure stays near 39% at the median and why holding periods are a few days.
The design leans on one assumption: that a fund with long-run upward drift tends to bounce after a washout. The test in the table below checks that assumption on 59 funds. Where the drift is up, the rules have something to harvest. Where the drift is down, as with inverse and volatility funds, the same rules buy into a persistent decline.
Because the rule set is only two lines, almost every difference between pages comes from the fund. Volatility, gap behaviour and trend direction decide whether a washout is followed by a bounce or by a second leg down. The trend and trailing stop template has the opposite bias, and its median return of 2.6% is lower than this one.
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 |
|---|---|---|---|---|---|---|
| TQQQ | 39.3% | 25.4% | −43.2% | 0.96 | 166 | 68% |
| SOXL | 39.2% | 33.3% | −58.5% | 0.81 | 167 | 58% |
| TECL | 30.2% | 38.2% | −51.0% | 0.77 | 162 | 65% |
| QLD | 26.2% | 23.6% | −29.5% | 0.93 | 164 | 68% |
| SSO | 24.6% | 21.9% | −25.5% | 1.11 | 172 | 70% (+1 open) |
| SPUU | 24.1% | 22.5% | −31.2% | 1.09 | 171 | 68% (+1 open) |
| FAS | 18.7% | 18.3% | −43.4% | 0.65 | 159 | 62% (+1 open) |
| ROM | 18.5% | 30.2% | −39.2% | 0.66 | 158 | 61% |
| SOXX | 17.0% | 31.1% | −23.8% | 0.77 | 162 | 61% |
| CLSE | 16.6% | 19.5% | −7.5% | 1.95 | 142 | 75% |
| VOOG | 15.9% | 15.7% | −16.6% | 1.11 | 161 | 70% |
| QQQM | 15.6% | 16.7% | −14.3% | 1.06 | 164 | 70% |
| QQQ | 14.2% | 16.7% | −14.3% | 0.98 | 164 | 69% |
| VOO | 14.1% | 14.4% | −12.6% | 1.28 | 172 | 72% (+1 open) |
| IOO | 14.0% | 16.7% | −11.7% | 1.24 | 169 | 71% (+1 open) |
| SPY | 13.2% | 14.6% | −12.7% | 1.21 | 172 | 71% (+1 open) |
| VV | 12.6% | 14.3% | −13.1% | 1.12 | 173 | 69% (+1 open) |
| XLK | 12.1% | 22.1% | −18.1% | 0.78 | 158 | 65% |
| IWM | 11.4% | 7.5% | −22.8% | 0.83 | 172 | 67% (+1 open) |
| XLF | 9.7% | 12.6% | −14.6% | 0.82 | 159 | 64% (+1 open) |
| TBF | 9.7% | 12.2% | −7.7% | 1.02 | 170 | 59% |
| QQQE | 9.5% | 9.8% | −13.2% | 0.72 | 156 | 62% |
| VOOV | 8.6% | 12.6% | −13.4% | 0.96 | 160 | 68% (+1 open) |
| RINF | 8.5% | 6.6% | −10.6% | 1.10 | 166 | 69% |
| XLP | 8.2% | 5.8% | −11.1% | 0.96 | 166 | 64% (+1 open) |
| XLY | 8.2% | 6.3% | −16.6% | 0.58 | 154 | 64% (+1 open) |
| VTV | 7.3% | 13.4% | −11.7% | 0.85 | 153 | 70% (+1 open) |
| VOX | 6.1% | 9.1% | −30.1% | 0.49 | 157 | 67% (+1 open) |
| CTA | 5.5% | 8.9% | −17.5% | 0.54 | 126 | 62% |
| IAU | 4.7% | 13.7% | −20.1% | 0.49 | 169 | 65% |
| KMLM | 4.0% | 7.1% | −14.2% | 0.49 | 151 | 62% |
| EEM | 3.9% | 6.6% | −15.9% | 0.36 | 162 | 60% (+1 open) |
| QAI | 3.4% | 3.9% | −6.8% | 0.78 | 152 | 65% (+1 open) |
| ALTY | 3.3% | 7.6% | −16.8% | 0.50 | 163 | 52% (+1 open) |
| USDU | 3.1% | 5.2% | −3.3% | 0.79 | 172 | 56% |
| IEF | 1.8% | −2.3% | −12.9% | 0.39 | 172 | 62% (+1 open) |
| AGG | 1.6% | −0.8% | −10.6% | 0.42 | 174 | 59% |
| IEI | 1.4% | −0.4% | −6.9% | 0.48 | 172 | 66% |
| BND | 1.4% | −0.8% | −9.1% | 0.39 | 172 | 60% |
| IGIB | 1.3% | 0.2% | −11.8% | 0.34 | 172 | 61% |
| TLT | 1.0% | −8.2% | −22.8% | 0.15 | 172 | 59% (+1 open) |
| UDN | 0.4% | −1.1% | −17.3% | 0.11 | 171 | 56% (+1 open) |
| PSQ | 0.1% | −13.8% | −33.8% | 0.08 | 172 | 57% (+1 open) |
| VIXM | 0.1% | −16.1% | −52.4% | 0.09 | 157 | 59% |
| SGOV | 0.0% | 3.2% | −0.1% | 0.42 | 52 | 46% |
| FXE | 0.0% | −0.9% | −12.8% | 0.02 | 170 | 56% (+1 open) |
| SH | −0.5% | −10.0% | −24.7% | -0.01 | 171 | 54% |
| SPDN | −1.4% | −9.7% | −26.3% | -0.11 | 169 | 53% |
| QID | −1.9% | −29.5% | −59.0% | 0.06 | 172 | 56% (+1 open) |
| UST | −2.7% | −8.5% | −31.1% | -0.23 | 173 | 58% (+1 open) |
| SDS | −3.6% | −21.6% | −47.9% | -0.10 | 169 | 53% |
| TMF | −4.0% | −31.2% | −55.1% | 0.01 | 170 | 56% (+1 open) |
| EEV | −5.9% | −16.1% | −59.8% | -0.12 | 162 | 53% |
| VXZ | −7.0% | −15.1% | −56.8% | -0.35 | 157 | 57% |
| SQQQ | −9.7% | −42.3% | −79.5% | -0.05 | 170 | 54% (+1 open) |
| REW | −10.9% | −36.1% | −72.4% | -0.21 | 166 | 55% (+1 open) |
| TECS | −26.6% | −46.7% | −90.6% | -0.46 | 165 | 51% (+1 open) |
| SOXS | −29.9% | −48.3% | −94.2% | -0.12 | 170 | 56% (+1 open) |
| UVXY | −30.3% | −48.7% | −89.2% | -0.29 | 162 | 51% |
Which funds worked and which did not
TQQQ leads the table at 39.3% a year against 25.4% for buy-and-hold, with a 43.2% maximum drawdown, 166 round trips and a 68% win rate. SOXL is also ahead of holding at 39.2% against 33.3%, but its drawdown is 58.5% and its win rate is only 58%. Both are leveraged funds, and in both the rules made money by stepping aside during long slides and re-entering after sharp ones.
The leveraged funds are not uniformly good. TECL returned 30.2% against 38.2% for buy-and-hold, and ROM returned 18.5% against 30.2%. In both cases the underlying sector ran strongly for most of the window and the strategy spent about 36% of the time invested, so it missed more of the trend than it gained from the dips. A mean-reversion rule pays for being out of the market in a strong uptrend.
The broad index funds show the same effect on a smaller scale. QQQ made 14.2% against 16.7% for holding. SPY made 13.2% against 14.6%. VOO made 14.1% against 14.4%. VOOG, a growth index fund, made 15.9% against 15.7%, the only one of these to finish ahead. Their drawdowns are between 12.6% and 16.6%. The trade-off on these funds is a small loss of return for a smaller drawdown, with win rates between 69% and 72%.
The weakest results are on inverse and volatility funds. UVXY lost 30.3% a year, SOXS lost 29.9%, and TECS lost 26.6%. Their drawdowns are 89.2%, 94.2% and 90.6%. These funds decay over time, so a washout in them is often a step in a long decline rather than a pause. Buy-and-hold lost more on all three, but that is a low bar. The QID page and the SOXL page show how those funds behave on their own.
The widest gaps in favour of the strategy are on the funds that fell the most. QID lost 29.5% a year when held and 1.9% a year under these rules. TMF lost 31.2% held and 4.0% under the rules. The strategy beat holding by a wide margin there, yet still finished with a negative return. Beating a fund that falls steadily and making money are different outcomes.
The sample is one window of 5.74 years that includes the 2022 bear market. The ranking of funds could change in another period, and a few high-ranking entries come from leveraged funds that had unusually large swings inside this window.
Where the test stops
The test is one window from 2021-01-04 to 2026-10-02 on daily decisions, with orders filled on minute bars and no margin. Prices are adjusted for splits and dividends. The headline has no fees or slippage, so the returns above are an upper bound for each fund under these rules.
The leveraged and inverse results depend on a short list of dates. The leveraged funds have the widest swings in the window, and a different start date could change the ranking. The broad index results are steadier but small. The 2022 sample is the only extended bear market in the window, so the evidence on how the rules behave in prolonged declines comes from one episode.
None of this is a forecast. The numbers describe what these rules did on these funds in this period. For comparison with other templates, the dip buyer page shows a different dip-buying rule with a median of 1.3%, and the 200-day regime filter shows a trend rule with a median of 1.8%.
Results by fund type
| Fund type | ETFs | Median CAGR | Median buy & hold | Median max DD | Beat holding |
|---|---|---|---|---|---|
| Broad index ETFs | 12 | 13.2% | 14.4% | −13.4% | 2 of 12 |
| Sector ETFs | 6 | 9.7% | 12.6% | −18.1% | 2 of 6 |
| Leveraged ETFs | 10 | 24.6% | 23.6% | −43.2% | 8 of 10 |
| Inverse ETFs | 11 | −3.6% | −21.6% | −59.0% | 10 of 11 |
| Bond ETFs | 7 | 1.4% | −0.8% | −10.6% | 6 of 7 |
| Commodity ETFs | 1 | 4.7% | 13.7% | −20.1% | 0 of 1 |
| Currency ETFs | 3 | 0.4% | −0.9% | −12.8% | 2 of 3 |
| Volatility products | 3 | −7.0% | −16.1% | −56.8% | 3 of 3 |
| Alternative-strategy ETFs | 6 | 5.5% | 7.6% | −14.2% | 1 of 6 |
Results by fund type
The category medians separate the funds into three groups.
Leveraged funds had a median of 24.6% a year against 23.6% for buy-and-hold, with a median drawdown of 43.2%. The strategy beat holding on 8 of 10. This is the group where the rules were most useful in return terms, because leveraged funds fall fast and rebound fast, which is the pattern RSI(2) looks for.
Broad index funds had a median of 13.2% against 14.4% for holding, a median drawdown of 13.4%, and a win over buy-and-hold on only 2 of 12. Sector funds had a median of 9.7% against 12.6%, with 2 of 6 ahead. Alternative-strategy funds had a median of 5.5% against 7.6%, with 1 of 6 ahead. In these groups the rules gave up return in exchange for lower exposure and shallower drawdowns.
Inverse funds had a median of negative 3.6% a year against negative 21.6% for holding. The strategy beat holding on 10 of 11, but the median drawdown was still 59.0%. Volatility products show the same shape: a median of negative 7.0% against negative 16.1%, a median drawdown of 56.8%, and a win over holding on all 3.
Bond funds had a median of 1.4% against negative 0.8% for holding, ahead on 6 of 7. The bond result reflects the period more than the rules. Long and intermediate bond funds fell during the rate rises, the strategy was out of them for more than half of the time, and its exposure of around 43% to 46% on those funds cost less than holding. Currency funds had a median of 0.4% against negative 0.9%, ahead on 2 of 3. The one commodity fund, IAU, made 4.7% against 13.7% for holding.
The pattern runs along the drift of the fund. Where holding paid well, the strategy trailed. Where holding lost money, the strategy lost less. It did not turn a falling asset into a winner. Related rules from the weekly profit target template had a higher median of 6.6%, so a faster exit with a target worked better across the whole universe.
Year by year, median across all ETFs
| Year | RSI(2) snapback | Buy & hold | ETFs with a gain |
|---|---|---|---|
| 2021 | 10.0% | 4.1% | 47 of 59 |
| 2022 | −2.5% | −12.7% | 24 of 59 |
| 2023 | 5.9% | 8.9% | 42 of 59 |
| 2024 | 2.3% | 9.7% | 37 of 59 |
| 2025 | 9.8% | 11.1% | 45 of 59 |
| 2026 | 2.0% | 3.7% | 36 of 59 |
Year by year across the universe
The yearly medians show how the result depends on the market, not on a steady edge.
In 2021 the median strategy return was 10.0% against 4.1% for buy-and-hold, and 47 of 59 funds finished positive. Dips in that year were followed by quick recoveries often enough for the rule to run well ahead of holding.
In 2022 the median was negative 2.5% against negative 12.7% for holding. Only 24 funds finished positive. The strategy lost far less than holding at the median, but a majority of funds still ended the year down. Buying washouts in a falling market produces losing entries in a row, and the data shows it.
In 2023 the median was 5.9% against 8.9% for holding, with 42 funds positive. In 2024 the median was 2.3% against 9.7%, with 37 positive. This was the widest gap in the wrong direction. In a steady uptrend with few deep dips, the strategy waited and earned little.
In 2025 the median was 9.8% against 11.1%, with 45 funds positive. The 2026 figure of 2.0% against 3.7% covers a partial year ending 2026-10-02, and 36 funds were positive.
Two things stand out. The strategy beat holding at the median in only 2 of the 6 calendar years, 2021 and 2022. In the four other years it trailed. The 34 of 59 count over the full window therefore leans on the funds that fell or went sideways.
The second observation is the spread in funds positive: 24 in 2022, 47 in 2021 and 45 in 2025. A strategy with a stable edge would show a narrower range. The result moves with the market regime. The monthly cycle template, with a median of 5.5%, has a different seasonal logic and fits this window differently.
Trade counts, exposure and cost sensitivity
The trade count is the main risk to the headline numbers. Funds with median results show 166 round trips, and most funds sit between 153 and 174. SGOV is the outlier at 52 round trips and 7.1% exposure, because a Treasury bill fund barely moves and RSI(2) rarely reaches either threshold. The next lowest counts are CTA at 126 and CLSE at 142, two funds with shorter histories than the rest.
Exposure explains the shallow drawdowns. Funds such as the large-cap index funds are invested about 36% of the time. Inverse funds are invested about 47% of the time, higher than the index funds, because those funds spend long periods with sustained down moves in RSI(2), and the strategy keeps buying into them.
The headline run is gross of costs. A strategy that makes 166 round trips pays the spread and any commission twice per round trip, once on entry and once on exit. A fund earning 1.4% a year, as the median bond fund did, has no room for that. A fund earning 39.3% a year, like TQQQ, has more room. The cost runs on each fund page use 5 and 10 basis points, and the fund pages are the right place to check whether a given fund keeps its return.
Changing the parameters
| Version | Median CAGR | Median max DD | Median round trips |
|---|---|---|---|
| Published rules | 4.7% | −17.5% | 166 |
| RSI(2) < 5 / > 70 | 5.8% | −16.8% | 159 |
| RSI(2) < 15 / > 70 | 4.3% | −18.2% | 173 |
| RSI(2) < 10 / > 60 | 5.1% | −17.5% | 174 |
| RSI(2) < 10 / > 80 | 5.6% | −17.8% | 155 |
What the parameter variants show
Each fund page runs four variants alongside the base rule of buying under 10 and selling over 70. The table here gives the median across funds.
Buying under 5 gave a median of 5.8% a year with a 16.8% median drawdown and 159 round trips. This is the best variant for return and the shallowest for drawdown. A stricter entry waits for a deeper washout, trades less, and avoids some of the weaker entries.
Buying under 15 gave a median of 4.3% with an 18.2% median drawdown and 173 round trips. A looser entry trades more and earns less. The base rule sits between them at 4.7%, a 17.5% median drawdown and 166 round trips.
The exit changes matter less. Selling over 60 gave 5.1% with 174 round trips. Selling over 80 gave 5.6% with 155 round trips and a median drawdown of 17.8%. Both exit variants finished above the base median, so the exit level does not point in one direction. Moving the exit changed the median by less than a point.
Three of the four variants finished above the base median and one, the looser entry, finished below it. All four stay in a band between 4.3% and 5.8%, so the base rule sits in the lower part of a narrow range and no setting changes the picture by a large amount. No setting turns this template into a strong performer on the median fund.
These are medians across 59 funds, not tuned results for any one of them. Choosing the best variant per fund would fit noise, since the window is 5.74 years and each fund has a few hundred trades at most. The per-fund tables on pages such as QQQ show the same four variants for a single fund.
Frequently asked questions
What is the RSI(2) snapback strategy?
Buy extreme 2-day RSI washouts under 10, exit as soon as RSI(2) recovers above 70. A short-horizon mean-reversion template popularized by Larry Connors' RSI-2 research. A 2-period RSI under 10 flags a sharp multi-day selloff. In assets with a persistent upward drift, those selloffs have tended to snap back within days. Trades are frequent and short. This is the highest-turnover template in the library.
Does RSI(2) snapback beat buy-and-hold?
Across 59 ETFs backtested 2021-01-04 to 2026-10-02, it beat same-ETF buy-and-hold on 34 of 59 (58%). Median CAGR was 4.7% with a median max drawdown of 17.5%. Per-ETF results vary widely; the table lists every one.
How often does RSI(2) trade?
Far more than RSI(14), with dozens of round trips per year on a volatile ETF. The backtest table on each page shows the exact count over the test window.
Is RSI(2) too fast for daily bars?
It is designed for daily bars. The 2-day window catches short, sharp washouts rather than long regimes.
How many trades does RSI(2) Dip Snapback make?
The median fund had 166 round trips between 2021-01-04 and 2026-10-02. Most funds sit between 153 and 174. SGOV is the exception with 52, because a Treasury bill fund rarely reaches the RSI thresholds.
Did RSI(2) Dip Snapback beat buy-and-hold?
It beat buy-and-hold of the same fund on 34 of 59 funds, with a median of 4.7% a year. It had a shallower drawdown on 58 of 59. The funds that beat holding were mostly leveraged, inverse, bond and volatility funds.
Which funds did RSI(2) Dip Snapback do best on?
TQQQ made 39.3% a year and SOXL made 39.2%. QLD made 26.2%, SSO 24.6% and SPUU 24.1%. These are leveraged index and sector funds, and their drawdowns ranged from 25.5% to 58.5%.
Does RSI(2) work on inverse and volatility ETFs?
Not in this test. SQQQ lost 9.7% a year, SOXS lost 29.9% and UVXY lost 30.3%. The strategy lost less than holding on most of them, but these funds decay and the drawdowns reached between 79.5% and 94.2%.
Which RSI(2) settings worked best?
Buying under 5 and selling over 70 gave the best median at 5.8% a year with a 16.8% median drawdown. Buying under 15 gave 4.3%. All four variants stayed in a narrow band, so the base rule of under 10 and over 70 sits in a narrow range rather than on a sharp peak.
How did RSI(2) Dip Snapback do in 2022?
The median return was negative 2.5% against negative 12.7% for buy-and-hold, and 24 of 59 funds finished positive. The strategy lost less than holding, but it kept buying dips in a falling market.
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.