DeployQuantUVXY trading strategies, backtested
ProShares Ultra VIX Short-Term Futures ETF: 1.5x short-term VIX futures: spikes in crashes, decays steadily otherwise. Every DeployQuant template run on UVXY over 5.7 years of minute data, same engine, same window, sorted by return.
UVXY is the ProShares Ultra VIX Short-Term Futures ETF. It aims for 1.5 times the daily move of a short-term VIX futures index, so it jumps when markets sell off and bleeds when they do not. From 2021-01-04 to 2026-10-02 the fund's own price history shows a total return of -99.94%, a CAGR of -72.83% and annualized volatility of 103.95%. Its maximum drawdown of 99.95% started on 2021-01-29 and reached its trough on 2026-09-22, with no recovery. Twelve templates were run on it with the same engine, window and $10,000 starting balance.
No template finished with a positive CAGR. Golden cross lost the least at -10.09% a year, and it did so with one round trip and 3.7% exposure. The monthly cycle lost the most at -68.72% and ended with $12.65. Buy-and-hold in the engine returned -48.74% a year, ended at $215.70 and had a maximum drawdown of 98.42%. Ten of the 12 templates beat that CAGR and 11 had a shallower drawdown. Since every template lost money, those counts mostly reflect how little time the rule spent in the fund.
The engine's holding line and the fund's own calendar years drift apart. The engine's holding returned -88%, -37.7%, -65.8%, -13.7%, -10% and -3.1% in 2021 to 2026. The fund's own years were -89.67%, -44.65%, -87.75%, -50.84%, -65.37% and -52.77%. The two are close in 2021 and 2022 and differ by a wide margin from 2023 on, and the facts do not say why. The comparison tables on this page use the engine's holding figure because each template runs in the same engine. The behaviour sections use the fund's own series.
| Strategy | CAGR | max DD | Sharpe | trades | win rate | final value |
|---|---|---|---|---|---|---|
| Golden Cross (SMA 50/200) | −10.1% | −51.3% | -0.52 | 1 | 0% | $5,431 |
| SMA-200 Trend + 15% Trailing Stop | −25.4% | −81.4% | -0.76 | 24 | 21% | $1,861 |
| 200-Day SMA Regime Filter | −29.8% | −86.9% | -0.91 | 12 | 0% | $1,314 |
| RSI(2) Dip Snapback | −30.3% | −89.2% | -0.29 | 162 | 51% | $1,261 |
| 20-Day Momentum + Trailing Stop | −36.2% | −93.3% | -0.60 | 125 | 28% | $760 |
| SMA 10/50 Trend | −38.0% | −94.0% | -0.44 | 16 | 6% | $640 |
| Weekly Entry + 7% Target | −38.4% | −97.5% | -0.51 | 280 | 46% (+1 open) | $621 |
| EMA 12/26 Trend | −39.0% | −96.0% | -0.41 | 19 | 16% | $583 |
| RSI(14) Mean Reversion | −48.5% | −98.0% | -0.55 | 12 | 50% (+1 open) | $220 |
| Drawdown Dip Buyer + 8% Target | −48.7% | −97.8% | -0.86 | 0 | – (+1 open) | $217 |
| 3-Month Momentum Switch | −49.8% | −98.1% | -1.05 | 27 | 11% | $192 |
| First-to-Last Day of Month | −68.7% | −99.9% | -0.69 | 69 | 33% (+1 open) | $13 |
| Buy & hold UVXY | −48.7% | −98.4% | -0.80 | – | – | $216 |
The ranking by loss
Golden cross made one round trip, entered on 2025-04-16 and exited on 2025-07-07 after 82 days with a loss of 46.62%. That one trade gave a CAGR of -10.09% and a drawdown of 51.34%. The rule held no position in any other year, so its yearly returns are 0% in 2021 to 2024 and 2026 and -45.7% in 2025. The 200-day average needs 200 sessions to form and UVXY closed above it on only 4.18% of sessions, so the rule had one chance in the window. Against holding it came out ahead in five of the six years. It lost the most in 2025, when it held a position and the engine's holding line returned -10%. In 2025 the rule trailed holding by that one trade.
Trend plus trailing stop was second at -25.39%, with 24 round trips and 5 wins, a 21% win rate. The average win was 13.14% and the average loss 11.31%, and the profit factor was 0.19. It was invested 3.7% of the time with an average hold of 3.2 days. Its best trade, 30.62%, lasted one day, from 2025-04-08 to 2025-04-09. Its worst, -26.52%, ran from 2024-11-01 to 2024-11-06. Fourteen of its 24 round trips ended in 2025. The drawdown was 81.39%, which started on 2022-03-07 and had not recovered by the end. The 15% trail kept each loss modest, and the repeated small losses still added up to most of the account.
The 200-day regime filter made 12 round trips and lost all 12, with an average loss of 14.74% and a CAGR of -29.78%. Its best trade lost 4.97%. Its worst two were -42.26% over 4 days from 2024-08-05 and -41.19% over 39 days from 2025-04-04. Both start on a day when the fund had just jumped. The first entry came on 2024-08-05, the day of UVXY's largest gain in the window at 62.04%, and the second on 2025-04-04, the day after a 37.03% gain. The filter buys strength by construction, since the price has to be above a long average, and a volatility fund's strength tends to reverse within days.
The RSI(2) snapback made 162 round trips with 82 wins, a 51% win rate, and a CAGR of -30.28%. The average win was 7.63% and the average loss 9.09%, and the profit factor was 0.76. It made 16.9% in 2022 and lost in the other five years. Its median trade returned 0.23%, so the typical trade was flat and the damage came from a few large losers: -40.9% from 2024-08-08 to 2024-08-22, -37.76% from 2023-11-01 to 2023-12-08 and -28.87% from 2021-10-04 to 2021-10-28. Its best trades were 32.05% on 2025-04-03 to 2025-04-04 and 30.13% on 2024-09-03 to 2024-09-04. Each of those lasted a day.
Momentum breakout made 125 round trips and won 35, a 28% win rate. The average win was 11.34% and the average loss 6.7%, with a profit factor of 0.47. It averaged 2.1 days in a trade and was invested 12.7% of the time. The CAGR was -36.17% and the drawdown 93.34%. Its best trade was 45.16% on 2024-08-01 to 2024-08-02, a single day, and its worst was -13.51%. The trailing stop kept single losses small. The rule lost because wins were rarer than losses and not large enough to compensate.
The two moving-average rules were close to each other. SMA 10/50 made 16 round trips and won 1, with an average win of 3.56% and an average loss of 16.74%. EMA 12/26 made 19 and won 3, with a longest losing streak of 12. Their CAGRs were -38.04% and -39.04%, their profit factors both 0.01, and their drawdowns 93.96% and 96.01%. They were invested about a fifth of the time, 19.7% and 19.6%, and the long holds were the ones that lost: SMA 10/50 lost 28.16% over 31 days from 2024-10-11 and EMA 12/26 lost 35.81% over 15 days from 2022-01-25.
The weekly 7% target made 280 round trips with a 46% win rate, an average win of 7.52% and an average loss of 7.45%. The profit factor was 0.84 and the CAGR -38.37%, with a drawdown of 97.45%. It was invested 48.5% of the time. The target fills often because the fund's average intraday range is 7.03%. The result was mixed by year, with -24.6%, -3.7%, -73.8%, -68% and -40% in 2021 to 2025 and 59.5% in 2026 so far. Its worst trade was -41.4% from 2024-08-05 to 2024-08-08, again on the day of the largest gain.
RSI(14) mean reversion and the dip buyer ended almost where holding did, at -48.55% and -48.7% against -48.74%. The dip buyer made one buy on 2021-02-02 at an adjusted price of $31,075 and never sold, so it is holding. RSI(14) mean reversion made 12 round trips with 6 wins, an average win of 18.93% and an average loss of 38.75%. It held for 102.1 days on average. Its worst trade, -83.44%, lasted 334 days from 2022-10-28 to 2023-09-27. It was holding a position opened on 2025-11-06 at the end, down 67.88%. It made money in 2022 and 2024, with 29.6% and 28.9%.
The 3-month momentum switch finished at -49.75%, slightly worse than holding. It made 27 round trips and won 3, a profit factor of 0.05, with a longest losing streak of 12. Its worst trade, -60.58%, began on 2024-08-05 and ended on 2024-08-14. The monthly cycle made 69 round trips and won 23, a 33% win rate, with an average win of 14.98% and an average loss of 18.87%. It was invested 95.2% of the time, so its -68.72% CAGR reflects the fund's decay with a rules wrapper on top.
How each strategy traded UVXY
| Strategy | Time in market | Avg hold (days) | Best trade | Worst trade | Profit factor | With 10 bps slippage |
|---|---|---|---|---|---|---|
| golden cross | 3.7% | 82 | −46.6% | −46.6% | – | −10.1% |
| trend + trailing stop | 3.7% | 3 | 30.6% | −26.5% | 0.19 | −26.0% |
| 200-day regime filter | 3.6% | 6 | −5.0% | −42.3% | – | −30.1% |
| RSI(2) snapback | 56.0% | 7 | 32.0% | −40.9% | 0.76 | −34.1% |
| momentum breakout | 12.7% | 2 | 45.2% | −13.5% | 0.47 | −38.9% |
| SMA 10/50 trend | 19.7% | 25 | 3.6% | −28.2% | 0.01 | −38.4% |
| weekly 7% target | 48.5% | 3 | 15.8% | −41.4% | 0.84 | −41.5% |
| EMA 12/26 trend | 19.6% | 21 | 9.3% | −35.8% | 0.01 | −39.5% |
| RSI mean reversion | 73.9% | 102 | 40.2% | −83.4% | 0.28 | −48.9% |
| dip buyer | 98.6% | – | – | – | – | −49.1% |
| 3-month momentum | 14.8% | 11 | 8.8% | −60.6% | 0.05 | −50.3% |
| monthly cycle | 95.2% | 28 | 41.5% | −37.3% | 0.37 | −69.5% |
Costs and fills
Exposure ran from 3.6% for the 200-day filter and 3.7% for golden cross and trend plus trailing stop to 95.2% for the monthly cycle and 98.6% for the dip buyer. Average holds were short for the active rules, 2.1 days for momentum breakout and 2.6 days for the weekly target, and long for RSI(14) mean reversion at 102.1 days and for golden cross at 82 days. Profit factors were below 1 for every template with losses. The highest was 0.84 for the weekly target and 0.76 for the RSI(2) snapback, and the lowest were 0.01 for SMA 10/50 and EMA 12/26 and 0.05 for the 3-month switch.
At 5 and 10 basis points per trade, the CAGR of most templates fell by a small amount. The RSI(2) snapback went from -30.28% to -32.2% and -34.07%. Momentum breakout went from -36.17% to -37.53% and -38.87%. The weekly target went from -38.37% to -36.75% at 5 basis points and -41.52% at 10. The 5 basis point result is better than the zero-cost result, so cost did not rank the three runs in order for this template. The low-trade rules barely moved: golden cross went from -10.09% to -10.11% and -10.13%.
UVXY traded an average of $458,284,172 a day, with a median minute volume of 1,145 shares. The fund is liquid at the daily level and a $10,000 order is small against it. The test fills on minute bars, with no fees in the headline run and no margin. On days with a 62.04% gain or a 31.94% drop, a live fill could differ from the bar price the backtest uses.
Months, streaks and drawdown length
Monthly extremes on UVXY land in a few clusters. April 2025 was the best month for four templates: the weekly target at 52.91%, momentum breakout at 39.76%, the RSI(2) snapback at 29.38% and trend plus trailing stop at 11.82%. The worst months were in August 2024 for the 200-day filter at -41.42%, the weekly target at -47.85% and the 3-month switch at -59.36%. Holding in the engine had its best month in April 2022 at 30.46% and its worst in March 2021 at -39.82%. RSI(14) mean reversion made 55.05% in January 2022 and lost 39.86% in March 2021, the same month as holding's worst.
The moving-average rules share a pattern. SMA 10/50 made 31.3% in March 2026 and EMA 12/26 made 31.28% in the same month, and their worst months, -32.74% in November 2024 and -32.77% in December 2024, sit one month apart.
Losing streaks describe the 200-day filter, which lost 12 trades in a row with no wins. EMA 12/26 and the 3-month switch each lost 12 in a row and SMA 10/50 lost 10. The weekly target had a winning streak of 7 and a losing streak of 8, and trend plus trailing stop lost 8 in a row after a best streak of 3. The RSI(2) snapback won as many as 5 in a row and lost as many as 6. The snapback made 162 trades, and its win rate was 51%.
Drawdowns on this fund rarely end. The monthly cycle fell 99.9% from 2021-01-29 and was still under water on 2026-10-02, 2,072 days later. The dip buyer's single position stayed down for 2,059 days. Trend plus trailing stop's 81.39% drawdown began on 2022-03-07 and reached its trough on 2026-03-31, after 1,485 days. Momentum breakout fell 93.34% from 2021-05-12 and bottomed on 2025-03-24, and the weekly target fell 97.45% from 2021-10-11 to 2025-02-19. Golden cross is the contrast: its 51.34% drawdown ran 73 days, from 2025-04-21 to 2025-07-03, because it had traded once.
By weekday, UVXY averaged -1.04% on Mondays and -0.8% on Wednesdays, with a positive average of 0.45% on Thursdays. Tuesday averaged -0.18% and Friday -0.04%. Over six years the weekday averages carry little weight beside a 7.03% average daily range.
RSI(2) below 10 appeared on 301 sessions. The baseline median 20-day return was -11.36% and the median after an RSI(2) signal was -10.02%, so the signal improved on the baseline by a small margin and left the forward return deeply negative.
How UVXY behaved
| Measure | UVXY |
|---|---|
| Data in this test | 2021-01-04 to 2026-10-02 (1444 sessions) |
| Total return, buy and hold | −99.9% |
| Annualized volatility | 104.0% |
| Deepest drawdown | −100.0% (2021-01-29 to 2026-09-22) |
| Up days | 40.3% |
| Average daily range | 7.03% |
| Average overnight gap | 2.53% |
| Correlation to SPY | -0.76 |
| Correlation to QQQ | -0.70 |
| Correlation to TLT | 0.00 |
| Sessions above the 200-day average | 4.2% |
| Crossings of the 200-day average | 24 |
| Falls of 10% or more from a 20-day high | 81 |
Calendar years
| Year | Return |
|---|---|
| 2021 | −89.7% |
| 2022 | −44.6% |
| 2023 | −87.8% |
| 2024 | −50.8% |
| 2025 | −65.4% |
| 2026 (part) | −52.8% |
Biggest single days
| Best day | Move |
|---|---|
| 2024-08-05 | 62.0% |
| 2021-11-26 | 39.0% |
| 2025-04-03 | 37.0% |
| 2024-08-02 | 36.2% |
| 2021-01-27 | 31.2% |
| Worst day | Move |
|---|---|
| 2025-04-09 | −31.9% |
| 2024-08-06 | −29.5% |
| 2024-12-20 | −22.8% |
| 2021-11-29 | −21.3% |
| 2024-08-09 | −19.9% |
Average return by calendar month
| Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
|---|---|---|---|---|---|---|---|---|---|---|---|
| −2.6% | −4.0% | −4.9% | −0.3% | −21.6% | −16.0% | −9.6% | −15.3% | 7.7% | −6.5% | −16.7% | −16.2% |
Most and least correlated funds
| Most correlated | Least correlated | ||
|---|---|---|---|
| VIXM | 0.93 | SPY | -0.76 |
| VXZ | 0.92 | SSO | -0.76 |
| SPDN | 0.76 | VOO | -0.76 |
| SH | 0.76 | VV | -0.76 |
| SDS | 0.76 | SPUU | -0.76 |
How UVXY behaved
UVXY rose on 40.26% of sessions. The average up day was 5.15% and the average down day was -4.02%. The average intraday range was 7.03% and the average overnight gap 2.53%. The overnight log return was -335.95% and the intraday log return was -399.82%, so the intraday session took slightly more of the decay than the overnight period, 54.34% against 45.66%. The lag-1 autocorrelation of daily returns was -0.05, close to zero, and the fund closed above its 200-day average on 4.18% of sessions and crossed it 24 times.
Beta to SPY was -4.84 and to QQQ -3.26, and the correlation to SPY was -0.76 and to QQQ -0.7. The beta to TLT was 0.01, so the long-bond fund had no linear link to it. The most correlated funds were VIXM at 0.93 and VXZ at 0.92, both longer-dated VIX products, followed by SPDN, SH and SDS at 0.76. The least correlated were SPY, SSO, VOO, VV and SPUU at -0.76.
By calendar month, September was the only strongly positive month at 7.72% on average, with April about flat at -0.28%. May averaged -21.61%, November -16.67%, December -16.22%, June -15.98% and August -15.29%. Each month has five or six observations. August included the 62.04% day of 2024-08-05 and still averaged a loss, because the spike was followed by sharp reversals. These are descriptions of this window.
RSI(14) was below 30 on 116 sessions and above 70 on 7. After an RSI(14) reading below 30 the median forward 5-day return was -1.79% against a baseline of -3.53%, and the 20-day figure was -5.82% against -11.36%. That is better than the baseline at both horizons, and still negative. After RSI(2) fell below 10, on 301 sessions, the median 5-day return was -2.29% and the 20-day return -10.02%. The RSI(14) mean reversion template used the first signal and held on average 102.1 days, long enough for the fund to fall 83.44% on one trade. Falls of 10% or more from a 20-day high occurred in 81 episodes covering 1,130 days, so the fund spent most of the window in or near such a fall.
Among the other volatility products, VIXM held at -16.06% a year and its best template, the RSI(2) snapback, returned 0.13%. VXZ held at -15.1% and its best template, golden cross, returned -1.15%. UVXY's best template, golden cross at -10.09%, also lost money. The mid-term products lose more slowly than UVXY, and the templates can reach break-even on them, which they did not do here. The test is one window with daily decisions, minute-bar fills, no margin and no fees in the headline run. Prices in the trade lists are adjusted for splits, so entry prices such as $31,075 are adjusted figures.
Years and spikes
Almost every template's yearly returns are negative, and the exceptions mark the spikes. In 2022 the RSI(2) snapback made 16.9% and RSI(14) mean reversion made 29.6%, while the engine's holding line lost 37.7%. In 2024 RSI(14) mean reversion made 28.9% against -13.7% for holding. In 2026 so far the weekly target made 59.5%, and EMA 12/26 made 3.7%. Everything else lost money in every year it held a position.
The fund's biggest days explain why. UVXY gained 62.04% on 2024-08-05, 38.99% on 2021-11-26, 37.03% on 2025-04-03, 36.17% on 2024-08-02 and 31.21% on 2021-01-27. The worst days were -31.94% on 2025-04-09, -29.48% on 2024-08-06, -22.81% on 2024-12-20, -21.32% on 2021-11-29 and -19.87% on 2024-08-09. Three of the five best days and three of the five worst fall in early August 2024 and in April 2025. A rule that happens to hold across a gain keeps it only if it also exits before the reversal, and almost none did.
Trend plus trailing stop returned 0% in 2021 and 2023 and then -37.1%, -46.8%, -39.3% and -8.4% in 2022, 2024, 2025 and 2026. The 200-day filter was flat in 2021 and 2023 and lost 24%, 58.5%, 52.6% and 11.9% in the other four. These rules lost more in the years they traded than in the years they sat out. The engine's holding line lost 88% in 2021 alone, so the flat years were the years in which sitting out paid the most.
Golden cross, with one trade in 2025, had its single bad year and was flat otherwise. The monthly cycle lost 85.6%, 29.2%, 83.8%, 56.9%, 71.4% and 44.6% in the six years. A rule that holds nearly every session cannot sit out the decay, so its yearly returns follow the fund's own years closely.
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.
Frequently asked questions
What was the best strategy for UVXY?
Of the 12 templates tested on UVXY over 2021-01-04 to 2026-10-02, the strongest by CAGR was golden cross at −10.1% (max drawdown 51.3%), versus −48.7% for buy-and-hold. The best result in hindsight is not a forecast. Check drawdowns and trade counts before drawing conclusions.
Did any strategy beat buying and holding UVXY?
10 of 12 templates beat UVXY buy-and-hold (−48.7% CAGR) on this window; 11 of 12 had a shallower maximum drawdown than holding (98.4%).
Which strategy worked best on UVXY?
Golden cross, at -10.09% a year against -48.74% for holding, with a maximum drawdown of 51.34%. It made one round trip, in 2025, and was invested 3.7% of the time. It still lost money.
Did any strategy make money on UVXY?
No. All 12 templates had a negative CAGR. The range ran from -10.09% for golden cross to -68.72% for the monthly cycle. Several templates made money in single years, for example RSI(14) mean reversion with 29.6% in 2022 and 28.9% in 2024.
Why did the 200-day filter lose on every trade?
It made 12 round trips and none won, with an average loss of 14.74%. A rule that buys when the price is above its 200-day average buys after a spike, and UVXY's spikes reversed within days. UVXY closed above the average on only 4.18% of sessions.
Why does the dip buyer match buy-and-hold on UVXY?
It bought once, on 2021-02-02, and never sold, so it held the fund for the whole window. Its CAGR was -48.7% against -48.74% for holding. Its 8% target never fired after that first buy.
Does mean reversion work on UVXY?
Not reliably. RSI(14) mean reversion returned -48.55% and the RSI(2) snapback returned -30.28%, and the RSI(14) rule had a trade that lost 83.44% over 334 days. The RSI(2) rule won 51% of its 162 trades and still lost money because the average loss was larger than the average win.
How volatile is UVXY?
Annualized volatility was 103.95%, the average intraday range was 7.03%, and the best day was a 62.04% gain on 2024-08-05. The worst day was -31.94% on 2025-04-09. The fund's own total return over the window was -99.94%.
Does holding UVXY in the engine match the fund's own return?
Not fully. Holding in the engine returned -97.84% over the window and the fund's own price series returned -99.94%. The two are close in 2021 and 2022 and differ from 2023 on, and the facts do not explain the gap.
Other volatility products
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.