DeployQuantVIXM trading strategies, backtested
ProShares VIX Mid-Term Futures ETF: mid-term VIX futures, a slower-decaying volatility position. Every DeployQuant template run on VIXM over 5.7 years of minute data, same engine, same window, sorted by return.
VIXM holds mid-term VIX futures, which makes it a bet on volatility rising that pays when markets fall hard and bleeds when they do not. Over 2021-01-04 to 2026-10-02 it bled. Buy-and-hold returned negative 16.06% a year, turned $10,000 into $3,660, and fell 68.59% from 2021-02-22 to 2026-09-22 without recovering. In the profile series the fund lost 65.95% in total.
Eleven of the 12 templates beat holding on CAGR and all 12 had a shallower drawdown. Only one made money: RSI(2) snapback at 0.13%, which is flat. The only template that finished below holding was the dip buyer at negative 16.68%, because it bought once in March 2021 and never sold.
For a fund like this the useful question is how much of the loss a template avoided. The answer in this window was most of it for the templates that stayed out, and almost none for the ones that stayed in. The headline runs have no fees or slippage, and VIXM trades thinly in the data, with a median of 362 shares a minute. The results describe one window with two sharp volatility spikes, August 2024 and April 2025.
| Strategy | CAGR | max DD | Sharpe | trades | win rate | final value |
|---|---|---|---|---|---|---|
| RSI(2) Dip Snapback | 0.1% | −52.4% | 0.09 | 157 | 59% | $10,075 |
| Golden Cross (SMA 50/200) | −1.8% | −24.8% | -0.04 | 3 | 33% | $9,002 |
| RSI(14) Mean Reversion | −7.4% | −51.6% | -0.29 | 14 | 64% (+1 open) | $6,438 |
| SMA-200 Trend + 15% Trailing Stop | −10.2% | −47.3% | -0.43 | 10 | 10% | $5,397 |
| Weekly Entry + 7% Target | −10.9% | −58.5% | -0.38 | 173 | 21% (+1 open) | $5,143 |
| 200-Day SMA Regime Filter | −12.2% | −53.6% | -0.60 | 30 | 17% | $4,749 |
| EMA 12/26 Trend | −12.9% | −58.5% | -0.49 | 31 | 10% | $4,510 |
| SMA 10/50 Trend | −13.1% | −57.0% | -0.52 | 24 | 21% | $4,476 |
| 3-Month Momentum Switch | −13.8% | −58.2% | -0.71 | 20 | 15% | $4,268 |
| First-to-Last Day of Month | −14.8% | −66.0% | -0.43 | 69 | 43% (+1 open) | $3,976 |
| 20-Day Momentum + Trailing Stop | −15.9% | −65.0% | -1.03 | 17 | 0% | $3,704 |
| Drawdown Dip Buyer + 8% Target | −16.7% | −65.3% | -0.51 | 0 | – (+1 open) | $3,507 |
| Buy & hold VIXM | −16.1% | −68.6% | -0.49 | – | – | $3,660 |
Which templates lost least on VIXM
The ranking by CAGR was RSI(2) snapback at 0.13%, the golden cross at negative 1.81%, RSI(14) mean reversion at negative 7.38%, the trailing-stop template at negative 10.18%, the weekly 7% target at negative 10.94%, the 200-day filter at negative 12.16%, EMA 12/26 at negative 12.95%, SMA 10/50 at negative 13.07%, 3-month momentum at negative 13.78%, the monthly cycle at negative 14.84%, the momentum breakout at negative 15.88% and the dip buyer at negative 16.68%.
RSI(2) snapback is the notable one. It made 157 round trips and won 59% of them. Its average win was 2.35% and its average loss was 3.22%, and its profit factor was 1.00, break-even. The yearly results were 16.8% in 2021, 45.9% in 2022, negative 39.2% in 2023, negative 5.2% in 2024, 13.4% in 2025 and negative 9.4% in 2026. The 2022 result came while holding lost 0.8%, so the template beat holding by 46.7 points that year. In 2023 it lost 39.2% while holding lost 43.7%. Its drawdown was 52.36%, from 2022-10-14 to 2024-08-27, and it had not recovered at the end. The result is a flat CAGR with a deep drawdown in the middle. Its worst trade lost 19.53% over 41 days from 2023-05-26 to 2023-07-06, and its best made 10.39% in one day, from 2025-04-03 to 2025-04-04.
The golden cross made 3 round trips. It lost 10.43% over 271 days from 2022-03-09 to 2022-12-05, lost 0.99% in 2026, and made 1.25% over 230 days from 2025-03-04 to 2025-10-20. It was invested 26.7% of the time. Its result of negative 1.81% is a function of being out of the fund in 2021, 2023 and 2024, when holding lost 18.8%, 43.7% and 13.0%. It had a 0% return in all three of those years. In 2022 and 2025 it trailed holding.
RSI(14) mean reversion won 64% of 14 round trips and lost money. Its average win was 6.07% and its average loss was 13.44%. One trade lost 38.75% over 338 days from 2022-10-31 to 2023-10-04, which is the 2023 loss of 37.1% in the yearly table. It also holds a position entered 2026-04-17 at 15.53 that was down 16.91% at the end. A high hit rate with this size of loss is the classic profile for buying a volatility fund after it falls, when it tends to keep falling.
The templates in the middle of the table were within a few points of each other and of holding. The weekly 7% target made 173 round trips and won 21%, with a longest losing streak of 16, an average win of 7.46% and an average loss of 2.34%. It made 10.7% in 2024 and 4.8% in 2025 and lost 38.9% in 2023. The trailing-stop template won 1 of 10 trades, a gain of 11.26% from 2024-12-19 to 2025-04-09, and lost 28.1% in 2024.
The trend templates were all negative and close together: the 200-day filter at negative 12.16% with 30 round trips and 17% wins, EMA 12/26 at negative 12.95% with 31 round trips and 10% wins, and SMA 10/50 at negative 13.07% with 24 round trips and 21% wins. EMA 12/26 had a longest losing streak of 15 and a profit factor of 0.14. A volatility fund spikes and decays, and a moving-average crossover enters after the spike has begun and exits after it has faded. Each of those templates held only about a third of the time or less, and lost money on most trades.
The momentum breakout is the extreme case. It made 17 round trips and none won, with a longest losing streak of 17. Its best trade lost 0.39%. It was invested 13.0% of the time and returned negative 15.88%, which is slightly worse than holding. A rule that enters on strength lost every time on this fund. Its worst trade lost 10.06% in one day, from 2024-08-06 to 2024-08-07, entered after the 21.09% jump on 2024-08-05.
Exposure, Sharpe and the worst month of each template
Time in the fund separated the templates more cleanly than their win rates did. The monthly cycle was invested 95.2% of the time and the weekly 7% target 75.3%. RSI(14) mean reversion was in for 65.2%, RSI(2) snapback for 48.6%, and the trailing-stop template for 42.3%. The golden cross, the 200-day filter and 3-month momentum were invested 26.7%, 26.0% and 24.0% of the time, and the momentum breakout only 13.0%. The two templates with the highest exposure also had the deepest drawdowns among the active rules, at 66.03% for the monthly cycle and 58.50% for the weekly target, against 24.81% for the golden cross.
Sharpe ratios tell the same story. Holding scored negative 0.49. RSI(2) snapback scored 0.09 and the golden cross negative 0.04, the only two readings near zero. RSI(14) mean reversion scored negative 0.29, the weekly target negative 0.38, the monthly cycle negative 0.43 and the trailing-stop template negative 0.43. The trend templates and 3-month momentum scored between negative 0.49 and negative 0.71, and the momentum breakout scored negative 1.03, the worst of the set. Several templates beat holding on CAGR while posting a Sharpe about the same as holding, which means they lost less in total without improving the return per unit of volatility.
The worst month for holding was June 2023 at negative 17.60%. The monthly cycle, RSI(2) snapback and RSI(14) mean reversion all had June 2023 as their worst month too, at negative 16.14%, negative 17.77% and negative 17.68%. That is three different rules taking the same hit, because all three were holding the fund that month. The trend and momentum templates had other worst months: August 2024 for the trailing-stop template at negative 14.78%, the 200-day filter at negative 16.38%, 3-month momentum at negative 21.02% and the momentum breakout at negative 21.50%. EMA 12/26 and SMA 10/50 had May 2025 at negative 11.01% and negative 11.13%, the month after the April 2025 spike that they had been holding through.
Best months were clustered too. April 2025 was the best month for eight of the 12 templates, with returns from 12.71% to 18.96%. The weekly 7% target made 18.96% that month, more than holding at 13.11%, with its best trade running from 2025-03-31 to 2025-04-04. RSI(2) snapback and the other mean-reversion rule had April 2022 as their best month, at 14.61% and 11.06%.
How each strategy traded VIXM
| Strategy | Time in market | Avg hold (days) | Best trade | Worst trade | Profit factor | With 10 bps slippage |
|---|---|---|---|---|---|---|
| RSI(2) snapback | 48.6% | 7 | 10.4% | −19.5% | 1.00 | −5.1% |
| golden cross | 26.7% | 186 | 1.3% | −10.4% | 0.10 | −1.9% |
| RSI mean reversion | 65.2% | 86 | 12.0% | −38.8% | 0.65 | −7.8% |
| trend + trailing stop | 42.3% | 89 | 11.3% | −12.2% | 0.13 | −10.5% |
| weekly 7% target | 75.3% | 8 | 13.5% | −9.6% | 0.79 | −16.9% |
| 200-day regime filter | 26.0% | 18 | 7.6% | −16.7% | 0.12 | −13.1% |
| EMA 12/26 trend | 33.5% | 23 | 5.8% | −9.0% | 0.14 | −13.9% |
| SMA 10/50 trend | 33.0% | 28 | 5.5% | −11.3% | 0.12 | −13.8% |
| 3-month momentum | 24.0% | 25 | 8.8% | −20.9% | 0.17 | −14.4% |
| monthly cycle | 95.2% | 28 | 14.5% | −16.5% | 0.60 | −16.9% |
| momentum breakout | 13.0% | 16 | −0.4% | −10.1% | – | −16.4% |
| dip buyer | 96.5% | – | – | – | – | −16.7% |
Costs, hold times and the single trades that decided results
At 10 basis points RSI(2) snapback fell from 0.13% to negative 5.10%. At 5 basis points it was negative 2.51%. With 314 fills and a profit factor of 1.00, any cost makes it a loser. The weekly 7% target, with 347 fills, fell from negative 10.94% to negative 14.15% and negative 16.89%, and its drawdown rose to 68.21%. At 10 basis points that is below holding at negative 16.06%. The monthly cycle fell from negative 14.84% to negative 16.85%, also below holding. The other templates moved by a few tenths of a point.
At 10 basis points the golden cross at negative 1.92% was the best result. RSI(14) mean reversion at negative 7.83% was second, the trailing-stop template at negative 10.51% was third, and the 200-day filter at negative 13.06% was fourth. Nothing finished positive.
Hold times on VIXM were short for the active templates. The weekly target averaged 7.9 days and RSI(2) snapback 6.6. The momentum breakout averaged 15.9 days and the 200-day filter 18.0. The longer holds were RSI(14) mean reversion at 85.9 days, the trailing-stop template at 88.7 and the golden cross at 186. The templates with longer holds had the larger single losses: negative 38.75% for RSI(14) and negative 10.43% for the golden cross.
The August 2024 spike is the cleanest example of how the templates responded. The best day for VIXM was 2024-08-05 at 21.09%, and the worst was 2024-08-06 at negative 13.32%. A template that held through both days kept most of the spike. A template that entered after the first day took the reversal. The 200-day filter's worst trade, a loss of 16.72% over 7 days from 2024-08-05 to 2024-08-12, was entered on the day of the spike. 3-month momentum lost 20.85% over 46 days from 2024-08-05 to 2024-09-20, the same entry date. The momentum breakout lost 10.06% from 2024-08-06 to 2024-08-07. Three templates bought the spike and lost on it. In the year-by-year table 2024 was negative for most: 3-month momentum lost 29.6%, the 200-day filter 33.8%, and the trailing-stop template 28.1%, against 13.0% for holding.
April 2025 produced the opposite result. The best month for holding was April 2025 at 13.11%, and the best days were 2025-04-10 at 12.18% and 2025-04-03 at 11.01%. RSI(2) snapback made 10.39% in one day there. The trend templates showed best months of 13.49% to 13.50% in the same month, since they were in the fund for it.
The dip buyer needs its own paragraph. It made a single entry on 2021-03-18 at an adjusted 38.27 and the 8% target was never reached. The position was down 66.28% at the end of the window. Its CAGR of negative 16.68% finished slightly below holding. A template with one trade has no win rate and no profit factor, and it is a test of one entry date.
The sample is small for the slow templates. The median minute volume was 362 shares and the average daily dollar volume was $3,186,899, and the cost runs use a flat charge per fill that does not model that thin volume.
How VIXM behaved
| Measure | VIXM |
|---|---|
| Data in this test | 2021-01-04 to 2026-10-02 (1444 sessions) |
| Total return, buy and hold | −66.0% |
| Annualized volatility | 29.8% |
| Deepest drawdown | −69.9% (2021-02-22 to 2026-09-22) |
| Up days | 45.3% |
| Average daily range | 2.06% |
| Average overnight gap | 0.77% |
| Correlation to SPY | -0.73 |
| Correlation to QQQ | -0.66 |
| Correlation to TLT | -0.01 |
| Sessions above the 200-day average | 30.2% |
| Crossings of the 200-day average | 60 |
| Falls of 10% or more from a 20-day high | 30 |
Calendar years
| Year | Return |
|---|---|
| 2021 | −19.2% |
| 2022 | −0.8% |
| 2023 | −44.9% |
| 2024 | −13.7% |
| 2025 | 5.5% |
| 2026 (part) | −15.4% |
Biggest single days
| Best day | Move |
|---|---|
| 2024-08-05 | 21.1% |
| 2025-04-10 | 12.2% |
| 2025-04-03 | 11.0% |
| 2024-08-02 | 9.2% |
| 2022-05-05 | 8.3% |
| Worst day | Move |
|---|---|
| 2024-08-06 | −13.3% |
| 2025-04-09 | −9.8% |
| 2025-05-12 | −7.5% |
| 2024-12-20 | −6.5% |
| 2024-08-09 | −5.6% |
Average return by calendar month
| Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
|---|---|---|---|---|---|---|---|---|---|---|---|
| −3.9% | 1.2% | 0.3% | 1.9% | −5.3% | −4.2% | −1.3% | −1.8% | 1.8% | −0.4% | −5.4% | −2.3% |
Most and least correlated funds
| Most correlated | Least correlated | ||
|---|---|---|---|
| VXZ | 0.96 | VOO | -0.73 |
| UVXY | 0.93 | SSO | -0.73 |
| SH | 0.73 | SPY | -0.73 |
| SPDN | 0.73 | VV | -0.72 |
| SDS | 0.72 | SPUU | -0.72 |
How VIXM behaved in this window
VIXM's annualized volatility was 29.83% with an average intraday range of 2.06% and an average overnight gap of 0.77%. It rose on 45.25% of days. The average up day was 1.35% and the average down day was negative 1.26%. It dropped on more days than it rose and rose by more on the up days, which is the pattern of a fund that decays and spikes.
Its beta to SPY was negative 1.32 and to QQQ negative 0.88, with correlations of negative 0.73 and negative 0.66. Correlation to TLT was negative 0.01. The funds most correlated with VIXM were VXZ at 0.96 and UVXY at 0.93, followed by SH and SPDN at 0.73 and SDS at 0.72. The least correlated were VOO, SSO and SPY at negative 0.73, and VV and SPUU at negative 0.72. A fund that moves inversely to equities most days is not a clean hedge: the correlation is negative 0.73, not negative 1.00, and the long-run return was negative 65.95%. The related volatility pages are UVXY and VXZ. Buy-and-hold on UVXY returned negative 48.74% a year in the same window and on VXZ negative 15.10%, so the mid-term funds decayed far more slowly than the short-term one. The golden cross was the best template on both of those pages, at negative 10.09% and negative 1.15%.
The return split is lopsided. The overnight log return was negative 85.58% and the intraday log return was negative 18.80%. The overnight share of the loss was 81.99% and the intraday share 18.01%. Most of the decline was incurred between the close and the next open.
The calendar years were negative 19.18% in 2021, negative 0.82% in 2022, negative 44.87% in 2023, negative 13.67% in 2024, 5.50% in 2025 and negative 15.40% in 2026 through 2026-10-02. The fund lost money in five of the six calendar years and gained only in 2025. 2022, a year in which US equity indexes fell hard, returned only negative 0.82%. RSI(2) snapback made 45.9% that year, so the swings inside a flat year were enough for a short-term rule to trade.
The drawdown from 2021-02-22 to 2026-09-22 was 69.88% in the profile table and 68.59% in the buy-and-hold run, with 2,038 days to the trough. The longest drawdown lasted 1,410 sessions and did not end. The only other drawdowns in the buy-and-hold list were 6.57% and 4.31% in January and February 2021. There were 30 falls of 10% or more from a 20-day high, over 203 days, which means the fund spent a large number of days in rapid drops.
The biggest days were 2024-08-05 at 21.09%, 2025-04-10 at 12.18%, 2025-04-03 at 11.01%, 2024-08-02 at 9.17% and 2022-05-05 at 8.29%. The worst were 2024-08-06 at negative 13.32%, 2025-04-09 at negative 9.75%, 2025-05-12 at negative 7.47%, 2024-12-20 at negative 6.47% and 2024-08-09 at negative 5.63%. The biggest up days sat next to large down days: 2024-08-05 and 2024-08-06, then 2025-04-09 and 2025-04-10. A daily rule has to be on the right side of each.
Monthly averages were weakest in November at negative 5.36%, May at negative 5.26% and June at negative 4.21%, and strongest in April at 1.87%, September at 1.79% and February at 1.24%. The data has 5 or 6 observations per month.
The fund was above its 200-day average in 30.20% of sessions and crossed it 60 times. RSI(14) fell below 30 in 122 sessions and above 70 in 13. After the oversold sessions the median forward 5-day return was negative 0.76% against a baseline of negative 0.39%, and the 20-day return was negative 1.32% against negative 1.11%. An oversold reading on VIXM was followed by results no better than any other day, and slightly worse. RSI(2) below 10 appeared in 233 sessions, with a 5-day median of negative 0.40% and a 20-day median of negative 1.34%, also near the baselines. The lag-1 autocorrelation was negative 0.07.
The test stops at one window, daily decisions, no margin, and flat per-fill costs. A window that began before a volatility spike or ended just after one would rank the templates differently.
Day of week, months of the year and the overnight gap
VIXM's day-of-week averages were small. Monday averaged negative 0.24%, Tuesday negative 0.09%, Wednesday negative 0.15%, Thursday 0.13% and Friday 0.04%. The Monday figure is the weakest. The data does not say why. With 1,444 sessions the gaps between days are small next to the 1.35% average up day, and they should not be read as a schedule.
The overnight figures pull in different directions. The average overnight gap was 0.77% while the overnight log return summed to negative 85.58%, so the gap statistic and the compounded total are measured differently and should not be compared directly. The compounded total shows that the fund lost far more between the close and the next open than inside the session, where the intraday log return was negative 18.80%. A rule that buys at the open and sells at the open, as these templates do, is exposed to the overnight leg whenever it is invested.
By calendar month the fund was positive on average in February, March, April and September, at 1.24%, 0.26%, 1.87% and 1.79%. The other eight months averaged negative, with January at negative 3.85%, December at negative 2.25%, and August at negative 1.84%. Each average rests on five or six observations, and the August figure contains the 2024 spike, so these numbers describe this window and not a seasonal pattern. The weakest stretch, May, June and November, includes June 2023, the worst single month for holding.
A final point on the signal statistics. The 233 sessions with RSI(2) below 10 outnumber the 122 with RSI(14) below 30, which helps explain why RSI(2) snapback traded 157 round trips and RSI(14) mean reversion only 14. The forward medians after both signals were negative, so the signal gave no entry advantage. RSI(2) snapback's flat result came from its exits: it sold quickly on a bounce, average hold 6.6 days, and avoided holding the fund through the slow decay.
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 VIXM?
Of the 12 templates tested on VIXM over 2021-01-04 to 2026-10-02, the strongest by CAGR was RSI(2) snapback at 0.1% (max drawdown 52.4%), versus −16.1% 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 VIXM?
11 of 12 templates beat VIXM buy-and-hold (−16.1% CAGR) on this window; 12 of 12 had a shallower maximum drawdown than holding (68.6%).
What was the best strategy for VIXM?
[RSI(2) snapback](/learn/strategies/rsi2-dip-snapback/vixm/) had the highest CAGR at 0.13%, against negative 16.06% for holding. Its drawdown was 52.36% and its profit factor was 1.00. At 10 basis points of slippage it fell to negative 5.10%.
Did any strategy beat buying and holding VIXM?
Eleven of the 12 templates beat holding on CAGR, and all 12 had a shallower drawdown than the 68.59% for holding. Only the dip buyer finished below holding, at negative 16.68%.
Does VIXM hedge a stock market sell-off?
It moved against the market on many days. Correlation to SPY was negative 0.73 and beta was negative 1.32. It still lost 65.95% over the window, and it returned negative 0.82% in 2022, a year when US stocks fell.
Why did VIXM lose so much over five years?
The data shows a decline in most years, with returns of negative 19.18% in 2021 and negative 44.87% in 2023. The overnight log return was negative 85.58%, which is most of the loss. The data does not say why.
How did the trend strategies do on VIXM?
They lost money. The [200-day filter](/learn/strategies/sma200-regime/vixm/) returned negative 12.16%, EMA 12/26 negative 12.95% and SMA 10/50 negative 13.07%. They won between 10% and 21% of their trades.
What happened on 2024-08-05?
VIXM rose 21.09%, its best day in the window, and fell 13.32% the next day. The 200-day filter and 3-month momentum entered on 2024-08-05 and lost 16.72% and 20.85% on those trades.
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.