DeployQuant3-Month Momentum Switch on QID
ProShares UltraShort QQQ: -2x daily Nasdaq-100. Backtest 2021-01-04 to 2026-10-02, $10,000 starting capital, computed by the same engine that runs live DeployQuant strategies.
QID is a fund that profits when the Nasdaq-100 drops, and the 3-month momentum switch asks whether QID's own momentum can be ridden. The template goes long when the trailing 63-day return is above 5% and sells when that return turns negative. QID is ProShares UltraShort QQQ, with -2x daily exposure to the Nasdaq-100.
Between 2021-01-04 and 2026-10-02, a $10,000 account following the rule ended at $3,389. The total return was -66.11% and the CAGR -17.18%. The maximum drawdown was 71.79% and the Sharpe ratio -0.49. Twenty-two round trips closed and only 3 were profitable, a win rate of 14%. The profit factor was 0.08. The account was invested 21.8% of the time.
Buy-and-hold on QID did worse on every summary line. It lost 86.52% in total, a CAGR of -29.46%, left $1,348 and drew down 87.33%. The rule was ahead of holding in 2021, 2023, 2024, 2025 and 2026. It was behind in 2022 by a wide margin, because 2022 was the year the fund gained.
On QID this template ranks 9th of 12. QID ranks 55th of 59 funds for the template, which has a median CAGR of 0% across all 59. No costs are charged in the headline run, the window is 5.74 years and every decision uses daily data.
Year by year
| Year | 3-month momentum | buy & hold |
|---|---|---|
| 2021 | −5.9% | −45.2% |
| 2022 | −30.3% | 63.3% |
| 2023 | −11.7% | −55.6% |
| 2024 | −7.7% | −31.8% |
| 2025 | −15.5% | −31.6% |
| 2026 | −25.0% | −29.5% |
The six years
Rule returns by year were -5.9% (2021), -30.3% (2022), -11.7% (2023), -7.7% (2024), -15.5% (2025) and -25% (2026 to date). Holding returned -45.2%, 63.3%, -55.6%, -31.8%, -31.6% and -29.5%. The rule lost money in every year, including the one year the fund rose.
The rule bought once in 2021, on 2021-05-13 at an adjusted 104.71, and sold on 2021-05-21 at 98.35 for -6.07%. May 2021 shows -5.91%, and every other month of the year is 0%. QID was falling that year and the 63-day return rarely cleared 5%. The gap to holding was 39.3 points in the rule's favor.
2022 is the year that mattered. QID returned 63.3% for a holder and the rule lost 30.3%, a gap of 93.6 points against the rule, which is the largest gap in the table. The rule made nine round trips and two won. It bought on 2022-01-24 at 84.02 and held to 2022-04-29 at 87.68, a gain of 4.36% over 95 days, the best trade of the window. After that the sequence went wrong. The rule bought on 2022-09-26 and sold three days later for -0.46%, bought on 2022-10-03 and sold two days later for -7.04%, then bought on 2022-10-07 at 106.19. That last position was held through a stretch where QID kept sliding and left on 2022-12-01 at 88.92, 55 days later. The 63-day return had to turn negative before the rule would exit, and by then the fund had fallen 16.26% from the entry price. Trades from 2022-06-09 to 2022-07-28 lost 5.25%, from 2022-10-03 to 2022-10-05 lost 7.04%, and from 2022-10-07 to 2022-12-01 lost 16.26%. January 2022 ended at -10.08%, March at -10.65%, July at -17.75%, October at -10.55%, November at -12.5% and December at -8.23%. April (21.64%) and June (14.43%) were the winning months.
2023 returned -11.7% for the rule against -55.6% for holding. The rule traded in January, September, October and November. The trade from 2023-09-27 to 2023-11-08 lost 8.83% over 42 days, and November ended at -11.21%.
2024 returned -7.7% against -31.8%. All three trades were in September and October and none won: -6.58% from 2024-09-09 to 2024-09-12, -0.09% on 2024-09-19 and -1.37% in early October.
2025 returned -15.5% against -31.6% from one trade. The rule bought on 2025-03-07 at 33.1 and sold on 2025-05-13 at 27.87 for -15.8%. March 2025 made 7.04%, then April lost 9.61% and May 12.62%.
2026 is a partial year ending 2026-10-02, with -25% against -29.5%. Five trades closed and none won. February lost 7.15%, March made 9.24%, April lost 15.46% and September lost 12.54%. The final trade was bought on 2026-09-16 at 14.4 and sold on 2026-09-22 at 13.08 for -9.17%.
Month by month
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2021 | 0.0% | 0.0% | 0.0% | 0.0% | −5.9% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% |
| 2022 | −10.1% | 6.7% | −10.7% | 21.6% | −0.6% | 14.4% | −17.8% | 0.0% | −0.5% | −10.6% | −12.5% | −8.2% |
| 2023 | −4.5% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% | −0.2% | 4.3% | −11.2% | 0.0% |
| 2024 | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% | −6.5% | −1.3% | 0.0% | 0.0% |
| 2025 | 0.0% | 0.0% | 7.0% | −9.6% | −12.6% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% |
| 2026 | 0.0% | −7.2% | 9.2% | −15.5% | 0.0% | 0.0% | 0.0% | 0.0% | −12.5% | 0.0% | – | – |
Month by month
Most months in the table show 0%, because the rule held nothing. The best month was April 2022 at 21.64% and the worst was July 2022 at -17.75%. June 2022 added 14.43%, March 2025 7.04% and March 2026 9.24%, and these are the only large gains.
The losing months include April 2026 at -15.46%, May 2025 at -12.62%, November 2022 at -12.5%, September 2026 at -12.54% and November 2023 at -11.21%. Every active month after 2022 either lost money or made less than 10%. A holder had a best month of 28.51% in April 2022 and a worst of -22.08% in April 2026. The rule was invested for April 2026 as well and lost 15.46% that month, so it shared in that fall.
QID's calendar-month averages show May at -9.75%, November at -8.36%, October at -5.32% and June at -4.94%, with September at 4.33%. The rule's November results in 2022 and 2023 match that pattern. Each month has 5 or 6 samples, so the figures describe this window. The fund's average return on Mondays was -0.34%, the weakest weekday, while Tuesday, Thursday and Friday sit near zero. A daily rule that buys at the open has no way to use these small tilts, and they are too small to build a rule on.
Every trade
3-month momentum on QID made 22 closed round trips, an average hold of 21 days, an average winner of 2.59%, an average loser of −5.94%, a profit factor of 0.08, a longest losing streak of 10. It held a position at the close on 21.8% of trading days.
| Entry | Entry price | Exit | Exit price | Return | Days held |
|---|---|---|---|---|---|
| 2021-05-13 | $104.71 | 2021-05-21 | $98.35 | −6.1% | 8 |
| 2022-01-24 | $84.02 | 2022-04-29 | $87.68 | 4.4% | 95 |
| 2022-05-02 | $92.60 | 2022-06-08 | $91.30 | −1.4% | 37 |
| 2022-06-09 | $92.93 | 2022-07-28 | $88.05 | −5.3% | 49 |
| 2022-09-26 | $106.27 | 2022-09-29 | $105.78 | −0.5% | 3 |
| 2022-10-03 | $110.91 | 2022-10-05 | $103.10 | −7.0% | 2 |
| 2022-10-07 | $106.19 | 2022-12-01 | $88.92 | −16.3% | 55 |
| 2022-12-08 | $96.21 | 2022-12-13 | $86.61 | −10.0% | 5 |
| 2022-12-19 | $101.63 | 2022-12-22 | $103.93 | 2.3% | 3 |
| 2022-12-29 | $109.72 | 2022-12-30 | $108.82 | −0.8% | 1 |
| 2023-01-05 | $108.98 | 2023-01-10 | $104.01 | −4.6% | 5 |
| 2023-09-22 | $58.78 | 2023-09-26 | $59.46 | 1.2% | 4 |
| 2023-09-27 | $59.93 | 2023-11-08 | $54.64 | −8.8% | 42 |
| 2024-09-09 | $38.15 | 2024-09-12 | $35.64 | −6.6% | 3 |
| 2024-09-19 | $33.59 | 2024-09-20 | $33.56 | −0.1% | 1 |
| 2024-10-04 | $32.96 | 2024-10-10 | $32.51 | −1.4% | 6 |
| 2025-03-07 | $33.10 | 2025-05-13 | $27.87 | −15.8% | 67 |
| 2026-02-05 | $20.68 | 2026-02-10 | $19.58 | −5.3% | 5 |
| 2026-02-13 | $20.56 | 2026-02-23 | $20.13 | −2.1% | 10 |
| 2026-03-04 | $20.22 | 2026-04-15 | $18.64 | −7.8% | 42 |
| 2026-09-02 | $14.46 | 2026-09-08 | $13.88 | −4.0% | 6 |
| 2026-09-16 | $14.40 | 2026-09-22 | $13.08 | −9.2% | 6 |
Prices are adjusted for splits and dividends, so they sit below the quotes printed at the time. An open position is marked at the last close.
The trades
The rule closed 22 trades with 44 fills. A typical trade lost 4.56% and lasted 6 days. The average hold was 20.7 days, the longest was 95 days and the shortest was 1 day. The longest losing streak was 10, and the longest winning streak was 1, so no two winners came back to back.
Only three trades won: 4.36% (2022-01-24 to 2022-04-29), 2.26% (2022-12-19 to 2022-12-22) and 1.16% (2023-09-22 to 2023-09-26). The average win was 2.59%. The average loss was -5.94%. The worst losses were -16.26% (2022-10-07 to 2022-12-01, 55 days), -15.8% (2025-03-07 to 2025-05-13, 67 days), -9.98% (2022-12-08 to 2022-12-13, 5 days), -9.17% (2026-09-16 to 2026-09-22) and -8.83% (2023-09-27 to 2023-11-08).
The template's caveat says a three-month lookback is slow, and the worst losses show it. The caveat also describes V-shaped crashes and recoveries, which is the shape of most QID rallies: a sharp move up followed by a drift down. The two largest losses were held for 55 and 67 days. The signal needs a 63-day return below 0% to exit, and a fund that spikes and falls back covers that distance after most of the damage is done. Entry has the same lag. A 63-day return above 5% on an inverse fund is often a rally that is close to its end.
The five best trades in the facts include two small losers (-0.09% and -0.46%), which shows how few winners there were. By exit year, 2021 had one trade and no wins, 2022 had nine and two, 2023 had three and one, 2024 had three and none, 2025 had one and none, and 2026 had five and none. The rule has not closed a winning trade since September 2023. Ten trades have closed since then, all at a loss.
The December 2022 chop and the 2026 entries
The busiest stretch of the test ran from 2022-12-08 to 2023-01-10. The rule bought on 2022-12-08 at 96.21 and sold on 2022-12-13 at 86.61, a loss of 9.98%. It bought again on 2022-12-19 at 101.63 and sold on 2022-12-22 at 103.93 for 2.26%. It bought on 2022-12-29 at 109.72 and sold the next day at 108.82 for -0.82%, then bought on 2023-01-05 at 108.98 and sold on 2023-01-10 at 104.01 for -4.56%. That is four round trips in about five weeks with one winner. Every re-entry came at a higher adjusted price than the exit before it, so the rule paid more to get back in than it received for getting out.
The 63-day return sits close to the entry and exit lines when a fund moves sideways with sharp swings, and QID did that in late 2022. December 2022 ended at -8.23% for the rule and January 2023 at -4.47%. The same shape returned in 2026. The rule bought on 2026-02-05 at 20.68 and lost 5.32%, bought on 2026-02-13 at 20.56 and lost 2.09%, and bought on 2026-03-04 at 20.22 and lost 7.81% over 42 days. In September it bought at 14.46 on 2026-09-02 for -4.01% and at 14.4 on 2026-09-16 for -9.17%.
The entry prices show the fund's decay inside the data. The first entry was at 104.71 in May 2021 and the last was at 14.4 in September 2026, all adjusted for splits and distributions. A position taken at any of those dates was in a fund whose price kept falling. The win rate of 14% and the profit factor of 0.08 describe 22 attempts to catch a rise in a series that mostly fell.
Largest drawdowns
| Peak | Low point | Depth | Days to low | Recovered | Days to recover |
|---|---|---|---|---|---|
| 2022-06-16 | 2026-09-22 | −71.8% | 1559 | not yet | – |
| 2022-03-14 | 2022-03-29 | −27.1% | 15 | 2022-05-09 | 41 |
| 2021-05-13 | 2022-02-02 | −18.4% | 265 | 2022-03-07 | 33 |
Buy-and-hold's deepest drawdown ran from 2021-03-08 to 2026-10-02 and reached −87.3%.
The drawdowns
The rule's deepest drawdown started on 2022-06-16 and reached 71.79% on 2026-09-22, after 1,559 days, without recovering. The second was 27.07% from 2022-03-14 to 2022-03-29, recovered by 2022-05-09. The third was 18.4% from 2021-05-13 to 2022-02-02, recovered on 2022-03-07, 33 days after the low.
The 71.79% figure is a long series of small and medium losses and no single trade. The compounding works against the account after each loss, because a smaller balance needs a larger gain to get back to the old peak, and the rule never found a gain large enough in the years after 2022. The worst single trade lost 16.26%. A drawdown that runs from June 2022 to September 2026 includes the whole stretch where the fund fell and the rule kept re-entering.
For comparison, holding QID drew down 87.33% from 2021-03-08 to the last day, 2,034 days with no recovery. The fund's own series fell 89.58% over a longest drawdown of 1,400 sessions. The rule's smaller number comes from being invested 21.8% of the time. The fund had 78 falls of 10% or more from a 20-day high over 553 days, so a strategy that sits in cash most of the time avoids part of that.
With trading costs
The headline run fills at the bar price. These runs charge slippage on every fill.
| Slippage per fill | CAGR | Max drawdown | Final value | Sharpe |
|---|---|---|---|---|
| None (headline) | −17.2% | −71.8% | $3,389 | -0.49 |
| 5 basis points | −17.5% | −72.3% | $3,309 | -0.51 |
| 10 basis points | −17.8% | −72.8% | $3,244 | -0.52 |
Costs
Slippage of 5 basis points per fill moved the CAGR to -17.52% and the ending balance to $3,309. At 10 basis points it was -17.81% and $3,244. The maximum drawdown went from 71.79% to 72.33% and then 72.8%. The Sharpe ratio moved from -0.49 to -0.508 and -0.519.
Only 44 fills occurred in 5.74 years, so costs have a small effect. QID trades $228,567,798 a day with a median minute volume of 5,631 shares, thinner than the larger inverse Nasdaq funds. The cost runs model slippage and do not model what happens on a gap day such as 2025-04-09, when QID fell 23.53%. The rule was long that day, inside the trade from 2025-03-07 to 2025-05-13, and took the full move at once. The other four of the fund's five worst days were 2022-11-10 at -14.71%, 2022-11-30 at -9.11%, 2022-07-27 at -8.47% and 2025-05-12 at -8.05%. The rule held the fund on all five of those days.
Changing the parameters
| Version | CAGR | Max drawdown | Round trips | Win rate | Final value |
|---|---|---|---|---|---|
| Published rules | −17.2% | −71.8% | 22 | 14% | $3,389 |
| Enter above 0% | −15.8% | −70.8% | 33 | 21% | $3,718 |
| Enter above 10% | −13.7% | −64.2% | 9 | 11% | $4,290 |
| Enter above 15% | −14.8% | −68.2% | 7 | 14% | $3,995 |
Other entry thresholds
All three variants change the entry level, and the exit stays at a negative 3-month return. Entering above 0% returned -15.83% a year with a 70.8% drawdown, 33 trades and 7 wins. Entering above 10% returned -13.71% with a 64.17% drawdown, 9 trades and one win. Entering above 15% returned -14.77% with a 68.18% drawdown, 7 trades and one win. The published 5% entry returned -17.18%.
Higher thresholds trade less and lose less. A 10% entry waits for a bigger move before buying, which filters out some of the weak rallies that produced the small losing trades in the default run. The cost is fewer chances, and the nine trades in that variant produced one winner. The 10% entry has the best CAGR, the best Sharpe at -0.383 and the shallowest drawdown. Between 10% and 15% the order reverses, so there is no clean trend in the table. With 7 to 33 trades and one to 7 winners, none of the variants has enough evidence to rank reliably.
Ending balances run from $3,389 to $4,290. All four are better than the $1,348 for holding and all four lose money. The variants show that the entry level is not the problem. A momentum switch on a decaying inverse fund loses at every entry level tested.
How QID behaved
| Measure | QID |
|---|---|
| Data in this test | 2021-01-04 to 2026-10-02 (1444 sessions) |
| Total return, buy and hold | −89.3% |
| Annualized volatility | 44.7% |
| Deepest drawdown | −89.6% (2021-03-08 to 2026-10-02) |
| Up days | 45.1% |
| Average daily range | 3.15% |
| Average overnight gap | 1.19% |
| Correlation to SPY | -0.94 |
| Correlation to QQQ | -1.00 |
| Correlation to TLT | -0.09 |
| Sessions above the 200-day average | 20.0% |
| Crossings of the 200-day average | 34 |
| Falls of 10% or more from a 20-day high | 78 |
How QID behaved
QID's own price lost 89.27% over the window, a CAGR of -32.21% with annualized volatility of 44.7%. Calendar returns were -46.44% in 2021, 66.4% in 2022, -57.25% in 2023, -34.03% in 2024, -34.98% in 2025 and -34.35% in 2026 so far. Only 45.11% of days were up. The average up day was 2.16% and the average down day -2.02%.
The leverage table describes how a -2x fund behaves over time. QQQ gained 151.07% over the window. Twice that inverted would be -302.14%, and the daily-rebalanced ideal was -93.34%. QID returned -89.27%. The gap comes from compounding, which costs an inverse fund most in rising, choppy years. In 2022, when QQQ fell 32.39%, QID rose 66.4%. In 2023, when QQQ rose 54.81%, QID fell 57.25%.
QID was above its 200-day average on 20% of sessions and crossed it 34 times. Beta to SPY was -2.57, beta to QQQ was -2 and correlation to QQQ was -1. The 14-day RSI fell below 30 on 116 sessions, and the median 20-day forward return was -3.09% against a baseline of -3.55%. RSI(2) fell below 10 on 248 sessions with -3.93%. Oversold readings did not predict a rebound.
Other inverse funds did better under the same rule. TBF returned 1.17% a year, PSQ -4.3% and SH -4.83%. SQQQ returned -25.78% and SOXS -41.12%. QID's -17.18% was 8th of 11 in the category, and the category median was -11.61%.
Overnight and intraday returns, and what oversold readings did
QID's log return split into -92.57% overnight and -127.02% during the session. Those are 42.16% and 57.84% of the total loss. The intraday part is the larger drag. The rule trades at the open, so each position it holds collects the whole session move from the fill onward, and the session was where most of the fund's decline accumulated.
The rule was flat for most of 2024 and 2025. QQQ gained 25.59% in 2024 while QID lost 34.03%, and QQQ gained 20.77% in 2025 while QID lost 34.98%. The rule's results in those years, -7.7% and -15.5%, came from a few days of exposure in a fund that was falling in both years, and the cash it held the rest of the time is the main reason it beat holding.
Oversold readings gave a short bounce and no more. After the 116 sessions with RSI(14) below 30, the median 5-day return was 0.12%, against a baseline of -1.17%. The 20-day median was -3.09%, against -3.55%. The fund closed with RSI(14) above 70 on only 19 sessions, which shows how seldom it ran hot. The lag-1 autocorrelation of daily returns was -0.03, so one day's move said almost nothing about the next.
A 3-month momentum rule asks for persistence over 63 sessions, and an inverse fund with 45.11% up days and a negative median forward return from any starting point rarely supplies it. The few large rallies, such as the best day of 12.12% on 2025-04-04 and 11% on 2022-09-13, were single sessions that the rule could only act on at the next open.
The rules
Hold while the trailing 3-month return is positive (above +5% to enter, below 0% to exit).
- WHEN the market opens · IF not invested AND the 63-day return > +5% · THEN buy with 98% of the sleeve
- WHEN the market opens · IF invested AND the 63-day return < 0% · THEN sell the whole position
Time-series momentum on a quarterly lookback, the horizon much of the academic momentum literature uses. The template enters after a +5% three-month run and exits when the same measure turns negative. The gap between entry (+5%) and exit (0%) is a buffer against flip-flopping around a single threshold.
Good for: assets with long, persistent cycles, such as index, sector and managed-futures ETFs.
Watch out: a three-month lookback is slow; V-shaped crashes and recoveries can see it exit near the bottom and re-enter well off the low.
How the rules fit this fund
The rule buys at the open when the 63-day return is above 5%, with 98% of the sleeve, and sells everything when it is below 0%. The 5-point gap is meant to stop flip-flopping. The template says it suits assets with long persistent cycles, such as index, sector and managed-futures ETFs.
QID's cycles are short. Its average intraday range is 3.15% and its average overnight gap is 1.19%, so a position can move several percent between one decision and the next. The rule checks only once a day, at the open, and has no stop between checks. A rise in an inverse fund comes from a drop in the Nasdaq-100 that lasts weeks, and the three-month window sees it after the move has run. The 22 trades show the pattern: short trades that stopped out for small losses, a few long ones that reversed.
The other templates on QID tell the same story from different angles. The RSI(2) snapback was best at -1.85%, followed by the momentum breakout at -3.08%, the golden cross at -6.03% and the trend plus trailing stop at -7.6%. The 200-day regime filter returned -10.46%, the SMA 10/50 trend -13.07%, the EMA 12/26 trend -13.63% and RSI mean reversion -14.18%. This rule was 9th at -17.18%. Behind it were the weekly 7% target at -26.24%, the dip buyer at -27.11% and the monthly cycle at -28.81%.
Every template lost money on QID. The test stops at one fund, one window and fixed parameters.
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
Did 3-month momentum beat buy-and-hold on QID?
Over 2021-01-04 to 2026-10-02, 3-month momentum on QID returned −17.2% annualized vs −29.5% for buy-and-hold: it beat buy-and-hold by 12.3% per year, with a maximum drawdown 15.5 points shallower than holding (71.8% vs 87.3%).
How many trades did it make?
22 completed round trips over 5.7 years (44 fills), with 14% of round trips closing profitably.
Why 63 days?
63 trading days is about one quarter, a common momentum lookback. You can sweep it in DeployQuant to see how the horizon changes results.
How did the 3-month momentum switch do on QID?
It turned $10,000 into $3,389 from 2021-01-04 to 2026-10-02, a CAGR of -17.18% with a 71.79% maximum drawdown. Buy-and-hold on QID returned -29.46% a year. The rule made 22 round trips and won 3.
Why is the win rate only 14%?
Three of 22 trades closed with a gain. The 63-day return signal is slow on a fund whose rallies last weeks, so entries come late and exits come after the reversal. The last winning trade ended on 2023-09-26.
Why did the rule lose 30.3% in 2022 when QID gained 63.3%?
The rule made nine round trips in 2022 and two won. Several trades entered after a rally and exited after a reversal, including -16.26% from 2022-10-07 to 2022-12-01. A slow lookback misses most of a fast move.
Did a higher entry threshold help?
Somewhat. Entering above 10% returned -13.71% and above 15% returned -14.77%, against -17.18% for the 5% default. Entering above 0% returned -15.83%. All lose money.
How much do trading costs matter?
Little. At 5 basis points the CAGR was -17.52%, and at 10 basis points it was -17.81%. The rule has only 44 fills in 5.74 years.
How does this rule compare with holding QID?
It lost less. The CAGR was -17.18% against -29.46%, and the drawdown 71.79% against 87.33%. Both lost money, and the rule trailed holding only in 2022.
Related
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.