DeployQuant3-Month Momentum Switch on QQQ
Invesco QQQ Trust: tracks the Nasdaq-100, a tech-heavy growth index. Backtest 2021-01-04 to 2026-10-02, $10,000 starting capital, computed by the same engine that runs live DeployQuant strategies.
The 3-month momentum switch buys QQQ when its 63-day return is above 5% and sells when that return falls below 0%. Both tests run at the market open. The gap between the two thresholds is meant to stop the rule flipping around a single line. From 2021-01-04 to 2026-10-02 it turned $10,000 into $15,727, a CAGR of 8.21%. Buy-and-hold of QQQ made 16.7% and ended at $24,265.
The rule made 12 closed round trips and held a 13th at the end of the data. Six of the 12 won, a win rate of 50%, and the winners averaged 12.54% against 3.8% for the losers, a profit factor of 3.46. The maximum drawdown was 21.53%, against 34.23% for holding. The Sharpe ratio was 0.65 against 0.83.
It beat holding in one year out of six. That was 2022, when it lost 18.1% and the fund lost 31.7%. In every other year it trailed, and the widest gap was 2024, at 23.3 points. On QQQ the rule ranked eighth of the 12 templates, and QQQ was fourteenth of 59 funds for this rule. Across all 59 funds the median CAGR of 3-month momentum was 0%, and the median across broad index funds was 8%, so QQQ at 8.21% is a typical result for its group.
The test is one window of 5.7 years with no fees in the headline run, so it describes this stretch of QQQ and nothing further.
Year by year
| Year | 3-month momentum | buy & hold |
|---|---|---|
| 2021 | 7.3% | 28.4% |
| 2022 | −18.1% | −31.7% |
| 2023 | 35.9% | 53.0% |
| 2024 | 1.8% | 25.1% |
| 2025 | 14.3% | 20.4% |
| 2026 | 13.3% | 22.1% |
How each year compared with holding
Holding QQQ gained 28.4% in 2021, lost 31.7% in 2022, then gained 53%, 25.1%, 20.4% and 22.1% in 2023, 2024, 2025 and the part of 2026 in the data. The rule returned 7.3%, minus 18.1%, 35.9%, 1.8%, 14.3% and 13.3%.
2021 was a gap of 21.1 points. The rule has no position in January, February or March, because a 63-day return needs a quarter of history and then needs to clear 5%. The first entry was on 2021-04-07. That trade lost 3.17% by 2021-05-11. The rule bought again on 2021-05-27 and held to 2021-10-05 for a gain of 6.32%. May lost 5.18% and September lost 5.56%, while June, July and August gained 6.09%, 2.75% and 4.09%.
2022 is the year the rule worked, and it worked partly. It lost 18.1% against 31.7%, 13.6 points ahead. The loss is spread across three months while the rule was invested: January at minus 6.12%, August at minus 5.59% and September at minus 7.55%. The rule held a position from 2021-10-29 to 2022-01-19, was out from February to July, then entered twice in late summer. The first of those was 2022-08-10 to 2022-08-30, a loss of 5.75% in 20 days. The second, 2022-09-13 to 2022-09-23, lost 7.88% in 10 days and is the worst trade of the test. Both entries came after the 63-day return rose above 5%, and both were reversed within 20 days.
2023 returned 35.9% against 53%. The rule bought on 2023-01-12 at 272.78 and held to 2023-09-22 at 353.09, a gain of 29.44% over 253 days, the best trade. Months of 9.16% in March, 7.51% in May and 6.24% in June built the year. The rule sold on 2023-09-22 after a minus 4.7% September, took no position in October and bought on 2023-11-15.
2024 returned 1.8% against 25.1%, the largest shortfall. The rule held 2023-11-15 to 2024-04-22 for 8.22%, then bought on 2024-05-20, sold on 2024-08-07 at a 0.87% loss, bought on 2024-08-19, sold on 2024-09-06 at a 2.93% loss and bought on 2024-10-23. April lost 5.9%, August 4.53%, September 3.13% and October 1.93%. The fund rose 25.1% over the same year and the rule was in and out of it three times.
2025 returned 14.3% against 20.4%. The rule held to 2025-02-28, sat out March, April and May, and bought on 2025-06-03 at 520.93. 2026 so far returned 13.3% against 22.1%. The rule held 2025-06-03 to 2026-01-30 for 19.99%, then 2026-04-22 to 2026-08-21 for 10.01%. May 2026 was the best month at 10.18% and July the worst of the year at minus 6.28%.
Month by month
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2021 | 0.0% | 0.0% | 0.0% | 2.2% | −5.2% | 6.1% | 2.8% | 4.1% | −5.6% | 0.2% | 1.9% | 1.1% |
| 2022 | −6.1% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% | −5.6% | −7.5% | 0.0% | 0.0% | 0.0% |
| 2023 | 5.5% | −0.4% | 9.2% | 0.5% | 7.5% | 6.2% | 3.8% | −1.5% | −4.7% | 0.0% | 0.4% | 5.5% |
| 2024 | 1.8% | 5.2% | 1.2% | −5.9% | −0.0% | 5.9% | −1.6% | −4.5% | −3.1% | −1.9% | 5.1% | 0.4% |
| 2025 | 2.1% | −4.3% | 0.0% | 0.0% | 0.0% | 5.1% | 2.3% | 0.9% | 5.2% | 4.7% | −1.6% | −0.6% |
| 2026 | 2.0% | 0.0% | 0.0% | 2.6% | 10.2% | −0.2% | −6.3% | 3.7% | 0.0% | 1.3% | – | – |
The months in cash
The month table has long stretches of 0.0%, and each shows a different cause. January to March 2021 is the warm-up and the 5% entry bar. February to July 2022 is six months in cash after the January 2022 sell, during which QQQ was falling. October to December 2022 is another three in cash after the second late-summer loss. October 2023 is a single month between the September sell and the November buy. March to May 2025 is three months out after the February sell, and February and March 2026 are two more.
The rule was out of the fund for most of the long 2022 decline and for all of the spring 2025 low. That is the benefit. The cost is that 0.0% months in the market's best weeks leave the rule behind. Holding's best month was April 2026 at 15.44% and the rule made 2.62% in that month because it entered on 2026-04-22. Holding's worst month was April 2022 at minus 13.14% and the rule made 0.0%.
The rule's own worst month was September 2022 at minus 7.55%, followed by July 2026 at minus 6.28%, January 2022 at minus 6.12% and April 2024 at minus 5.9%. It has no month worse than 8%, against 13.14% for holding. The best months were May 2026 at 10.18%, March 2023 at 9.16% and May 2023 at 7.51%.
Every trade
3-month momentum on QQQ made 12 closed round trips and one position still open at the end of the test, an average hold of 106 days, an average winner of 12.54%, an average loser of −3.80%, a profit factor of 3.46, a longest losing streak of 3. It held a position at the close on 60.9% of trading days.
| Entry | Entry price | Exit | Exit price | Return | Days held |
|---|---|---|---|---|---|
| 2021-04-07 | $319.16 | 2021-05-11 | $309.05 | −3.2% | 34 |
| 2021-05-27 | $322.74 | 2021-10-05 | $343.15 | 6.3% | 131 |
| 2021-10-29 | $369.48 | 2022-01-19 | $361.32 | −2.2% | 82 |
| 2022-08-10 | $316.05 | 2022-08-30 | $297.88 | −5.8% | 20 |
| 2022-09-13 | $293.95 | 2022-09-23 | $270.80 | −7.9% | 10 |
| 2023-01-12 | $272.78 | 2023-09-22 | $353.09 | 29.4% | 253 |
| 2023-11-15 | $380.61 | 2024-04-22 | $411.89 | 8.2% | 159 |
| 2024-05-20 | $446.07 | 2024-08-07 | $442.20 | −0.9% | 79 |
| 2024-08-19 | $469.53 | 2024-09-06 | $455.77 | −2.9% | 18 |
| 2024-10-23 | $488.92 | 2025-02-28 | $494.95 | 1.2% | 128 |
| 2025-06-03 | $520.93 | 2026-01-30 | $625.08 | 20.0% | 241 |
| 2026-04-22 | $648.46 | 2026-08-21 | $713.40 | 10.0% | 121 |
| 2026-09-28 | $739.86 | open | – | 1.3% | – |
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 13 trades
Six of the 12 closed trades won. Ranked by size, 29.44% (2023-01-12 to 2023-09-22), 19.99% (2025-06-03 to 2026-01-30), 10.01% (2026-04-22 to 2026-08-21) and 8.22% (2023-11-15 to 2024-04-22) are the four above 8%, with 6.32% in 2021 and 1.23% in late 2024 completing the winners. The median trade gained 1.23% and the median hold was 121 days. The longest hold was 253 days and the shortest was 10.
The six losers are small: 7.88%, 5.75%, 3.17%, 2.93%, 2.21% and 0.87%. The losers averaged 3.8% and the winners averaged 12.54%, which gives the profit factor of 3.46. The shortfall to holding came from days spent in cash while the fund rose.
The trade list also shows the V-shape failure the template warns about. In August 2024 the rule sold on 2024-08-07 at 442.20, the day the fund's drawdown reached its low, and bought back on 2024-08-19 at 469.53, a higher price. It then sold on 2024-09-06 at 455.77 for a loss of 2.93%. In 2022 the exit on 2022-01-19 was a small loss of 2.21%, but the two entries in August and September lost 5.75% and 7.88% in a rebound that failed.
The open position was entered on 2026-09-28 at 739.86 and was up 1.31% at the end. In total the rule made 25 fills, which is 12 round trips and one open buy.
Re-entry prices and streaks
Comparing each exit with the next entry shows where the cash periods paid and where they cost. After the exit at $270.80 on 2022-09-23 the rule bought again at $272.78 on 2023-01-12, so the months in cash gave up almost nothing in price and avoided the rest of the 2022 decline. Most of the other re-entries came in higher. The rule sold at $353.09 on 2023-09-22 and bought at $380.61 on 2023-11-15. It sold at $494.95 on 2025-02-28 and bought at $520.93 on 2025-06-03. It sold at $625.08 on 2026-01-30 and bought at $648.46 on 2026-04-22. A momentum rule waits for a quarter of gains to confirm the trend, so it buys after the rebound has started.
The three trades with the shortest holds, 10, 18 and 20 days, all lost: 7.88%, 2.93% and 5.75%. A trade that ends within three weeks is one where the 63-day return fell back below 0% almost as soon as it had cleared 5%, and each of these three came in the middle of a bounce that failed. The loss streak peaked at 3 trades, from 2021-10-29 through the two 2022 entries, and the win streak also reached 3, the last three closed trades in the list. The sequence from 2024-10-23 to 2026-08-21 shows the rule at its steadiest, with gains of 1.23%, 19.99% and 10.01% and no losing trade between them.
Largest drawdowns
| Peak | Low point | Depth | Days to low | Recovered | Days to recover |
|---|---|---|---|---|---|
| 2021-09-07 | 2023-01-19 | −21.5% | 499 | 2023-06-02 | 134 |
| 2024-07-10 | 2024-10-31 | −14.8% | 113 | 2025-09-19 | 323 |
| 2026-06-02 | 2026-07-29 | −10.8% | 57 | not yet | – |
Buy-and-hold's deepest drawdown ran from 2021-11-19 to 2022-11-03 and reached −34.2%.
Drawdowns
The maximum drawdown was 21.53%, from 2021-09-07 to 2023-01-19, a fall that took 499 days and recovered by 2023-06-02, 134 days later. The peak is earlier than holding's peak of 2021-11-19, because the rule was invested through the autumn of 2021 and gave up gains in January 2022, August 2022 and September 2022. Holding's drawdown over the 2022 decline was 34.23%, 12.7 points deeper than the rule's.
The second drawdown was 14.81%, from 2024-07-10 to 2024-10-31, and it took 323 days to recover, to 2025-09-19. That is the 2024 sequence of in-and-out trades. Holding's drawdown over the same stretch began on the same date, reached 13.31% on 2024-08-07 and recovered by 2024-11-06, 91 days later. The rule fell further and took far longer to recover, which is the one episode where the switch did worse than holding. The third was 10.85%, from 2026-06-02 to 2026-07-29, with no recovery by the end of the data, and it is the fall that followed July 2026's minus 6.28%.
Holding's second drawdown, 22.38% from 2025-02-19 to 2025-04-08, is the episode where the rule did well. The rule sold on 2025-02-28 and was in cash for March, April and May.
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) | 8.2% | −21.5% | $15,727 | 0.65 |
| 5 basis points | 8.0% | −21.9% | $15,573 | 0.64 |
| 10 basis points | 7.6% | −22.3% | $15,261 | 0.61 |
Execution costs
Costs lower the result slightly. With 5 basis points per fill the CAGR was 8.02% and the final value $15,573. With 10 basis points the CAGR was 7.64% and the final value $15,261. The headline run ended at $15,727. The drawdown moved from 21.53% to 21.9% and 22.26%. With 25 fills over the window the cost is small, and the runs show that the rule does not depend on cheap execution. QQQ's average daily dollar volume was $18,772,024,579.
Cost runs and the Sharpe ratio
The Sharpe ratio was 0.65 in the headline run, 0.637 with 5 basis points per fill and 0.613 with 10. The CAGR fell from 8.21% to 8.02% and then 7.64%. Each fill carries the same cost whether the trade lasts 10 days or 253, so an aggressive slippage assumption hurts the years with many entries, 2022 and 2024, more than the long 2023 and 2025 holds. The ending values were $15,727, $15,573 and $15,261.
Changing the parameters
| Version | CAGR | Max drawdown | Round trips | Win rate | Final value |
|---|---|---|---|---|---|
| Published rules | 8.2% | −21.5% | 12 | 50% | $15,727 |
| Enter above 0% | 10.8% | −20.0% | 26 | 42% | $18,037 |
| Enter above 10% | 4.5% | −22.9% | 9 | 67% | $12,874 |
| Enter above 15% | 3.1% | −17.6% | 7 | 57% | $11,939 |
Changing the entry threshold
The exit stays at a 63-day return below 0% in all three variants, and the entry bar changes. Entering above 0%, with no buffer, returned 10.82% with a drawdown of 20.02%, a Sharpe ratio of 0.791 and 26 trades, 11 of them winners. It ended at $18,037. Entering above 10% returned 4.5% with a drawdown of 22.87%, 9 trades and 6 wins. Entering above 15% returned 3.14% with a drawdown of 17.65%, 7 trades and 4 wins.
The 5% entry sits between them on trades and does worse than the no-buffer version. The buffer that is meant to avoid flip-flopping cost return here: the version with no buffer made more trades, 26, and still gained more, 10.82% against 8.21%, with a shallower drawdown of 20.02%. The higher thresholds of 10% and 15% waited for a stronger run and entered later, and they returned less than the published rule. The result is one fund over one window and three alternatives.
The trade-off visible in the table is that higher entry bars reduce trades and also reduce return, which is the opposite of what a buffer is supposed to do when markets whip. On this fund, with long up-years and one down-year, waiting for confirmation cost more than whipsaws did.
Sharpe ratios and trade counts across the thresholds
The entry-above-0% version has a Sharpe ratio of 0.791 against 0.65 for the published rule, and it made 26 trades with 11 winners. The 10% and 15% versions have ratios of 0.431 and 0.337. The ratio falls in step with return as the entry bar rises, so the higher bars gave up return without buying a smoother ride. The 15% version did have the shallowest drawdown at 17.65%, but a drawdown that low came with a CAGR of 3.14% and a long time spent waiting. Each variant is one path through one window, and the ordering of the variants may differ on another fund. On this fund the rule with no buffer was the best of the four on CAGR, Sharpe ratio and ending value.
How QQQ behaved
| Measure | QQQ |
|---|---|
| Data in this test | 2021-01-04 to 2026-10-02 (1444 sessions) |
| Total return, buy and hold | 151.1% |
| Annualized volatility | 22.4% |
| Deepest drawdown | −35.0% (2021-11-19 to 2022-11-03) |
| Up days | 54.8% |
| Average daily range | 1.58% |
| Average overnight gap | 0.59% |
| Correlation to SPY | 0.94 |
| Correlation to TLT | 0.09 |
| Sessions above the 200-day average | 75.2% |
| Crossings of the 200-day average | 20 |
| Falls of 10% or more from a 20-day high | 21 |
What QQQ's behaviour meant for the rule
QQQ's annualized volatility was 22.39% and it rose on 54.82% of days. The lag-1 autocorrelation of daily returns was minus 0.04, close to zero, so yesterday's move said little about today's. The fund was above its 200-day average on 75.18% of sessions and crossed it 20 times. It fell 10% or more from a 20-day high 21 times over 82 days.
About 64.64% of QQQ's return came overnight, and the rule trades at the open. The rule's cash days therefore missed the larger part of the fund's return along with the intraday part.
The 2022 decline, 32.39% for the calendar year and 35% from peak to trough over 2021-11-19 to 2022-11-03, is the only long fall in the window. Other declines were short: 22.38% over 48 days in early 2025 and 13.31% over 28 days in mid-2024. A three-month lookback needs a quarter of weakness before the 63-day return turns negative. The rule sold in January 2022 and sat out the rest of the fall. In 2024 it sold on 2024-08-07 and re-entered 12 days later at a higher price.
The best day was 2025-04-09 at 11.75% and the worst was 2025-04-04 at minus 6.1%. A rule that sold at the open on 2025-02-28 was out of the market for both days.
Seasonality, RSI statistics and time in the market
The rule was invested 60.9% of the time. QQQ's average return by calendar month was highest in May at 5.19% and November at 4.42%, followed by October at 2.77% and June at 2.7%, and lowest in September at -1.65% and February at -0.77%. The rule's September record has some of the largest losses in the table: -5.56% in 2021, -7.55% in 2022 and -4.7% in 2023, then -3.13% in 2024. Four Septembers in a row lost money for the rule while the fund's September average was only mildly negative. Each calendar month has five or six observations, so the pattern describes this window.
The RSI statistics describe the same churn from another angle. RSI(14) fell below 30 on only 15 sessions and rose above 70 on 145, so overbought readings were far more common than oversold ones. After the 15 oversold sessions the median 5-day return was 3.69%, compared with 0.56% for all days. The rule never trades on those readings. They show that the sharp falls in QQQ were followed by quick rebounds, and a rule keyed to a 63-day return reacts too slowly to catch them. The weekday averages were close to zero, from 0.03% to 0.19%, so the daily open fill carries no weekday cost or benefit visible in the data.
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.
Against the other templates and funds
On QQQ the rule ranked eighth. Weekly 7% target made 16.52% with a drawdown of 19.41%, monthly cycle 14.73%, RSI(2) snapback 14.17% with a 14.27% drawdown, golden cross 13.6%, the 200-day filter 12.9%, trend plus trailing stop 11.27% and SMA 10/50 9.03%. Behind it were EMA 12/26 at 7.85%, momentum breakout at 7.84%, RSI mean reversion at 6.4% and the dip buyer at 5.37%. Golden cross and the 200-day filter did better than the 3-month rule, and so did SMA 10/50, while EMA 12/26 did slightly worse.
Among broad index funds the rule returned 8.8% on VOOG, 8.43% on IOO, 8.25% on VOO, 8.21% on SPY and 8% on QQQM. On the other end it lost money on IWM at minus 1.59% with 20 round trips, on EEM at minus 1.62% and on QQQE at minus 2.86% with 19. QQQ made 12 round trips and had a drawdown of 21.53%, against 18.58% on SPY. The full list is on the 3-month momentum page and the other templates on this fund are on the QQQ page.
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 QQQ?
Over 2021-01-04 to 2026-10-02, 3-month momentum on QQQ returned 8.2% annualized vs 16.7% for buy-and-hold: it trailed buy-and-hold by 8.5% per year, with a maximum drawdown 12.7 points shallower than holding (21.5% vs 34.2%).
How many trades did it make?
12 completed round trips over 5.7 years (25 fills), with 50% 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 QQQ?
It returned 8.21% a year and turned $10,000 into $15,727, against 16.7% and $24,265 for buy-and-hold. The maximum drawdown was 21.53% for the rule and 34.23% for holding.
What are the entry and exit rules?
It buys at the open when the 63-day return is above 5% and sells at the open when that return is below 0%. The 5% entry and 0% exit leave a gap between the two thresholds.
Did the rule avoid the 2022 decline?
Partly. It lost 18.1% in 2022 against 31.7% for holding. Three trades in 2022 all lost, with the worst a 7.88% loss from 2022-09-13 to 2022-09-23, after the rule had been in cash from February to July.
How many trades did it make on QQQ?
12 closed round trips and one open position, 25 fills in all. Six won. The best was 29.44% from 2023-01-12 to 2023-09-22 and the worst was a loss of 7.88%.
Does a lower entry threshold help on QQQ?
Entering above 0% returned 10.82% with a 20.02% drawdown and 26 trades. Entering above 10% returned 4.5% and above 15% returned 3.14%. In this window, a lower bar did better than the published 5%.
Why did the rule trail holding so much in 2024?
It returned 1.8% against 25.1%. It made three entries in the year, sold on 2024-08-07 and bought back on 2024-08-19 at a higher price, and lost on both trades that started in May and August.
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.