DeployQuantSMA 10/50 Trend 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 SMA 10/50 trend rule holds QQQ while the 10-day average is above the 50-day average and goes to cash when the 10-day falls below. It reacts in weeks, where the golden cross reacts in months. From 2021-01-04 to 2026-10-02 it made 18 closed round trips, held a nineteenth position open at the end, and turned $10,000 into $16,425. Buy-and-hold turned the same amount into $24,265. The CAGR was 9.03% against 16.7%, and the maximum drawdown was 20.8% against 34.2%.
The rule ranked seventh of the 12 templates on QQQ. Across the 59 funds, QQQ was the twelfth best result for this template, and the template's median CAGR across all 59 was 2.12%. QQQ is a good fund for this rule in relative terms, and the rule still returned less than holding. Seven of the 18 closed trades won, a win rate of 39%. The average winner gained 13.48% and the average loser lost 3.39%, which gives a profit factor of 2.51.
The rule was invested 65.0% of the time. That left about a third of the window in cash, and QQQ rose through most of it. The gaps to holding were widest in 2023 and 2021, the years the fund rose most. The rule beat holding in 2022 only.
The numbers come from one window of 5.74 years with no fees in the headline run.
Year by year
| Year | SMA 10/50 trend | buy & hold |
|---|---|---|
| 2021 | 10.7% | 28.4% |
| 2022 | −19.6% | −31.7% |
| 2023 | 30.6% | 53.0% |
| 2024 | 11.7% | 25.1% |
| 2025 | 14.4% | 20.4% |
| 2026 | 10.7% | 22.1% |
Six years against buy-and-hold
The year table has one positive gap and five negative ones.
In 2021 the rule returned 10.7% against 28.4%, a gap of 17.7 points. Its first trade began on 2021-04-09, so January to March read 0.0%. That first trade lost 2.53% over 39 days and was closed on 2021-05-18. The rule then bought again on 2021-06-02 and held for 120 days, for a gain of 8.77%, and bought a third time on 2021-10-26.
In 2022 it returned negative 19.6% against negative 31.7%, a gain over holding of 12.1 points. It was the one year in which the rule's speed paid off, but it did not avoid the decline. Four round trips in 2022 all lost: 5.95% from 2022-03-29 to 2022-04-20, 4.20% from 2022-07-22 to 2022-09-07, 7.23% from 2022-11-18 to 2022-12-23, and a small loss of 0.14% that began in October 2021 and closed on 2022-01-11. A fast crossover bought each short rally and sold the following drop. The rule cut the damage compared with holding, and still lost almost a fifth of the account.
In 2023 it returned 30.6% against 53.0%, a gap of 22.4 points, the largest in the table. The year's biggest trade was the entry on 2023-01-24, held 205 days to 2023-08-17 for 27.24%. After that the rule made two short trades in the autumn, losing 2.20% and 3.15%, and re-entered on 2023-11-13.
In 2024 it returned 11.7% against 25.1%, a gap of 13.4 points. The rule made three trades, with a win of 13.63% that closed on 2024-04-18, a win of 4.50% from 2024-05-15 to 2024-07-31, and a loss of 3.01% from 2024-08-27 to 2024-09-05. The rule sold on 2024-07-31, inside the fund's drawdown from 2024-07-10 to 2024-08-07. In 2025 it returned 14.4% against 20.4%, a gap of 6.0 points, the smallest of the five negative gaps. The 2025 year contained the second-best trade, 24.06% from 2025-05-07 to 2025-11-24, and one loss of 7.28% from 2025-01-24 to 2025-03-04.
In 2026 it returned 10.7% against 22.1%, a gap of 11.4 points. The rule made 2 trades to the end of the data: a gain of 8.64% from 2026-04-16 to 2026-07-17, and an open position from 2026-08-14 that was up 2.28%.
The pattern across the six years is that the rule was slow enough to miss the early part of each recovery and fast enough to leave before the late part of each rise. Both effects cost return, and only 2022 gave anything back.
Month by month
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2021 | 0.0% | 0.0% | 0.0% | 1.2% | −3.6% | 6.3% | 2.7% | 4.0% | −4.6% | 1.6% | 1.9% | 1.1% |
| 2022 | −4.6% | 0.0% | −1.9% | −3.9% | 0.0% | 0.0% | 2.8% | −4.9% | −1.8% | 0.0% | 2.1% | −8.9% |
| 2023 | 2.6% | −0.4% | 9.3% | 0.5% | 7.6% | 6.3% | 3.8% | −5.1% | −2.1% | −3.1% | 3.1% | 5.4% |
| 2024 | 1.8% | 5.2% | 1.2% | −3.8% | 0.7% | 5.9% | −2.3% | 0.6% | −2.5% | −0.8% | 5.2% | 0.5% |
| 2025 | −0.6% | −2.6% | −2.7% | 0.0% | 7.7% | 6.3% | 2.4% | 0.9% | 5.3% | 4.7% | −5.2% | −1.7% |
| 2026 | 1.2% | −1.0% | 0.0% | 4.3% | 10.0% | −0.2% | −5.6% | −2.2% | 3.2% | 1.3% | – | – |
What the monthly table shows
The month table has a run of zeros at the start and a few scattered zeros after, and each zero is a month spent in cash. January to March 2021 are zeros, as are February, May, June, and October 2022, April 2025, and March 2026. The rule was invested 65.0% of the time.
The best month was May 2026 at 10.0%. Buy-and-hold's best month was April 2026 at 15.4%, and the rule's April return was 4.3%, since the entry on 2026-04-16 came in the middle of the month. The worst month was December 2022 at negative 8.9%, when the position bought on 2022-11-18 gave up its gains and was sold on 2022-12-23 for a loss of 7.23%. Buy-and-hold's worst month was April 2022 at negative 13.1%, and the rule lost 3.8% in that month, a smaller loss.
The table shows the effect of the lag on recoveries. In March 2023 the rule returned 9.3%, in May 7.6%, in June 6.3%, and in July 3.8%. The fund had started recovering months earlier, but the rule's entry on 2023-01-24 came after the fund had already risen from its November 2022 low.
The losing months are mostly concentrated near turns. August 2023 returned negative 5.1% right after the sell on 2023-08-17, September and October 2023 returned negative 2.1% and negative 3.1% from two short trades, and November 2025 returned negative 5.2% in the month of the exit on 2025-11-24. December 2025 and January 2026 show the fast average whipsawing: the rule sold on 2025-12-29 for a loss of 0.81% and bought back on 2025-12-30, the next day, then sold on 2026-02-10 for a loss of 0.82%. A rule with a 10-day average can flip on one session when the two averages are close together. July 2026 returned negative 5.6% and August negative 2.2%, with the exit on 2026-07-17 partway through the first of those months.
Every trade
SMA 10/50 trend on QQQ made 18 closed round trips and one position still open at the end of the test, an average hold of 73 days, an average winner of 13.48%, an average loser of −3.39%, a profit factor of 2.51, a longest losing streak of 4. It held a position at the close on 65.0% of trading days.
| Entry | Entry price | Exit | Exit price | Return | Days held |
|---|---|---|---|---|---|
| 2021-04-09 | $322.53 | 2021-05-18 | $314.36 | −2.5% | 39 |
| 2021-06-02 | $321.92 | 2021-09-30 | $350.16 | 8.8% | 120 |
| 2021-10-26 | $368.10 | 2022-01-11 | $367.57 | −0.1% | 77 |
| 2022-03-29 | $358.69 | 2022-04-20 | $337.34 | −6.0% | 22 |
| 2022-07-22 | $298.26 | 2022-09-07 | $285.73 | −4.2% | 47 |
| 2022-11-18 | $280.56 | 2022-12-23 | $260.28 | −7.2% | 35 |
| 2023-01-24 | $280.91 | 2023-08-17 | $357.44 | 27.2% | 205 |
| 2023-09-07 | $363.23 | 2023-09-21 | $355.25 | −2.2% | 14 |
| 2023-10-19 | $359.09 | 2023-10-23 | $347.77 | −3.1% | 4 |
| 2023-11-13 | $370.48 | 2024-04-18 | $420.96 | 13.6% | 157 |
| 2024-05-15 | $442.41 | 2024-07-31 | $462.32 | 4.5% | 77 |
| 2024-08-27 | $468.07 | 2024-09-05 | $453.99 | −3.0% | 9 |
| 2024-09-23 | $478.13 | 2025-01-16 | $513.94 | 7.5% | 115 |
| 2025-01-24 | $528.35 | 2025-03-04 | $489.87 | −7.3% | 39 |
| 2025-05-07 | $477.99 | 2025-11-24 | $593.01 | 24.1% | 201 |
| 2025-12-05 | $622.71 | 2025-12-29 | $617.64 | −0.8% | 24 |
| 2025-12-30 | $618.32 | 2026-02-10 | $613.26 | −0.8% | 42 |
| 2026-04-16 | $636.83 | 2026-07-17 | $691.84 | 8.6% | 92 |
| 2026-08-14 | $732.83 | open | – | 2.3% | – |
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.
Seven winners and eleven losers
The trade list is the most useful table on the page for a rule that trades this often. The 18 closed trades split cleanly: 7 wins averaging 13.48% and 11 losses averaging 3.39%.
The winners were 27.24% (2023-01-24 to 2023-08-17, 205 days), 24.06% (2025-05-07 to 2025-11-24, 201 days), 13.63% (2023-11-13 to 2024-04-18, 157 days), 8.77% (2021-06-02 to 2021-09-30, 120 days), 8.64% (2026-04-16 to 2026-07-17, 92 days), 7.49% (2024-09-23 to 2025-01-16, 115 days), and 4.50% (2024-05-15 to 2024-07-31, 77 days). Every winner held for at least 77 days. The two largest held about 200 days each.
The losers held for shorter periods. The worst were 7.28% (2025-01-24 to 2025-03-04, 39 days), 7.23% (2022-11-18 to 2022-12-23, 35 days), 5.95% (2022-03-29 to 2022-04-20, 22 days), and 4.20% (2022-07-22 to 2022-09-07, 47 days). The shortest was a loss of 3.15% in 4 days, from 2023-10-19 to 2023-10-23. The median trade lost 0.81% and the median hold was 47 days. The two longest holds, 205 and 201 days, were both winners.
This is the trend-following profile: a low win rate, a high ratio of average win to average loss, and a result dominated by two or three long holds. The 27.24% and 24.06% trades supply much of the template's gain, and losing either one would have moved the CAGR a lot. The trade list also shows what a faster signal costs. The golden cross on QQQ made 2 round trips in the same window and returned 13.6%. The 10/50 rule made 18 and returned 9.03%. The extra crossovers produced 11 losing trades, and the extra winners did not match the single long hold of the slower rule. The golden cross page for QQQ shows the slower alternative trade by trade.
The longest losing streak was 4 trades and the longest winning streak was 2. The rule held a position at the close on 65.0% of days, with an average hold of 73.3 days. The best trade was 27.24% and the worst was negative 7.28%. The position open at the end was up 2.28% and was marked at the last close.
Largest drawdowns
| Peak | Low point | Depth | Days to low | Recovered | Days to recover |
|---|---|---|---|---|---|
| 2021-11-19 | 2022-12-23 | −20.8% | 399 | 2023-06-14 | 173 |
| 2023-07-18 | 2023-10-23 | −10.4% | 97 | 2024-01-19 | 88 |
| 2026-06-02 | 2026-09-15 | −10.3% | 105 | not yet | – |
Buy-and-hold's deepest drawdown ran from 2021-11-19 to 2022-11-03 and reached −34.2%.
Three drawdowns, and two that ended on a sell date
The largest drawdown was 20.8%, from the peak on 2021-11-19 to the low on 2022-12-23. It took 399 days to reach the low and 173 more to recover, ending on 2023-06-14. Buy-and-hold's low came earlier, on 2022-11-03, and the rule's low came later. The reason is on the trade list. The rule bought on 2022-11-18, after the fund's low, into a rally, and sold on 2022-12-23 for a loss of 7.23%. The drawdown trough is the exit date of that trade. A fast crossover can buy a bounce in a downtrend, and the cost shows up as a new low for the strategy while the fund is still below its peak.
The second drawdown was 10.4%, from 2023-07-18 to 2023-10-23, recovered on 2024-01-19. That trough also lands on an exit date, 2023-10-23, which closed a 4-day trade with a loss of 3.15%. The 97 days to the low and 88 days to recover describe a stretch where the rule sold, bought back twice, and lost small amounts each time.
The third was 10.25%, from 2026-06-02 to 2026-09-15, with no recovery by the end of the data. The rule sold on 2026-07-17 with a gain of 8.64% and bought back on 2026-08-14, and the open position was up 2.28% at the end.
Compared with holding, the largest drawdown was 13.4 points shallower. Holding had drawdowns of 34.2%, 22.4%, and 13.3%. The rule's were 20.8%, 10.4%, and 10.3%, which shows what the faster exit does. The cost is the return. The Sharpe ratio was 0.68 against 0.83 for holding, so the shallower drawdowns did not make up for the lower return on a risk-adjusted basis.
The 2025 drawdown for holding, 22.4% from 2025-02-19 to 2025-04-08, does not appear in the rule's list. The rule sold on 2025-03-04, lost 7.28% on that trade, and was in cash for April. That is the clearest case where the 10/50 rule helped, and the comparison with the golden cross on QQQ, which held through the same decline to a drawdown of 22.2%, shows the value of the faster signal.
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) | 9.0% | −20.8% | $16,425 | 0.68 |
| 5 basis points | 8.5% | −21.1% | $15,985 | 0.65 |
| 10 basis points | 8.3% | −21.5% | $15,798 | 0.64 |
Cost runs with 37 fills
The rule made 37 fills. At 5 basis points of slippage per fill the CAGR was 8.51% and the final value was $15,985. At 10 basis points the CAGR was 8.29% and the final value was $15,798, against $16,425 in the headline run. The maximum drawdown went from 20.8% to 21.14% and then 21.48%, and the Sharpe ratio went from 0.68 to 0.653 and 0.635.
The cost effect is smaller than on the high-turnover templates and larger than on golden cross, which made only 5 fills. Between the two, the 10/50 rule shows how the cost of a signal scales with its speed. Each trade cost the same flat amount, and the faster rule made more of them. The cost runs here are flat offsets and do not model the spread on specific days. QQQ trades an average of $18,772,024,579 a day, so a $10,000 position is a very small part of its volume.
Changing the parameters
| Version | CAGR | Max drawdown | Round trips | Win rate | Final value |
|---|---|---|---|---|---|
| Published rules | 9.0% | −20.8% | 18 | 39% | $16,425 |
| SMA 5/50 | 12.8% | −18.3% | 22 | 36% | $19,925 |
| SMA 20/50 | 7.5% | −28.7% | 13 | 54% | $15,133 |
| SMA 10/100 | 12.2% | −17.5% | 9 | 78% | $19,394 |
Faster and slower versions of the same rule
We tested three variants: a 5-day fast average, a 20-day fast average, and a 100-day slow average.
The published rule returned 9.03% with a 20.8% drawdown and 18 round trips. SMA 5/50 returned 12.76% with an 18.25% drawdown and 22 round trips, of which 8 won. SMA 20/50 returned 7.48% with a 28.7% drawdown and 13 round trips, of which 7 won. SMA 10/100 returned 12.23% with a 17.54% drawdown and 9 round trips, of which 7 won. Final values were $16,425, $19,925, $15,133, and $19,394.
Two of the three variants beat the published rule on both return and drawdown. The 5/50 rule was faster than the published one and did better. The 10/100 rule was slower on the slow leg and did better, with 9 round trips against 18 and a win rate of 7 of 9. The 20/50 rule was slower on the fast leg and did worse, with the deepest drawdown of any version at 28.7%. These results do not point in one direction. Faster helped in one case and slower helped in another, which says the published 10/50 setting sits in a noisy part of the grid. With 18 trades or fewer in each test, a few crossings account for the differences.
The variant table does not show a setting that works across funds. The SMA 10/50 hub page lists the published rule on all 59 funds, where the median CAGR was 2.12%. A setting that is best on QQQ in this window is a result about QQQ and about these dates.
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 |
Why a short-term crossover trails on QQQ
QQQ closed above its 200-day average on 75.2% of sessions and crossed the line 20 times in the window. A fund that spends three quarters of its time in an uptrend rewards holding, and a rule that leaves after every pullback pays for each exit. The fund had 21 falls of 10% or more from a 20-day high, spread over 82 days.
The fund's lag-1 autocorrelation was negative 0.04, so there was nothing in day-to-day returns for a short-term rule to follow, and the rule depended on multi-week trends. Annualized volatility was 22.4%, with an average up day of 1.01% and an average down day of negative 1.06%. The up-day and down-day averages are close, and the fund's gain came from its up days outnumbering its down days, at 54.8% of sessions.
The peers show a similar picture for the template on broad index funds. SMA 10/50 returned 9.16% on IOO, 9.05% on QQQM, 8.95% on VOOG, 8.18% on VV, 7.97% on VOO, and 7.94% on SPY. The large-cap funds had much shallower drawdowns, 12.58% on VV, 12.69% on VOO, and 12.68% on SPY, against 20.8% on QQQ. The weakest results were IWM at negative 1.2%, with 21 round trips, and EEM at negative 3.89%, with 23 round trips and a drawdown of 41.17%. QQQ's 12th place of 59 puts it in the group of funds where the rule worked best. The median across the broad index funds was 7.97%.
For the same fund under other rules, the 200-day filter on QQQ returned 12.9%, the EMA 12/26 trend returned 7.85%, and the weekly 7% target returned 16.52%. SMA 10/50 finished below the 200-day filter and above the EMA 12/26 trend.
The rules
A faster moving-average crossover: hold while the 10-day average is above the 50-day.
- WHEN the market opens · IF not invested AND SMA(10) > SMA(50) · THEN buy with 98% of the sleeve
- WHEN the market opens · IF invested AND SMA(10) < SMA(50) · THEN sell the whole position
A faster version of the golden cross. The 10- and 50-day averages catch intermediate trends measured in weeks rather than years. It enters recoveries earlier and exits breakdowns earlier, with more whipsaw trades in sideways markets. Useful for seeing how signal speed changes a strategy's results.
Good for: trending assets with multi-week swings, such as leveraged index ETFs.
Watch out: several false signals a year is normal; each whipsaw costs a small loss and they add up in flat markets.
Rule mechanics on a fast signal
Both rules compare the two averages at the open. A buy needs the 10-day above the 50-day, and a sell needs it below. Orders fill on the minute bar after the signal, and the position uses 98% of the sleeve.
There is no buffer between the two conditions. When the averages are close, a one-day move can flip the signal in either direction, and the rule has no way to wait for confirmation. The pair of trades on 2025-12-05 to 2025-12-29 and 2025-12-30 to 2026-02-10 is an example in this data: the rule sold on one day and bought back the next. Both trades lost under 1%.
The rule also has no stop. A position stays open until the averages cross, so a sharp fall that reverses within days does not trigger an exit, while a slow fall that lasts weeks does. The 7.28% loss from 2025-01-24 to 2025-03-04 is the clearest case: the position was held through a decline for 39 days before the signal came.
The other 11 templates in the test use different mechanisms. The trend with trailing stop template adds a price-based exit to a 200-day entry and returned 11.27% on QQQ. The 3-month momentum rule uses a return threshold and returned 8.21%. The choice between them comes down to which kind of delay a holder is willing to accept, and the table on QQQ shows the cost of each over one window.
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 SMA 10/50 trend beat buy-and-hold on QQQ?
Over 2021-01-04 to 2026-10-02, SMA 10/50 trend on QQQ returned 9.0% annualized vs 16.7% for buy-and-hold: it trailed buy-and-hold by 7.7% per year, with a maximum drawdown 13.4 points shallower than holding (20.8% vs 34.2%).
How many trades did it make?
18 completed round trips over 5.7 years (37 fills), with 39% of round trips closing profitably.
Why 10 and 50 days?
A common intermediate-trend pairing. It reacts within weeks and ignores single bad days. Both windows are editable parameters in DeployQuant.
Does SMA 10/50 beat buy-and-hold on QQQ?
No. From 2021-01-04 to 2026-10-02 the rule returned 9.03% a year against 16.7% for holding. The maximum drawdown was 20.8% against 34.2%. It beat holding only in 2022, by 12.1 points.
How many trades did SMA 10/50 make on QQQ?
It made 18 closed round trips and had one position open at the end. Seven won, a win rate of 39%. The average winner gained 13.48% and the average loser lost 3.39%.
What was the best trade for SMA 10/50 on QQQ?
The entry on 2023-01-24, held 205 days to 2023-08-17, gained 27.24%. The second best was 24.06% from 2025-05-07 to 2025-11-24. The worst was a loss of 7.28% from 2025-01-24 to 2025-03-04.
Is 10/50 the best setting on QQQ?
Not in this test. SMA 5/50 returned 12.76% and SMA 10/100 returned 12.23%, both with shallower drawdowns than the published rule. SMA 20/50 returned 7.48% with a 28.7% drawdown.
How do trading costs change the SMA 10/50 result?
With 37 fills, the CAGR went from 9.03% to 8.51% at 5 basis points and 8.29% at 10 basis points. The final value fell from $16,425 to $15,798 at the higher setting.
How does SMA 10/50 compare with golden cross on QQQ?
Golden cross returned 13.6% with 2 round trips and a 22.2% drawdown. SMA 10/50 returned 9.03% with 18 round trips and a 20.8% drawdown. The faster rule traded much more and returned less over this window.
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.