DeployQuantWeekly Entry + 7% Target on SOXL
Direxion Daily Semiconductor Bull 3X Shares: 3x daily leveraged semiconductors, one of the more volatile ETFs listed. Backtest 2021-01-04 to 2026-10-02, $10,000 starting capital, computed by the same engine that runs live DeployQuant strategies.
SOXL is the Direxion Daily Semiconductor Bull 3X Shares, a 3x daily leveraged semiconductor fund and one of the more volatile ETFs in the study. The weekly 7% target buys it at the first open of each week with 98% of the sleeve, rests a limit order 7% above the entry, and sells on Thursday at 2:00pm if the trade is losing. From 2021-01-04 to 2026-10-02 the rule turned $10,000 into $44,050, a CAGR of 29.47%, with a max drawdown of 71.36% and a Sharpe ratio of 0.74. Buy-and-hold on SOXL ended at $52,126, a CAGR of 33.32%, with a max drawdown of 89.62% and a Sharpe ratio of 0.82.
The rule made 282 round trips and won 169 of them, a 60% win rate. The average win was 7.58% and the average loss was 8.93%, so the profit factor was 1.31. The average hold was 2.5 days and the median hold was 2 days. The strategy was invested 45.4% of the time. It finished 3.85 points of CAGR behind holding and with a drawdown 18.26 points shallower.
On SOXL the rule ranks 4th of the 12 templates, behind the golden cross at 45.93%, the RSI(2) snapback at 39.24% and the monthly cycle at 32.86%. Across the 59 funds SOXL ranks 4th for this strategy, and the median CAGR of the rule over all 59 funds is 6.61%. Within the leveraged group it is third for this rule, behind TECL at 35.49% and ROM at 30.27%. Everything is a backtest of one window with no fees in the headline run.
Year by year
| Year | weekly 7% target | buy & hold |
|---|---|---|
| 2021 | −34.6% | 119.0% |
| 2022 | −13.2% | −84.8% |
| 2023 | 65.5% | 211.2% |
| 2024 | −14.7% | −12.1% |
| 2025 | 62.2% | 53.6% |
| 2026 | 251.7% | 285.2% |
Year by year
The rule returned 34.6% negative in 2021, 13.2% negative in 2022, 65.5% in 2023, 14.7% negative in 2024, 62.2% in 2025 and 251.7% in 2026 to date. Buy-and-hold returned 119% in 2021, 84.8% negative in 2022, 211.2% in 2023, 12.1% negative in 2024, 53.6% in 2025 and 285.2% in 2026.
It beat holding in two years. In 2022 it lost 13.2% against holding's 84.8% loss, a gap of 71.6 points. In 2025 it returned 62.2% against 53.6%. It trailed in 2021 by 153.6 points, in 2023 by 145.7 points, in 2024 by 2.6 points and in 2026 by 33.5 points. The pattern is the usual one for a capped-gain rule. In years when the fund more than doubled, the 7% target took small gains and left the long run to holding. In the year the fund lost 84.8%, the rule's early exits cut the loss to a fraction.
The 2021 result looks odd. The fund gained 121.74% on its price series that year and the rule lost 34.6%. The rule traded 46 times in 2021 and won 23, so half of the trades won, and the losses outweighed the wins. The worst trade of the whole backtest, 30.27% negative, fell in March 2021, and the monthly table shows 2021-03 at 24.8% negative and 2021-05 at 14.7% negative. The fund ended the year well ahead of where it started, and the weekly rule, entering only at Monday opens and exiting at targets, did not capture that gain.
The 2026 figure of 251.7% is a partial year through 2026-10-02 and it is the largest single year in the table. It comes from monthly returns of 18.26% in January, 30.34% in February, 19.78% in March and 49.38% in April, which is also the rule's best month of the window. The rule ended 2026 with 28 wins in 38 trades.
Month by month
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2021 | 12.9% | −8.3% | −24.8% | −5.5% | −14.7% | 1.2% | −4.8% | −12.4% | −17.6% | 21.5% | 12.8% | 9.2% |
| 2022 | −2.3% | 22.9% | 6.2% | −5.5% | 0.9% | −45.0% | 30.4% | 0.1% | −23.1% | 3.4% | 23.5% | 1.3% |
| 2023 | 9.6% | 17.9% | 16.3% | −9.7% | −10.7% | 23.9% | 11.8% | −6.0% | −11.1% | −9.6% | 16.2% | 12.2% |
| 2024 | −4.1% | 18.6% | 15.2% | −16.6% | 2.7% | 9.8% | −21.7% | −4.2% | 4.6% | −1.7% | −7.5% | −2.9% |
| 2025 | 5.2% | 4.8% | −26.2% | −8.0% | 42.4% | 28.5% | 0.8% | 14.9% | 24.9% | 24.8% | −31.0% | −4.9% |
| 2026 | 18.3% | 30.3% | 19.8% | 49.4% | 10.8% | 0.5% | 2.3% | −0.3% | 0.6% | 11.8% | – | – |
The monthly record
The worst month was 2022-06 at 44.96% negative. Holding's worst month was 2026-07 at 56.9% negative, and the rule's July 2026 return was 2.25% positive, which is a large gap in the rule's favour. The next worst months for the rule were 2025-11 at 30.97% negative, 2025-03 at 26.2% negative, 2021-03 at 24.8% negative and 2022-09 at 23.12% negative. The best months were 2026-04 at 49.38%, 2025-05 at 42.37%, 2022-07 at 30.39% and 2026-02 at 30.34%.
The 2022-06 loss is worth a closer look. The rule bought in the week of 2022-06-13 at 16.6 and sold at 12.86 on 2022-06-16 for a loss of 22.53%, which is the third worst trade of the backtest. The following month, 2022-07, returned 30.39%. A time-based exit that sells on Thursday realizes the loss, and a rebound after the exit goes uncaptured until the next Monday entry. The same pattern shows in 2025-11, a loss of 30.97% after the strong 2025-10 gain of 24.81%.
The July 2026 data shows the other side of the Thursday rule. SOXL had its worst drawdown of the year in that month, and holding lost 56.9%. The rule's third drawdown, 33.51% from 2026-07-24 to 2026-07-29, was smaller than the 69.27% that holding suffered from 2026-06-22 to 2026-07-29, and it had not recovered by the end of the window.
Every trade
weekly 7% target on SOXL made 282 closed round trips, an average hold of 3 days, an average winner of 7.58%, an average loser of −8.93%, a profit factor of 1.31, a longest losing streak of 6. It held a position at the close on 45.4% of trading days.
Best 10 round trips
| Entry | Entry price | Exit | Exit price | Return | Days held |
|---|---|---|---|---|---|
| 2026-04-06 | $54.52 | 2026-04-08 | $66.23 | 21.5% | 2 |
| 2022-12-12 | $11.91 | 2022-12-13 | $14.34 | 20.4% | 1 |
| 2026-09-28 | $147.27 | 2026-10-02 | $165.39 | 12.3% | 4 |
| 2025-09-15 | $29.28 | 2025-09-18 | $32.73 | 11.8% | 3 |
| 2023-06-12 | $22.70 | 2023-06-13 | $25.26 | 11.3% | 1 |
| 2025-09-29 | $35.07 | 2025-10-02 | $38.93 | 11.0% | 3 |
| 2021-02-01 | $34.37 | 2021-02-02 | $37.93 | 10.4% | 1 |
| 2022-09-06 | $12.56 | 2022-09-09 | $13.85 | 10.3% | 3 |
| 2022-11-14 | $12.69 | 2022-11-15 | $13.97 | 10.1% | 1 |
| 2026-08-24 | $112.37 | 2026-08-27 | $123.25 | 9.7% | 3 |
Worst 10 round trips
| Entry | Entry price | Exit | Exit price | Return | Days held |
|---|---|---|---|---|---|
| 2021-03-01 | $39.64 | 2021-03-04 | $27.64 | −30.3% | 3 |
| 2024-07-15 | $64.90 | 2024-07-18 | $49.75 | −23.3% | 3 |
| 2022-06-13 | $16.60 | 2022-06-16 | $12.86 | −22.5% | 3 |
| 2024-04-08 | $43.88 | 2024-04-18 | $34.43 | −21.5% | 10 |
| 2022-08-29 | $15.12 | 2022-09-01 | $11.96 | −20.9% | 3 |
| 2025-04-14 | $11.05 | 2025-04-16 | $8.81 | −20.3% | 2 |
| 2026-08-17 | $151.50 | 2026-08-20 | $121.09 | −20.1% | 3 |
| 2021-05-10 | $34.21 | 2021-05-13 | $27.94 | −18.3% | 3 |
| 2026-07-27 | $139.92 | 2026-07-30 | $114.47 | −18.2% | 3 |
| 2022-09-12 | $14.40 | 2022-09-15 | $11.84 | −17.8% | 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 trades that decided the result
The best trade gained 21.48% from 2026-04-06 to 2026-04-08, entering at an adjusted 54.52 and exiting at 66.23. A gain larger than 7% means the fill came above the limit price, which happens when a fund opens above a resting limit order. The second best gained 20.4% in a single day, from 2022-12-12 to 2022-12-13. The third gained 12.3% from 2026-09-28 to 2026-10-02, entering at 147.27, which is the last trade of the window. The first trade of the backtest gained exactly 7% from 2021-01-04 to 2021-01-06, entering at 30.86.
The worst trades were all short. The worst lost 30.27% from 2021-03-01 to 2021-03-04 and the next lost 23.34% from 2024-07-15 to 2024-07-18, both over 3 days. The third lost 22.53% from 2022-06-13 to 2022-06-16 and the fifth lost 20.9% from 2022-08-29 to 2022-09-01. The fourth, from 2024-04-08 to 2024-04-18 for a loss of 21.54%, ran 10 days. The longest hold in the window was 17 days. A loss of 20% to 30% on a trade held for three days is the cost of a three-times fund that falls fast and a rule with no price stop.
The median trade gained 6.99% over 2 days. That is almost exactly the target, and it says that the typical win ended at the limit. The streaks were long: the longest winning streak ran 12 trades and the longest losing streak ran 6. The yearly trade counts were 46 in 2021, 49 in 2022, 51 in 2023, 47 in 2024, 51 in 2025 and 38 in 2026. The share of winners was 23 of 46 in 2021, 30 of 49 in 2022, 29 of 51 in 2023, 28 of 47 in 2024, 31 of 51 in 2025 and 28 of 38 in 2026.
The win rate was therefore at its lowest in 2021, at half of the trades, and at its highest in 2026, which matches the 251.7% return that year. The win rate rose and fell with the fund's trend.
Largest drawdowns
| Peak | Low point | Depth | Days to low | Recovered | Days to recover |
|---|---|---|---|---|---|
| 2021-02-12 | 2022-10-12 | −71.4% | 607 | 2025-09-02 | 1056 |
| 2025-10-31 | 2025-12-17 | −38.6% | 47 | 2026-02-25 | 70 |
| 2026-07-24 | 2026-07-29 | −33.5% | 5 | not yet | – |
Buy-and-hold's deepest drawdown ran from 2021-12-27 to 2022-10-14 and reached −89.6%.
Three drawdowns
The deepest drawdown was 71.36%, from a peak on 2021-02-12 to a trough on 2022-10-12, which is 607 days, with recovery on 2025-09-02, 1,056 days after the trough. Holding fell 89.62% from 2021-12-27 to 2022-10-14 and recovered on 2026-02-25. The rule's peak came earlier, because it was down 34.6% in 2021 while the fund was still rising.
The second drawdown was 38.56%, from 2025-10-31 to 2025-12-17, and it recovered on 2026-02-25 after 70 days. The third was 33.51% in 5 days, from 2026-07-24 to 2026-07-29, and it had not recovered by 2026-10-02. For comparison holding had drawdowns of 69.27% from 2026-06-22 to 2026-07-29 and 43.04% from 2026-02-25 to 2026-03-30.
A drawdown of 71.36% is deep for any rule. The underlying fund's own max drawdown was 90.49% from 2021-12-27 to 2022-10-14, and it took until 2026-02-25 to recover, a stretch of 1,043 sessions. The rule cut that depth without removing it.
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) | 29.5% | −71.4% | $44,050 | 0.73 |
| 5 basis points | 23.0% | −71.5% | $32,774 | 0.67 |
| 10 basis points | 20.1% | −72.7% | $28,637 | 0.63 |
Costs on 282 round trips
With 564 fills, the rule is sensitive to costs. At 5 basis points the CAGR fell from 29.47% to 22.97%, and the end value fell from $44,050 to $32,774. At 10 basis points the CAGR was 20.11% and the end value was $28,637. The max drawdown moved from 71.36% to 71.45% and 72.71%, and the Sharpe ratio from 0.74 to 0.665 and 0.633. Both cost runs finished well behind holding at 33.32%.
The reason the cost is large is the size of each fill relative to the edge. The average win was 7.58%, and each round trip pays the cost twice on a fund that moves 5% a day. SOXL is liquid, with average daily dollar volume of about $2,243,524,454 and a median minute volume of 101,543.5 shares, so a bid-ask cost of 5 or 10 basis points is plausible for orders of this size, though fills at the 7% limit price depend on the book at the time.
Changing the parameters
| Version | CAGR | Max drawdown | Round trips | Win rate | Final value |
|---|---|---|---|---|---|
| Published rules | 29.5% | −71.4% | 282 | 60% | $44,050 |
| 5% target | 16.4% | −64.1% | 292 | 67% | $23,929 |
| 10% target | 36.5% | −71.2% | 254 | 52% | $59,639 |
| 15% target | 52.3% | −75.2% | 218 | 44% | $111,878 |
Changing the profit target
Three variants changed the target and left the other rules alone.
A 5% target gave a CAGR of 16.41% with a max drawdown of 64.14%, on 292 trades with 195 winners. The Sharpe ratio was 0.556. A tighter target won more often and returned much less. A 10% target gave 36.48%, a drawdown of 71.23% and a Sharpe ratio of 0.804, on 254 trades with 133 winners. A 15% target gave 52.29% with a drawdown of 75.21% and a Sharpe ratio of 0.932, on 218 trades with 96 winners, and an end value of $111,878.
The pattern is steady. As the target rose from 5% to 7%, 10% and 15%, the CAGR rose from 16.41% to 29.47%, 36.48% and 52.29%, while the number of trades fell and the win rate fell, from 195 wins in 292 trades to 96 wins in 218. The drawdown rose each step as well. The 10% and 15% variants both finished ahead of holding's 33.32%, which the 7% template did not. On a fund that moved 5% on an average day and rose 450.98% in total, a wider target let winners run farther.
That result is specific to the window. SOXL's price series rose in a few violent stretches, and the 15% variant benefitted from them. On the leveraged funds that fell over the window, such as TMF at 11.83% negative or UST at 8.16% negative, the base rule lost money, and these variants were not run there. The test also changed one setting at a time, and the 15% result rests on 96 winning trades in one period. The tests do not show whether the same ordering would hold in a different five-year window.
How SOXL behaved
| Measure | SOXL |
|---|---|
| Data in this test | 2021-01-04 to 2026-10-02 (1444 sessions) |
| Total return, buy and hold | 451.0% |
| Annualized volatility | 113.3% |
| Deepest drawdown | −90.5% (2021-12-27 to 2022-10-14) |
| Up days | 53.4% |
| Average daily range | 7.77% |
| Average overnight gap | 3.17% |
| Correlation to SPY | 0.80 |
| Correlation to QQQ | 0.87 |
| Correlation to TLT | 0.05 |
| Sessions above the 200-day average | 58.5% |
| Crossings of the 200-day average | 44 |
| Falls of 10% or more from a 20-day high | 102 |
What SOXL did and how the other leveraged funds compare
SOXL is built to return three times the daily move of a semiconductor index. In the leverage table the underlying is SOXX, with a realized beta of 2.97 and an R-squared of 1. Over the window SOXL returned 450.98% in total and SOXX returned 388.35%. A daily-rebalanced 3x of SOXX would have returned 854.1%, and the fund's actual result was lower than that ideal. The yearly returns of the fund were 121.74% in 2021, 85.64% negative in 2022, 226.54% in 2023, 12.33% negative in 2024, 54.98% in 2025 and 289.88% in 2026 to date. In 2024 SOXX gained 13% and SOXL lost 12.33%, which is the decay of a daily-reset fund in a choppy year.
The fund rose on 53.36% of sessions. Its beta to SPY was 5.49 and to QQQ was 4.42, with correlations of 0.8 and 0.87. Its closest relatives by correlation are SOXX at 1.00, XLK at 0.9, TECL at 0.9, ROM at 0.9 and QQQM at 0.88. The best day was 2025-04-09 at 55.98% and the worst was 2026-06-05 at 30.71% negative, with 2025-04-03 at 29.79% negative close behind. The overnight log return was 261.62% and the intraday log return was 94.21% negative, so all of the fund's gain came between the close and the next open. A rule that buys at the open and exits at a limit order during the day collects the intraday part, which was negative in this window, and the rule's exposure of 45.4% left it out of many of the overnight gaps.
The fund was above its 200-day average on 58.55% of sessions and crossed it 44 times. RSI(14) fell below 30 on only 7 sessions, and the median forward 5-day return after those was 11.92%, with a 20-day median of 16.81%. RSI(2) below 10 occurred on 140 sessions with medians of 1.7% and 6.65%, against baselines of 1.22% and 2.96%. In the median case, dips in SOXL were followed by large rebounds, which suits a rule that buys every week.
Among the other leveraged funds, the weekly target returned 35.49% on TECL with a 43.24% drawdown, 30.27% on ROM, 29.44% on TQQQ, 25.62% on QLD, 23.28% on SPUU, 21.27% on SSO and 13.73% on FAS. SOXL's drawdown of 71.36% was the deepest in the group, ahead of TMF at 60.65% and FAS at 53.69%. Against the other templates on SOXL, the golden cross returned 45.93% with a 69.3% drawdown, the RSI(2) snapback 39.24% with 58.51%, and the monthly cycle 32.86% with 90.24%. The strategy overview and the SOXL page hold the full tables.
The rules
Buy at the first open of each week, rest a +7% profit target, and exit Thursday afternoon if the trade is losing.
- WHEN the first session of the week opens · IF not invested · THEN buy with 98% of the sleeve (once per week)
- WHILE invested · a managed limit order rests at entry price × 1.07
- WHEN it's 2:00pm on the week's second-to-last session · IF the position is losing · THEN sell everything
A weekly swing template: enter Monday, aim for +7%, and do not carry a loser into the weekend. The profit target rests at the broker as a real limit order the whole time (DeployQuant maintains it as a managed order). The Thursday-afternoon exit gives losing trades a time deadline instead of a price stop.
Good for: volatile assets that regularly swing 7% within a week, such as leveraged ETFs.
Watch out: the time-based exit realizes many small losses by design. The template needs the +7% winners to outnumber them, which the per-ETF results test directly.
A weekly rule on a fund that moves 5% on an average day
SOXL's average up day was 5.31% and its average down day was 5.29% negative. The average intraday range was 7.77%, the annualized volatility was 113.25%, and the fund fell 10% within 20 days on 102 separate occasions, covering 769 trading days. A 7% target on a fund like this is a small move, one that the fund can make in a single session. That explains the median hold of 2 days and the 60% win rate: most weeks, the limit order was reached within a day or two of the entry.
The Thursday exit is what cuts losing weeks. If the trade is under water at 2:00pm on the second-to-last session of the week, the rule sells, so a losing position does not carry into the weekend. On SOXL the rule's losses are large when they come. The average loss was 8.93%, larger than the 7.58% average win, and the worst trade lost 30.27% from 2021-03-01 to 2021-03-04. A time exit with no price stop allows a three-times fund to fall for three days before the rule acts.
The design therefore relies on the win rate staying high. With a 60% win rate the profit factor was 1.31. The template description says it needs the 7% winners to outnumber the small losses, and the data on SOXL shows winners outnumbering losers but with losers that are larger than winners. The result is positive and below holding. The strategy exited at the target more often than it exited at a loss, and the few big losses took back much of the gain.
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 weekly 7% target beat buy-and-hold on SOXL?
Over 2021-01-04 to 2026-10-02, weekly 7% target on SOXL returned 29.5% annualized vs 33.3% for buy-and-hold: it trailed buy-and-hold by 3.9% per year, with a maximum drawdown 18.3 points shallower than holding (71.4% vs 89.6%).
How many trades did it make?
282 completed round trips over 5.7 years (564 fills), with 60% of round trips closing profitably.
Why exit on Thursday?
It is a time deadline. The template does not hold a losing trade over the weekend gap. On holiday-shortened weeks the exit moves to the week's second-to-last session automatically.
Does the 7% target rest at the broker?
Yes. Deployed live, the target is a real GTC limit order that DeployQuant places and maintains.
How did the weekly 7% target do on SOXL?
It returned a CAGR of 29.47% over 2021-01-04 to 2026-10-02, ending at $44,050 from $10,000, with a max drawdown of 71.36% and a Sharpe ratio of 0.74. Buy-and-hold returned 33.32% with an 89.62% drawdown. The rule ranks 4th of 12 templates on SOXL.
What was the win rate and why does the rule still trail holding?
It won 169 of 282 trades, a 60% win rate. The average win was 7.58% and the average loss was 8.93%, so profit factor was 1.31. The 7% cap leaves the large rallies to holding, and the rule was invested 45.4% of the time.
How much does a different target change the result?
A 5% target gave 16.41%, a 10% target gave 36.48% and a 15% target gave 52.29% with a drawdown of 75.21%. The 10% and 15% targets beat holding's 33.32%. This is one window and one fund.
What happens with trading costs?
At 5 basis points the CAGR was 22.97% and at 10 basis points it was 20.11%, down from 29.47%. With 564 fills the rule is sensitive to costs. Both cost runs finished behind holding.
What was the worst trade?
A loss of 30.27% from 2021-03-01 to 2021-03-04, entering at an adjusted 39.64 and exiting at 27.64. The rule exits losing trades only on Thursday afternoon, so a three-times fund can fall for several days first.
Did the rule protect against drawdowns?
Partly. Its max drawdown was 71.36% against 89.62% for holding. In 2022 it lost 13.2% against 84.8% for holding. It still had a deep drawdown, with 607 days from peak to trough and 1,056 days to recover.
How reliable is this result?
It rests on one window with SOXL's large rallies in 2023, 2025 and 2026. The headline run has no fees or slippage, and the results are hypothetical and are not a forecast.
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.