DeployQuantSOXX trading strategies, backtested
iShares Semiconductor ETF: the semiconductor industry, a high-momentum sector. Every DeployQuant template run on SOXX over 5.7 years of minute data, same engine, same window, sorted by return.
SOXX holds the semiconductor industry. In the 5.7 years from 2021-01-04 to 2026-10-02 it returned 31.11% a year in the test and turned $10,000 into $47,348. It paid for that return with deep drawdowns: 45.06% from 2021-12-27 to 2022-10-14, then 40.93% from 2024-07-10 to 2025-04-08, then 28.99% from 2026-06-22 to the 2026-07-29 low, which had not recovered by the end of the data.
Twelve templates were run on it. None beat buy-and-hold on CAGR, and all twelve had a shallower maximum drawdown than holding. The best was weekly 7% target at 30.54% a year, 0.57 points behind holding, with a 22.76% drawdown and a Sharpe ratio of 1.03 against 0.92 for the fund. The worst was SMA 10/50 trend at 7.65%, and it was the only template that never beat holding in any calendar year.
The pattern is a fund that rewards owning it through everything and a set of templates that give up part of that return in exchange for a smaller loss in 2022. The window is one stretch of 5.7 years, the headline numbers carry no fees, and the order of the templates describes SOXX over this window only.
| Strategy | CAGR | max DD | Sharpe | trades | win rate | final value |
|---|---|---|---|---|---|---|
| Weekly Entry + 7% Target | 30.5% | −22.8% | 1.03 | 176 | 39% | $46,183 |
| First-to-Last Day of Month | 29.2% | −44.5% | 0.90 | 69 | 62% (+1 open) | $43,478 |
| Golden Cross (SMA 50/200) | 28.2% | −28.8% | 1.02 | 3 | 100% (+1 open) | $41,701 |
| Drawdown Dip Buyer + 8% Target | 27.5% | −39.3% | 0.93 | 19 | 100% (+1 open) | $40,303 |
| SMA-200 Trend + 15% Trailing Stop | 19.8% | −36.4% | 0.77 | 14 | 57% (+1 open) | $28,222 |
| 200-Day SMA Regime Filter | 19.2% | −39.2% | 0.77 | 19 | 26% (+1 open) | $27,422 |
| EMA 12/26 Trend | 18.0% | −34.0% | 0.77 | 22 | 50% (+1 open) | $25,811 |
| RSI(2) Dip Snapback | 17.0% | −23.8% | 0.77 | 162 | 61% | $24,642 |
| 20-Day Momentum + Trailing Stop | 16.7% | −30.2% | 0.78 | 18 | 61% (+1 open) | $24,308 |
| 3-Month Momentum Switch | 14.3% | −35.4% | 0.64 | 16 | 44% | $21,542 |
| RSI(14) Mean Reversion | 12.8% | −35.0% | 0.64 | 14 | 86% | $19,948 |
| SMA 10/50 Trend | 7.7% | −40.3% | 0.42 | 17 | 53% (+1 open) | $15,268 |
| Buy & hold SOXX | 31.1% | −45.1% | 0.92 | – | – | $47,348 |
How each template compared with holding SOXX
Holding gained 43.6% in 2021, lost 34.5% in 2022, then gained 65.3%, 12.8%, 40.1% and 95% in 2023, 2024, 2025 and the part of 2026 in the data. The 2026 figure matters for every comparison below. A 95% gain in nine months is the largest single-year move in the table, and every template that was in cash for any part of it fell far behind. The gap to holding in 2026 was minus 30.5 points for weekly 7% target, minus 33.8 for EMA 12/26 and minus 88.7 for RSI(2) snapback.
Weekly 7% target made 176 round trips with a 39% win rate and a profit factor of 1.63. The average win was 7.33% and the average loss was 2.93%. Its best trade ran from 2026-03-30 to 2026-04-08 for 12.83%, and its worst was from 2022-08-22 to 2022-09-01, a loss of 10.95%. In 2022 it lost 6.9% against 34.5% for the fund, and in 2024 it gained 26.3% against 12.8%. Those were the two years it beat holding. It was exposed 72.7% of the time.
Monthly cycle returned 29.17% and had a 44.51% drawdown, 0.55 points shallower than holding, so it is a template that mostly reproduces the fund. It beat holding in 2024, 2025 and 2026. Its best trade was April 2026, a gain of 36.49%, and its worst was June 2022 at minus 18.29%.
Golden cross made 3 round trips, all winners, at 28.24% a year with a Sharpe of 1.02 and a 28.82% drawdown. The trade that mattered ran from 2023-01-30 to 2024-11-12 for a gain of 70.09% over 652 days. Its first trade, from 2021-10-19 to 2022-03-23, made 3.08% and was out before most of the 2022 fall. It bought again on 2025-07-09 and was still holding at the end of the data with a gain of 141.84%. Its drawdown of 28.82% is the one that began on 2026-06-22. Golden cross is the template that does the most with the fewest trades on this fund, and three trades is a small sample.
The drawdown dip buyer returned 27.48% from 19 round trips, every one a win at about 8%. That 100% win rate has a cost on the other side: the strategy holds a losing position until the target is hit, and its drawdown was 39.25%. It had an open position at the end that was down 5.54%, entered on 2026-06-15. In 2023 it gained 89.6% against 65.3% for the fund, a lead of 24.3 points over the fund, and the gap in 2026 was minus 33.5 points.
The two trend templates built on the 200-day average, trend plus trailing stop and the 200-day filter, returned 19.81% and 19.21%. The 200-day filter had 19 round trips, a 26% win rate and a profit factor of 1.04. Eight of those round trips closed in 2022 and only one won. In 2025 it closed 4 and won none. The trailing-stop version had 14 round trips and a 57% win rate with a 36.41% drawdown. Its worst trade was the entry on 2024-07-25 that stopped out on 2024-08-05 for a loss of 14.84%.
EMA 12/26 returned 17.96% with a 50% win rate and 22 round trips. Its best trade ran 197 days from 2025-05-07 to 2025-11-20 for 55.59%, and its worst was a loss of 10.65% in June 2022. RSI(2) returned 17.01% with 162 round trips and the shallowest drawdown among templates after weekly 7% target, 23.81%. Momentum breakout returned 16.73%, 3-month momentum returned 14.3% and RSI(14) mean reversion returned 12.78% with a win rate of 86% and a profit factor of 4.99. That last template made 14 round trips and one of them, entered 2022-01-21, lost 17.06% over 181 days, which dwarfs every other trade that template made.
SMA 10/50 returned 7.65% with 18 round trips and a 40.32% drawdown. It had no year where it beat holding, because even 2022 was a loss of 37.1% against 34.5% for the fund. It was the template with the deepest cut in CAGR, and its 17 round trips included 5 closed in 2022 of which only 1 won.
Other sector funds sit lower. XLK held at 22.07% and its best template, monthly cycle, made 20.86%. XLF held at 12.58% and RSI mean reversion made 11.82%. XLY held at 6.27% and weekly 7% target made 10.19%, so a template beat the fund there. The same is true for XLP, where RSI(2) made 8.22% against 5.8%, and VOX, where golden cross made 11.2% against 9.08%. On SOXX the fund itself rose so fast that holding was the hard number to beat.
How each strategy traded SOXX
| Strategy | Time in market | Avg hold (days) | Best trade | Worst trade | Profit factor | With 10 bps slippage |
|---|---|---|---|---|---|---|
| weekly 7% target | 72.7% | 8 | 12.8% | −10.9% | 1.63 | 25.6% |
| monthly cycle | 95.2% | 28 | 36.5% | −18.3% | 2.26 | 26.1% |
| golden cross | 60.3% | 270 | 70.1% | 1.0% | – | 27.9% |
| dip buyer | 64.4% | 66 | 9.6% | 8.0% | – | 27.4% |
| trend + trailing stop | 69.0% | 99 | 66.2% | −14.8% | 2.28 | 19.1% |
| 200-day regime filter | 60.9% | 42 | 24.5% | −7.7% | 1.04 | 18.3% |
| EMA 12/26 trend | 60.6% | 58 | 55.6% | −10.7% | 3.33 | 17.2% |
| RSI(2) snapback | 37.1% | 5 | 11.7% | −11.4% | 1.57 | 10.7% |
| momentum breakout | 51.4% | 58 | 49.7% | −10.0% | 2.96 | 16.0% |
| 3-month momentum | 58.4% | 77 | 137.4% | −11.7% | 3.27 | 13.5% |
| RSI mean reversion | 30.7% | 47 | 16.0% | −17.1% | 4.99 | 12.3% |
| SMA 10/50 trend | 59.5% | 73 | 34.5% | −15.1% | 1.75 | 7.0% |
Time in market, trade quality and costs
Exposure ran from 95.2% for monthly cycle to 30.7% for RSI(14). Between those sat weekly 7% target at 72.7%, trend plus trailing stop at 69%, the dip buyer at 64.4%, the 200-day filter at 60.9%, EMA 12/26 at 60.6%, golden cross at 60.3%, SMA 10/50 at 59.5%, 3-month momentum at 58.4%, momentum breakout at 51.4% and RSI(2) at 37.1%. Because 2026 contributed 95% in a single partial year, the templates with the most days in the market had the best chance to reach it.
Median holding periods were 3 days for weekly 7% target and 4 days for RSI(2), against 155 days for golden cross. The best single trade on the page is the 3-month momentum entry on 2025-06-04, which was held 446 days and gained 137.38%. That one trade carries much of the template's result. The 200-day filter and golden cross each hold a similar entry from June and July 2025 at the end of the data, with open gains of 175.24% and 141.84%.
Costs hit the busy templates. RSI(2) went from 17.01% with no costs to 13.78% with 5 basis points per trade and 10.69% with 10. That is the largest drop in the table, since it makes about 162 round trips. Weekly 7% target went from 30.54% to 28.62% and 25.63%. Monthly cycle went to 27.62% and 26.14%. EMA 12/26 went to 17.49% and 17.21%, and golden cross barely moved, 28.21% and 27.95%. The dip buyer was almost unchanged at 27.44% and 27.4%. At 10 basis points golden cross led at 27.95%, then the dip buyer at 27.4% and monthly cycle at 26.14%, with weekly 7% target fourth at 25.63%.
SOXX is liquid, with $1,048,392,630 of average daily dollar volume and a median minute volume of 5,853 shares. The cost runs show what a slower fill would cost the high-turnover templates. The fund moves 2.58% in an average day, with an average overnight gap of 1.07%, so the 10 basis point run is a reasonable check on the busy templates.
Open positions and the second drawdown
Several templates end the window inside a position that is large compared with anything they closed. The 200-day filter entered on 2025-06-04 at 213.96 and was up 175.24% at the end. The golden cross entered on 2025-07-09 at 243.51 and was up 141.84%. The 3-month momentum switch held the same 2025-06-04 entry until 2026-08-24 and closed it for 137.38%, the largest single trade on the page. A reader comparing CAGRs sees 28.24% for the golden cross, but a large part of it sits in one open position that had not yet met a sell signal.
The fund's own history has a second drawdown that the 2022 discussion misses. From the 2024-07-10 peak to the 2025-04-08 trough, holding lost 40.93%, nearly as deep as the 45.06% fall of 2022, and it recovered on 2025-09-18, 163 days later. The weekly 7% target, with a 22.76% drawdown, and the RSI(2) snapback, at 23.81%, had the shallowest maximum drawdowns. The trend plus trailing stop made a 14.84% loss on 2024-07-25 to 2024-08-05, its worst trade, right after the 2024 peak. The 200-day filter's worst trade was only 7.7% in January 2025, and its 2025 return was 18.6% against 40.1% for holding.
The third drawdown began on 2026-06-22 and reached 28.99% by 2026-07-29 without recovering by the end of the data. July 2026 was the worst month for most templates, including holding at -21.13%, the golden cross at -21%, the dip buyer at -20.53% and the trend plus trailing stop at -29.44%. The last of these is worse than holding because the stop-and-reenter rule took the trade losses of 12.58% on 2026-07-06 to 2026-07-17 and 10.62% from 2026-07-20 to 2026-07-29 in a falling fund, then reentered on 2026-07-30.
The RSI(14) mean reversion template stands apart on quality. It won 86% of its 14 round trips with a profit factor of 4.99, and its best trades were 15.96% from 2025-11-21 to 2025-12-10 and 15.07% from 2026-09-03 to 2026-09-23. Its largest loss was 17.06% from 2022-01-21 to 2022-07-21. At 12.78% a year it earned less than most of the other templates, and the 14 trades are a small sample on a rule that needs a rare oversold reading.
How SOXX behaved
| Measure | SOXX |
|---|---|
| Data in this test | 2021-01-04 to 2026-10-02 (1444 sessions) |
| Total return, buy and hold | 388.4% |
| Annualized volatility | 38.1% |
| Deepest drawdown | −45.8% (2021-12-27 to 2022-10-14) |
| Up days | 53.9% |
| Average daily range | 2.58% |
| Average overnight gap | 1.07% |
| Correlation to SPY | 0.79 |
| Correlation to QQQ | 0.87 |
| Correlation to TLT | 0.05 |
| Sessions above the 200-day average | 70.7% |
| Crossings of the 200-day average | 38 |
| Falls of 10% or more from a 20-day high | 43 |
Calendar years
| Year | Return |
|---|---|
| 2021 | 44.7% |
| 2022 | −35.1% |
| 2023 | 67.0% |
| 2024 | 13.0% |
| 2025 | 40.7% |
| 2026 (part) | 95.9% |
Biggest single days
| Best day | Move |
|---|---|
| 2025-04-09 | 18.8% |
| 2022-11-10 | 10.5% |
| 2026-07-30 | 8.8% |
| 2026-06-11 | 8.3% |
| 2025-05-12 | 7.3% |
| Worst day | Move |
|---|---|
| 2026-06-05 | −10.4% |
| 2025-04-03 | −10.1% |
| 2025-04-10 | −8.3% |
| 2026-06-23 | −7.9% |
| 2025-01-27 | −7.8% |
Average return by calendar month
| Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 4.5% | 2.5% | −0.3% | 1.6% | 11.4% | 4.8% | −0.5% | −1.6% | −0.4% | 2.0% | 8.4% | 1.4% |
Most and least correlated funds
| Most correlated | Least correlated | ||
|---|---|---|---|
| SOXL | 1.00 | SOXS | -1.00 |
| XLK | 0.90 | TECS | -0.90 |
| TECL | 0.90 | REW | -0.89 |
| ROM | 0.90 | QID | -0.87 |
| QQQM | 0.87 | SQQQ | -0.87 |
Why SOXX behaves the way it does
SOXX's annualized volatility was 38.08% and its beta to SPY was 1.84, against 1.49 to QQQ. The correlation to SPY was 0.79 and to QQQ 0.87. Its correlation to TLT was 0.05. The high beta explains the size of the 2022 loss, 35.07% for the calendar year, and the 95.9% gain in 2026.
The average up day was 1.78% and the average down day was minus 1.78%, with 53.92% of days up. The distribution is wide. The average intraday range was 2.58% and the average overnight gap was 1.07%. About 86.53% of the fund's return came overnight and 13.47% came intraday. A rule that buys at the open and holds only inside the day has little to capture on this fund.
The best day was 2025-04-09 at 18.84% and the second was 2022-11-10 at 10.49%. The worst were 2026-06-05 at minus 10.45% and 2025-04-03 at minus 10.11%. Large moves in both directions are close together: April 2025 holds both the 10.11% fall and the 18.84% rebound, and June 2026 holds the 10.45% fall and an 8.3% rebound on 2026-06-11. A daily rule that reacts to one large move faces a reversal within days. The fund fell 10% or more from a 20-day high 43 times, across 197 days.
The fund spent 70.68% of sessions above its 200-day average and crossed it 38 times, so the fund changed sides of its long-term average often. That is consistent with the 19 round trips and 26% win rate of the 200-day filter.
The RSI statistics describe why the mean reversion templates did as well as they did. RSI(14) went under 30 on only 7 sessions, and the median forward 5-day return after those sessions was 6.1% against 0.63% for all days. The forward 20-day return was 11.58% against 2.11%. Seven observations is a small sample, but the direction fits the 86% win rate of RSI(14) mean reversion. RSI(2) went under 10 on 133 sessions, with a median forward 5-day return of 0.89% and a forward 20-day return of 3.58%. That is better than the baseline, though modest next to the fund's 2.58% average daily range. The lag-1 autocorrelation was minus 0.09, a small negative number that fits with short-term bounces.
May averaged 11.44% and November 8.44%, against minus 1.61% in August and minus 0.5% in July. Each month has five or six observations, so these averages rest on a handful of years. The 2026 data includes April, the best month for holding at 40.13%, and July, the worst at minus 21.13%. Several templates show that same April and July pair as their best and worst months, which says how much of the six-year result sits in those two months.
The calendar years show a fund that alternates. SOXX gained 44.66% in 2021, lost 35.07% in 2022, gained 66.96% in 2023, then added only 13% in 2024 before rising 40.68% in 2025. The 2024 year is where several templates did well relative to holding. Golden cross made 18.5% and weekly 7% target made 26.3%, against 12.8% for holding. The modest year-end figure sits on top of a 40.93% drawdown that began on 2024-07-10 and bottomed on 2025-04-08, so the fund fell hard inside the stretch. It took 163 days from that low to recover, against 425 days after the 2022 low.
The drawdown column separates templates that look alike on CAGR. Weekly 7% target at 30.54% had a 22.76% drawdown, and monthly cycle at 29.17% had a 44.51% drawdown. The Sharpe ratio shows the same split: 1.03 for weekly 7% target against 0.90 for monthly cycle and 0.92 for holding.
Most-correlated funds are SOXL at 1.00, XLK and TECL at 0.9, ROM at 0.9 and QQQM at 0.87. SOXL is the 3x version of this index. Its page shows how the same templates did with leverage.
Calendar averages, autocorrelation and sector peers
The weekday averages are small next to the daily range: 0.32% on Monday, 0.03% on Tuesday, 0.23% on Wednesday, 0.13% on Thursday and 0% on Friday. The fund's average up day and down day were each 1.78%, and the lag-1 autocorrelation was -0.09. A negative figure means a large move had a small tendency to partly reverse the next day, which fits the RSI(2) snapback winning 61% of its 162 round trips. The effect was small: the template ended behind holding at 17.01% against 31.11%.
The month-of-year table has May at 11.44% and November at 8.44% as the strong months, and August at -1.61% and July at -0.5% as the weak ones. The best and worst calendar months of the whole window were April 2026 at 40.13% for holding and July 2026 at -21.13%. Both fall in the final year, which makes the final months of the data matter more than any seasonal average. The 2026 return of 95.9% in nine months is larger than the full-year returns of 2021, 2023, 2024 and 2025.
Intraday and overnight returns split 21.22% and 136.36% in log terms. For a rule that buys at the open and sells later in the same session, that means most of the fund's gain was not available inside the session. The templates that hold overnight, such as the golden cross and the monthly cycle, collected the overnight part. The short-hold rules, such as the RSI(2) snapback at 37.1% exposure, collected less of it.
The sector peers help to place the semiconductor result. XLK held at 22.07% and its monthly cycle made 20.86%. XLY held at 6.27% and the weekly 7% target made 10.19%. VOX held at 9.08% and the golden cross made 11.2%. XLP held at 5.8% and the RSI(2) snapback made 8.22%. On the slower sector funds a template could beat holding. On SOXX none did, and the weekly 7% target came closest at 30.54% a year against 31.11%.
SOXX traded about $1,048,392,630 a day with a median minute volume of 5,853 shares. The fund had a 38.08% annualized volatility and spent 70.68% of sessions above its 200-day average. The test starts at a single point, 2021-01-04, and the strongest months fall near its end, so the results describe this window and do not extend to other periods.
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 SOXX?
Of the 12 templates tested on SOXX over 2021-01-04 to 2026-10-02, the strongest by CAGR was weekly 7% target at 30.5% (max drawdown 22.8%), versus 31.1% 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 SOXX?
0 of 12 templates beat SOXX buy-and-hold (31.1% CAGR) on this window; 12 of 12 had a shallower maximum drawdown than holding (45.1%).
What was the best strategy for SOXX from 2021 to 2026?
Weekly 7% target, at 30.54% a year with a 22.76% maximum drawdown. Buy-and-hold made 31.11% with a 45.06% drawdown. The template ended at $46,183 against $47,348 for holding.
Did any strategy beat buy-and-hold on SOXX?
None of the 12 templates beat SOXX buy-and-hold on CAGR. All 12 had a shallower maximum drawdown than the 45.06% for holding. SMA 10/50 trend was furthest behind at 7.65% a year.
Why did golden cross do well on SOXX with so few trades?
It made 3 round trips and all of them won. One held 652 days from 2023-01-30 to 2024-11-12 and gained 70.09%, and an open position from 2025-07-09 was up 141.84% at the end. Three trades is a small sample, so the 28.24% CAGR depends heavily on those two.
How much did SOXX fall in 2022?
The calendar year was a loss of 35.07%. The deepest drawdown was 45.06% in the test, from 2021-12-27 to 2022-10-14, and the fund took until 2023-12-13 to recover.
How much do trading costs matter on SOXX?
RSI(2) snapback fell from 17.01% to 10.69% at 10 basis points per trade, the biggest drop of any template. Golden cross and the dip buyer changed by well under a point. The headline numbers have no fees.
How volatile is SOXX?
Yes. Annualized volatility was 38.08% and the beta to SPY was 1.84. The average daily range was 2.58%, and the fund fell 10% or more from a 20-day high on 43 occasions, across 197 days.
Other sector etfs
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.