DeployQuantSPY trading strategies, backtested
SPDR S&P 500 ETF Trust: the most-traded S&P 500 ETF and a common benchmark for US large caps. Every DeployQuant template run on SPY over 5.7 years of minute data, same engine, same window, sorted by return.
SPY is the most-traded S&P 500 fund. All 12 DeployQuant strategy templates ran on it from 2021-01-04 to 2026-10-02. Each run starts with $10,000. Buy-and-hold ended at $21,820, a 14.56% annualized return, with a 23.99% maximum drawdown and a Sharpe ratio of 0.95. None of the 12 templates beat that return. Ten had a shallower maximum drawdown than holding, and all 12 finished with a positive return.
The closest was the weekly 7% target at 13.88% a year, followed by the RSI(2) snapback at 13.17% and the monthly cycle at 12.84%. The weakest was the dip buyer at 2.61%. SPY gives a clean test of what a rule adds to a diversified index fund, because the fund itself returned 125.29% over 1444 sessions.
| Strategy | CAGR | max DD | Sharpe | trades | win rate | final value |
|---|---|---|---|---|---|---|
| Weekly Entry + 7% Target | 13.9% | −24.1% | 1.01 | 83 | 30% (+1 open) | $21,085 |
| RSI(2) Dip Snapback | 13.2% | −12.7% | 1.21 | 172 | 71% (+1 open) | $20,342 |
| First-to-Last Day of Month | 12.8% | −22.4% | 0.88 | 69 | 70% (+1 open) | $20,007 |
| EMA 12/26 Trend | 10.2% | −11.8% | 0.98 | 19 | 58% (+1 open) | $17,463 |
| 20-Day Momentum + Trailing Stop | 9.1% | −13.1% | 0.91 | 3 | 67% (+1 open) | $16,475 |
| Golden Cross (SMA 50/200) | 8.7% | −18.1% | 0.77 | 2 | 50% (+1 open) | $16,150 |
| 200-Day SMA Regime Filter | 8.3% | −18.0% | 0.84 | 15 | 40% (+1 open) | $15,805 |
| 3-Month Momentum Switch | 8.2% | −18.6% | 0.86 | 9 | 78% (+1 open) | $15,728 |
| SMA 10/50 Trend | 7.9% | −12.7% | 0.79 | 17 | 59% (+1 open) | $15,507 |
| SMA-200 Trend + 15% Trailing Stop | 5.5% | −31.1% | 0.50 | 3 | 33% (+1 open) | $13,562 |
| RSI(14) Mean Reversion | 4.4% | −17.1% | 0.42 | 13 | 77% | $12,780 |
| Drawdown Dip Buyer + 8% Target | 2.6% | −16.0% | 0.29 | 2 | 100% | $11,593 |
| Buy & hold SPY | 14.6% | −24.0% | 0.95 | – | – | $21,820 |
What separated the templates
The 12 templates split into three groups by how they trade. The first group stays almost fully invested and trades on a calendar or a target. The weekly 7% target held a position 87.5% of the time and the monthly cycle 95.2%. Their results sit close to buy-and-hold because they are in the market for nearly every session, and they differ from holding mainly in what they skip. The weekly target made 83 round trips with a 30% win rate, an average winner of 7.08% and an average loser of 1.62%. Those numbers show the design: a rule that exits losers on Thursday afternoon and sells winners at 7% wins rarely and loses small. Its profit factor was 1.99 and its Sharpe ratio 1.01, the second highest after the RSI(2) snapback.
The second group is the trend rules: the EMA 12/26 trend, the golden cross, the 200-day regime filter, the SMA 10/50 trend, the momentum breakout and the 3-month momentum switch. They returned between 7.94% and 10.2% a year. Their maximum drawdowns ranged from 11.84% for the EMA rule to 18.58% for the 3-month switch. The EMA 12/26 rule did best of the six at 10.2% with an 11.84% drawdown, and it made 19 round trips against 2 for the golden cross and 3 for the momentum breakout.
The third group is the mean-reversion rules and the dip buyer, which hold cash most of the time. The RSI(2) snapback held a position 36.2% of days, the RSI(14) mean reversion 26.8% and the dip buyer 22.8%. The RSI(2) rule is the exception that worked: 172 round trips, a 71% win rate, an average hold of 4.4 days and a 12.69% drawdown, the shallowest of any template that traded more than 20 times. The RSI(14) rule returned 4.37% with 13 round trips, and the dip buyer returned 2.61% from 2 trades, both winners at 8%. With exposure that low, the weaker rules left the fund's gains on the table.
No template beat buy-and-hold over the whole window, and year-level wins are few. Only the weekly target (2022, 2024, 2025) and the monthly cycle (2024, 2025) beat holding in more than one calendar year. Nine other templates beat it in 2022 alone, and the trailing stop strategy beat it in no year. The shape is consistent with a fund that rose in five of six calendar years. Rules that step aside have one year in which to pay off.
How other broad index funds compare
The same templates on neighbouring funds show that SPY's result is typical for large caps. QQQ held for 16.7% a year and its best template, the weekly 7% target, returned 16.52%. IOO held for 16.72% and its best template, the monthly cycle, returned 15.8%. VOO held for 14.42% and its best was the RSI(2) snapback at 14.08%. In each of those three the best template finished just under buy-and-hold.
The exceptions are funds with weaker buy-and-hold results. IWM held for 7.5% a year and the RSI(2) snapback returned 11.38% there. EEM held for 6.58% and the weekly target returned 11.62%. VOOG held for 15.73% and the weekly target returned 19.51%. On SPY, where holding returned 14.56%, no template cleared it.
The 2022 bear market and the April 2025 drop
SPY lost 17.8% on a buy-and-hold basis in 2022, with a drawdown of 23.99% from 2022-01-03 to 2022-10-12. The RSI(2) snapback lost 2.5% that year, the best result in the table by a wide margin. The EMA 12/26 trend lost 8.3%, the golden cross lost 8.4%, the momentum breakout lost 10.6%, the SMA 10/50 trend lost 10% and the 200-day filter lost 12.4%. The dip buyer lost 4.3% in 2022 from a position opened on 2022-05-10, which closed on 2023-06-09 for an 8% gain after 395 days.
The trailing stop strategy lost 29.8% in 2022, worse than holding, because both of its stop-outs came after declines and its re-entries came inside rallies. The two calendar-based rules lost about what holding lost: the monthly cycle lost 19.6% and the weekly target 16.5%, against 17.8%. Staying invested through the calendar brought the damage along.
April 2025 is the second test. Buy-and-hold fell 18.41% from the peak on 2025-02-19 to a low on 2025-04-08, and the best day of the window, 9.39% on 2025-04-09, followed. The monthly cycle's drawdown over those weeks was 20.27%, deeper than holding's, and the golden cross had an 18.13% drawdown from the same peak. The RSI(2) snapback had a smaller one of 8.21% from 2025-02-21 to 2025-03-13 and bought on 2025-04-07, a trade that returned 9.16% in 3 days and is the best trade the rule made. The weekly target bought on 2025-04-21 at $513.02 and sold on 2025-05-01 at $551.20 for 7.44%.
The two drops are a good pair for comparison because they have different shapes. 2022 was a slow decline with rallies inside it, and rules that exited on weakness or bought oversold readings came out ahead. April 2025 was a fall of 48 days followed by a recovery, and buy-and-hold regained its high on 2025-06-26. The 200-day filter sold on 2025-03-11 and returned 10.2% for 2025. The golden cross sold on 2025-04-17 and returned 0.2%. Buy-and-hold returned 17.4%.
How each strategy traded SPY
| Strategy | Time in market | Avg hold (days) | Best trade | Worst trade | Profit factor | With 10 bps slippage |
|---|---|---|---|---|---|---|
| weekly 7% target | 87.5% | 21 | 7.9% | −6.9% | 1.99 | 10.9% |
| RSI(2) snapback | 36.2% | 4 | 9.2% | −6.3% | 2.18 | 6.8% |
| monthly cycle | 95.2% | 28 | 9.1% | −9.0% | 2.00 | 10.0% |
| EMA 12/26 trend | 71.2% | 78 | 20.3% | −3.5% | 5.11 | 9.5% |
| momentum breakout | 64.6% | 281 | 23.3% | −8.1% | 3.36 | 9.0% |
| golden cross | 67.7% | 480 | 34.9% | −3.0% | 11.00 | 8.7% |
| 200-day regime filter | 67.2% | 82 | 33.4% | −3.6% | 3.49 | 7.7% |
| 3-month momentum | 65.7% | 137 | 26.3% | −7.6% | 4.95 | 7.9% |
| SMA 10/50 trend | 69.4% | 83 | 18.9% | −5.9% | 4.00 | 7.4% |
| trend + trailing stop | 75.9% | 362 | 31.9% | −14.4% | 1.11 | 5.4% |
| RSI mean reversion | 26.8% | 43 | 8.5% | −8.3% | 2.62 | 3.8% |
| dip buyer | 22.8% | 239 | 8.0% | 8.0% | – | 2.6% |
Trading style and costs
The trading table shows how differently the 12 templates use the fund. Average hold ranges from 4 days for the RSI(2) snapback to 480 for the golden cross. Best and worst trades show where each rule's risk sits. The golden cross had a best trade of 34.9%, an 811-day hold, and a worst of 3.04%. The 200-day filter's best was 33.4% and its worst 3.6%. The trailing stop's worst trade was a 14.37% loss. The weekly target's best trade of 7.87% is close to its 7% target, so it cannot have a large winner by design, and its worst trade was a 6.89% loss in May and June 2022.
Profit factors vary from 1.11 on the trailing stop strategy to 11 on the golden cross. The golden cross figure comes from two trades, one winner of 34.93% and one loser of 3.04%, and carries little information. The RSI(2) snapback's 2.18 comes from 172 round trips and means more. The EMA 12/26 trend's 5.11 comes from 19 round trips with an average winner of 7.24% and an average loser of 2.02%.
Costs separate the busy rules from the quiet ones. The headline run has no slippage. With 10 basis points per fill, the RSI(2) snapback fell from 13.17% to 6.75% a year, and its Sharpe ratio from 1.21 to 0.66. At 5 basis points it returned 9.98%. This is the most cost-sensitive result in the table, because the rule makes 345 fills in 5.7 years, each with a small edge. The weekly target went from 13.88% to 12.87% at 5 basis points and 10.87% at 10. The monthly cycle went from 12.84% to 11.44% and then 10.02%.
The slow rules barely moved. The golden cross went from 8.71% to 8.68% and then 8.65%. The momentum breakout went from 9.09% to 9.05% and 9.01%, and the dip buyer did not change at 2.61%. The EMA 12/26 trend went from 10.2% to 9.74% and 9.49%. At 10 basis points the ranking changes. The weekly target is still first at 10.87%, and the RSI(2) snapback at 6.75% falls behind the monthly cycle at 10.02%, the EMA 12/26 trend at 9.49%, the 200-day filter at 7.69% and the SMA 10/50 trend at 7.37%. Whether 10 basis points is realistic for SPY depends on order size and timing. The test shows the direction and the size of the effect: busier rules lose more of their edge.
The RSI(2) snapback has the highest Sharpe ratio, 1.21, with exposure of 36.2%, against 0.95 for buy-and-hold at full exposure. Time out of the market lowers volatility as well as return, and for this rule the Sharpe ratio rose.
Long trades and open positions
The slow templates owe their results to a handful of long trades. The golden cross made 34.93% from 2023-01-27 to 2025-04-17, and that single trade carries its 8.71% CAGR. The trend plus trailing stop made 31.85% from 2022-12-01 to 2025-04-04. The momentum breakout made 23.3% from 2023-11-27 to 2025-03-13. The 200-day filter made 33.39% from 2023-11-02 to 2025-03-11. All four exited in March or April 2025, in the weeks around the 2025-04-08 low.
Each of these templates then re-entered in May or June 2025 and ends the window in a position larger than any it closed in the same period. The momentum breakout entered on 2025-05-06 at 549.06 and was up 40.18%. The trailing-stop template entered on 2025-05-13 at 573.39 and was up 34.23%. The golden cross entered on 2025-06-30 at 608.92 and was up 26.4%. These open gains are part of the CAGR in the table, so each of these results would change if a sell signal had arrived earlier or later.
The RSI(2) snapback and the weekly 7% target work at the other extreme. Their best trades were small: 9.16% from 2025-04-07 to 2025-04-10 for the snapback, and 7.87% from 2023-10-30 to 2023-11-14 for the weekly target. Their worst were 6.31% from 2025-02-24 to 2025-03-18 and 6.89% from 2022-05-23 to 2022-06-16. With 172 and 83 round trips, no single trade decided either result. That is the main reason the snapback's 71% win rate and 2.18 profit factor are more informative than the golden cross's 11, which rests on two trades.
The EMA 12/26 trend sits between the two groups. It made 19 round trips with a 58% win rate, and its best trade was 20.29% from 2025-05-05 to 2025-11-20. Its worst trade lost 3.47% from 2023-09-01 to 2023-09-21. A profit factor of 5.11 on a rule with a 10.2% CAGR means the losses were small and the wins modest.
How SPY behaved
| Measure | SPY |
|---|---|
| Data in this test | 2021-01-04 to 2026-10-02 (1444 sessions) |
| Total return, buy and hold | 125.3% |
| Annualized volatility | 16.4% |
| Deepest drawdown | −24.5% (2022-01-03 to 2022-10-12) |
| Up days | 54.4% |
| Average daily range | 1.15% |
| Average overnight gap | 0.44% |
| Correlation to QQQ | 0.94 |
| Correlation to TLT | 0.08 |
| Sessions above the 200-day average | 78.0% |
| Crossings of the 200-day average | 30 |
| Falls of 10% or more from a 20-day high | 11 |
Calendar years
| Year | Return |
|---|---|
| 2021 | 30.5% |
| 2022 | −18.2% |
| 2023 | 26.2% |
| 2024 | 24.9% |
| 2025 | 17.7% |
| 2026 (part) | 13.8% |
Biggest single days
| Best day | Move |
|---|---|
| 2025-04-09 | 9.4% |
| 2022-11-10 | 5.5% |
| 2025-05-12 | 3.3% |
| 2022-06-24 | 3.1% |
| 2022-11-30 | 3.1% |
| Worst day | Move |
|---|---|
| 2025-04-04 | −6.0% |
| 2025-04-03 | −4.8% |
| 2022-09-13 | −4.3% |
| 2022-05-18 | −4.0% |
| 2022-06-13 | −3.8% |
Average return by calendar month
| Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1.4% | 0.1% | 0.8% | 0.6% | 3.0% | 1.3% | 3.1% | 0.7% | −2.2% | 2.9% | 4.0% | 0.2% |
Most and least correlated funds
| Most correlated | Least correlated | ||
|---|---|---|---|
| VOO | 1.00 | SDS | -1.00 |
| SSO | 1.00 | SH | -1.00 |
| VV | 1.00 | SPDN | -1.00 |
| SPUU | 0.99 | SQQQ | -0.94 |
| IOO | 0.96 | QID | -0.94 |
What SPY's behaviour says about the order
SPY's daily statistics point to the same ranking. The fund was up on 54.4% of days, the average up day and average down day were both 0.74%, and lag-1 autocorrelation was negative 0.02. Yesterday's return told almost nothing about today's. A rule that waits for a trend to confirm gets in after a move has started, and a rule that buys weakness relies on the next few days reversing.
The RSI figures from the fund show the second effect. The RSI(2) fell below 10 on 146 sessions, and the median 5-day forward return after those sessions was 0.89% against a baseline median of 0.42%. The median 20-day forward return was 2.29% against a baseline of 1.73%. RSI(14) fell below 30 on only 18 sessions, with a median 5-day forward return of 3.74% and a 20-day median of 2.86%. The RSI(14) sample is small, which fits the 13 round trips and 26.8% exposure of the RSI(14) mean reversion rule. The RSI(2) sample is larger, and it fits the 71% win rate and the 122 winning round trips.
The 200-day average figures explain the trend rules. SPY closed above its 200-day average on 77.99% of sessions and crossed it 30 times. Rules based on that line were invested most of the time and traded in clusters when the fund hovered near the average. The fund made 11 falls of 10% or more from a 20-day high, spread over 23 sessions, so fast declines were rare and slow ones were the larger problem.
The best days of the window cluster around the worst. The five best days include 9.39% on 2025-04-09 and 5.48% on 2022-11-10, and the worst include 5.98% on 2025-04-04 and 4.78% on 2025-04-03. A rule that is out of the market after a fall misses the rebound day as well. Overnight gaps averaged 0.44% and made up 62.46% of the log return, with 37.54% earned during the session.
Calendar patterns in the data are thin. September averaged negative 2.21% and November 4.01%, each from 5 or 6 years. The fund's correlation to QQQ was 0.94 and to TLT 0.08. VOO, SSO and VV had a correlation of 1.00, and the best template on VOO was also the RSI(2) snapback, at 14.08%. The IOO page shows the same templates on a global large-cap fund with a correlation of 0.96 to SPY.
The limits are those of the test: 5.7 years, one bear market and one sharp drop, no fees in the headline runs, daily decisions, and rule counts as low as 2 round trips for the slow templates. The ranking describes this window of SPY and gives no forecast.
Drawdowns, overnight returns and the funds that mirror SPY
Holding SPY had three drawdowns of note. The first, 23.99% from 2022-01-03 to 2022-10-12, took 427 days to recover, until 2023-12-13. The second, 18.41% from 2025-02-19 to 2025-04-08, recovered on 2025-06-26. The third, 8.76% from 2026-01-27 to 2026-03-30, recovered on 2026-04-14. The longest drawdown lasted 488 sessions. The fund had 11 falls of 10% or more from a 20-day high, spread over 23 days.
Overnight log returns were 49.65% and intraday 29.84%, a split of 62.46% overnight and 37.54% intraday. A rule that closes every position before the close keeps only the intraday share. The weekday averages were 0.14% on Monday, 0.01% on Tuesday, 0.08% on Wednesday, 0.02% on Thursday and 0.06% on Friday, which are small compared with a 0.74% average daily move.
The oversold readings had more evidence than on most funds. RSI(14) fell below 30 on 18 sessions, with a median forward 5-day return of 3.74% and a 20-day median of 2.86%, against baselines of 0.42% and 1.73%. RSI(2) below 10 on 146 sessions had a 20-day median of 2.29%. The 18 sessions are not enough to size a rule on, which fits the RSI(14) mean reversion template making 13 round trips and a 4.37% CAGR with 26.8% exposure. The RSI(2) rule had the larger sample and the stronger result.
SPY traded about $30,453,859,411 a day on average, with a median of 107,870 shares a minute. A $10,000 account is small against that volume, so the cost runs of 5 and 10 basis points work as a stress test of turnover.
The funds most correlated with SPY were VOO, SSO and VV at 1, SPUU at 0.99 and IOO at 0.96. Among the peers, VV held at 14.26% and its best template, the weekly 7% target, made 14.94%. VTV held at 13.44% and its monthly cycle made 12.08%. QQQE held at 9.83% and the weekly target made 11.21%. On SPY, the same weekly target made 13.88% against 14.56%.
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 SPY?
Of the 12 templates tested on SPY over 2021-01-04 to 2026-10-02, the strongest by CAGR was weekly 7% target at 13.9% (max drawdown 24.1%), versus 14.6% 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 SPY?
0 of 12 templates beat SPY buy-and-hold (14.6% CAGR) on this window; 10 of 12 had a shallower maximum drawdown than holding (24.0%).
Which strategy worked best on SPY?
The weekly 7% target had the highest return at 13.88% a year. The RSI(2) snapback had the best risk profile, with a Sharpe ratio of 1.21 and a 12.69% drawdown. None of the 12 templates beat buy-and-hold's 14.56%.
Did any strategy beat buy-and-hold on SPY?
None of the 12 beat holding on return from 2021-01-04 to 2026-10-02. Ten had a shallower maximum drawdown than holding's 23.99%. The weekly 7% target and the monthly cycle beat holding in more than one calendar year.
How did the strategies handle the 2022 bear market?
The RSI(2) snapback lost 2.5% in 2022 against 17.8% for buy-and-hold. The EMA 12/26 trend lost 8.3% and the golden cross 8.4%. The trend plus trailing stop strategy lost 29.8% and was the only template worse than holding.
Do trading costs matter on SPY?
They matter most for the busy rules. At 10 basis points per fill the RSI(2) snapback fell from 13.17% to 6.75% a year. The golden cross moved from 8.71% to 8.65%.
Why did the RSI(2) snapback do well on SPY?
SPY's median 5-day return after RSI(2) fell below 10 was 0.89% against a 0.42% baseline, across 146 sessions. The rule made 172 round trips with a 71% win rate and spent 36.2% of days in the market.
Is a buy-and-hold of SPY hard to beat in this window?
In this data it was. SPY rose in five of the six calendar years, and a rule that sits in cash has to make up the return it misses. Only the 2022 decline rewarded stepping aside.
Other broad index 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.