DeployQuantFAS trading strategies, backtested
Direxion Daily Financial Bull 3X Shares: 3x daily leveraged financials. Every DeployQuant template run on FAS over 5.7 years of minute data, same engine, same window, sorted by return.
FAS is a 3x daily leveraged fund on US financials. Between 2021-01-04 and 2026-10-02 buying and holding it made 18.34% a year, turned $10,000 into $26,295, and went through a 66.26% max drawdown. The underlying sector fund, XLF, returned 102.88% in total over the same dates, and FAS returned 180.7% on its price series. The path in between ran 125.73% in 2021, -43.13% in 2022, 84.17% in 2024 and -15.01% so far in 2026.
Twelve templates ran on FAS with the same engine and the same $10,000. Three beat buy-and-hold on CAGR. The drawdown dip buyer made 24.15%, the RSI(14) mean reversion made 22.26%, and the RSI(2) snapback made 18.66%, a hair above the fund. Ten of twelve had a shallower max drawdown than holding, and eight finished with a positive CAGR.
All three winners are rules that buy weakness. The four trend and breakout templates lost money or came close to it, and the trend plus trailing stop lost 11.05% a year with a 68.57% max drawdown. That is a clear split, and the profile data below shows what in this fund separates the two groups.
The headline run has no fees or slippage. Cost runs at 5 and 10 basis points change the ranking at the top, and they are covered in the trading section.
| Strategy | CAGR | max DD | Sharpe | trades | win rate | final value |
|---|---|---|---|---|---|---|
| Drawdown Dip Buyer + 8% Target | 24.1% | −65.3% | 0.69 | 18 | 100% (+1 open) | $34,627 |
| RSI(14) Mean Reversion | 22.3% | −50.1% | 0.71 | 20 | 85% (+1 open) | $31,705 |
| RSI(2) Dip Snapback | 18.7% | −43.4% | 0.65 | 159 | 62% (+1 open) | $26,703 |
| First-to-Last Day of Month | 17.0% | −66.7% | 0.58 | 69 | 57% (+1 open) | $24,611 |
| Weekly Entry + 7% Target | 13.7% | −53.7% | 0.54 | 232 | 42% | $20,936 |
| EMA 12/26 Trend | 8.7% | −59.8% | 0.42 | 22 | 36% | $16,140 |
| SMA 10/50 Trend | 4.6% | −57.0% | 0.30 | 19 | 58% | $12,956 |
| 20-Day Momentum + Trailing Stop | 3.9% | −45.4% | 0.28 | 45 | 42% | $12,419 |
| 3-Month Momentum Switch | −2.5% | −58.1% | 0.09 | 25 | 40% | $8,665 |
| 200-Day SMA Regime Filter | −2.7% | −54.8% | 0.08 | 31 | 29% | $8,525 |
| Golden Cross (SMA 50/200) | −5.5% | −49.0% | 0.04 | 5 | 20% (+1 open) | $7,233 |
| SMA-200 Trend + 15% Trailing Stop | −11.1% | −68.6% | -0.18 | 28 | 39% | $5,105 |
| Buy & hold FAS | 18.3% | −66.3% | 0.60 | – | – | $26,295 |
Three winners, all of them buying weakness
The dip buyer is the top row, and its trade log is the most unusual in the set. It made 18 round trips and won all 18, with an average win of 8.19%. Thirteen of the 18 closed in 2021 alone, for a year return of 156.8% against 122.8% held. After that it made 1 trade in 2022, 1 in 2024, 2 in 2025 and 1 in 2026. The 2022 trade is the one that matters for risk. It entered on 2022-01-10 at an adjusted 131.74 and did not exit until 2024-11-06 at 145.11, a 1031 day hold that returned 10.15%. During that hold the account sat through the 65.26% max drawdown. The strategy has no stop, so a win rate of 100% and a max drawdown of 65.26% are both true at once.
The position that was open at the end of the window had entered on 2026-09-17 at 163.20 and was 13.64% under water. Exposure was 86.7% of days. The dip buyer is invested almost all the time on this fund because FAS had 55 falls of 10% or more from a 20-day high, spread over 434 days. The 8% target is reached quickly in a high-volatility fund, which is why 2021 produced so many trades.
RSI(14) mean reversion is the more varied result. It made 22.26% with 20 round trips, 17 of them winners, for 85%. The average win was 14.14% and the average loss was -24.66%, and the profit factor was 2.40. The best trade, from 2025-04-07 to 2025-04-30, returned 37.75% from an adjusted 87.44 to 120.45. The worst, from 2022-04-12 to 2022-08-04, lost 31.35%. Its max drawdown was 50.07%, which is shallower than the fund's 66.26%, and it was invested on 37.9% of days. The yearly returns were 75.7% in 2021, 9.3% in 2022, -10.5% in 2023, 30.9% in 2024, 63.8% in 2025 and -14.0% in 2026. It made money in 2022 when FAS lost 42.7% held, which is the main reason it sits above the fund.
The RSI(2) snapback made 18.66% from 159 round trips with a 62% win rate and a 43.37% max drawdown. Its edge over the fund is small, 0.32 points of CAGR, and its drawdown is shallower by a wide margin. It beat holding in 2022, 2025 and 2026 and trailed in 2021 and 2024, the two years the fund more than doubled or nearly did. The worst trades were -25.61% from 2023-03-08 to 2023-03-22 and -20.29% from 2022-09-15 to 2022-09-29. The average loss of -4.28% is larger than the average win of 3.91%, so the 62% win rate carried it, with a profit factor of 1.48.
The monthly cycle made 16.98% and trailed the fund by a little, with 95.2% exposure and a 66.69% max drawdown, effectively the same drawdown as holding. The weekly 7% target made 13.73% with 232 round trips and a 42% win rate, and spent 63.6% of days invested. Its average hold was 4.7 days. It won in 2022, 2023 and 2025 and lost badly to the fund in 2021, when it made 46.6% against 122.8%.
The trend group is where the leverage hurts. The EMA 12/26 made 8.70% and the SMA 10/50 made 4.61%. The momentum breakout made 3.85%. The 3-month momentum switch lost 2.46% and the 200-day regime filter lost 2.74%. The golden cross lost 5.49% in 5 round trips, with one winner of 75.16% from 2023-12-14 to 2025-05-06 and four losers. The 200-day filter shows the same shape on a larger scale: its single best trade returned 123.52% from 2023-11-15 to 2025-03-12, and it still finished at a loss because only 29% of its 31 round trips won.
How each strategy traded FAS
| Strategy | Time in market | Avg hold (days) | Best trade | Worst trade | Profit factor | With 10 bps slippage |
|---|---|---|---|---|---|---|
| dip buyer | 86.7% | 100 | 10.2% | 8.0% | – | 24.1% |
| RSI mean reversion | 37.9% | 39 | 37.8% | −31.4% | 2.40 | 21.4% |
| RSI(2) snapback | 37.5% | 5 | 26.9% | −25.6% | 1.48 | 12.3% |
| monthly cycle | 95.2% | 28 | 39.9% | −31.1% | 1.22 | 14.3% |
| weekly 7% target | 63.6% | 5 | 16.1% | −17.8% | 1.12 | 4.0% |
| EMA 12/26 trend | 57.3% | 55 | 59.0% | −19.6% | 1.62 | 7.9% |
| SMA 10/50 trend | 56.8% | 63 | 47.7% | −27.2% | 1.34 | 3.9% |
| momentum breakout | 46.9% | 22 | 55.8% | −10.3% | 1.14 | 2.3% |
| 3-month momentum | 56.0% | 47 | 54.2% | −21.4% | 0.89 | −3.2% |
| 200-day regime filter | 51.3% | 35 | 123.5% | −22.4% | 0.83 | −3.7% |
| golden cross | 51.7% | 203 | 75.2% | −29.7% | 0.71 | −5.6% |
| trend + trailing stop | 53.9% | 40 | 49.4% | −14.5% | 0.58 | −11.8% |
How costs and hold times change the order
The dip buyer and the RSI(14) template are the least sensitive to costs among the leaders. The dip buyer goes from 24.15% to 24.14% at 5 basis points and 24.13% at 10, with 37 fills. RSI(14) goes from 22.26% to 21.85% and 21.42%, with 41 fills. Both stay well above the fund's 18.34% in the cost runs.
The RSI(2) snapback is the opposite case. It makes 319 fills, and it falls from 18.66% to 15.46% at 5 basis points and to 12.30% at 10. At either cost level it is below buy-and-hold, and at 10 basis points it ends at $19,468 against $26,295. The 0.32 point lead over the fund in the headline run is smaller than the cost of one assumption about slippage. The test uses fixed costs of 5 and 10 basis points, and real execution on a fund that trades about $99,675,304 a day and a median of 1,451 shares per minute could be better or worse than that. FAS is thinner than the broad index funds, and a thin fund is where a cost assumption matters most.
The weekly template shows the sharpest decline. From 13.73% it drops to 10.96% at 5 basis points and 4.02% at 10, and its max drawdown rises from 53.69% to 56.21%. With 464 fills, it pays costs more than any other template here. The monthly cycle goes from 16.98% to 15.64% and then 14.33%.
Hold times show why. The dip buyer averaged 100.3 days with a median of 11, because the 2022 trade lasted 1031 days and skewed the average while most trades closed in a few days. RSI(2) averaged 5.0 days with a longest hold of 16. The weekly template averaged 4.7 days. The trend rules held for 34.6 to 62.7 days on average, the golden cross for 202.8.
The worst single trades are an argument for caution about any of these numbers. The weekly template lost 17.75% in 3 days from 2023-03-06, and 17.37% in 3 days from 2025-03-03. Those are the March 2023 and March 2025 drops. The fund's worst days were -21.93% on 2025-04-04 and -14.98% on 2025-04-03, and its best was 22.51% on 2025-04-09. A rule that is invested through those three sessions takes the whole swing, and that is true of every template that was invested on those dates.
The account-level risk is different from the per-trade numbers. A 100% win rate on the dip buyer says nothing about the 65.26% drawdown. Every template with a drawdown under the fund's has a smaller one because it was out of the market for part of the fall, not because it limited losses inside the fall.
What the trend and breakout rules traded on FAS
Four trend rules entered FAS within days of each other in November 2023, and that one move produced each rule's best trade. The 200-day filter entered on 2023-11-15 at 57.83 and left on 2025-03-12 at 129.26 for 123.52%. The trend plus trailing stop entered the same day and exited on 2024-04-12 at 86.38 for 49.36%. The EMA 12/26 rule entered on 2023-11-10 at 53.41 and exited on 2024-04-16 at 84.95 for 59.05%. The momentum breakout rule entered on 2023-11-16 at 58.7 and exited on 2024-04-10 at 91.46 for 55.81%. The only difference was the exit rule. The 15% trailing stop and the EMA cross sold in April 2024, and the 200-day filter held for 483 days.
The same rules then had to survive 2022. EMA 12/26 made 5 round trips in 2022 and won none. Its worst trade lost 19.56% in 20 days from 2022-01-04 at 121.27, and the next lost 14.13% from 2022-02-09. The trailing stop won 2 of 7 in 2022 and 0 of 4 in 2023. The 200-day filter won 2 of 9 in 2022, 1 of 5 in 2023 and 3 of 11 in 2025. Its median trade lost 1.77% and its median hold was 6 days, which shows how many of its 31 trades were short whipsaws around one long hold. The fund crossed its 200-day average 61 times, with annualized volatility of 54.99%.
The trailing stop shows the risk on the open end. On 2026-09-17 it entered at 163.2 and exited on 2026-10-01 at 139.48, a 14.53% loss in 14 days. The template's worst month was September 2026 at negative 22.65%, and the trailing stop template finished 2026 with a return of negative 25.4%. The result of negative 11.05% a year comes from 28 round trips with a 39% win rate and a median trade of negative 6.32%.
The weekly 7% target made 232 round trips, held 4.7 days on average and won 42%. Its best trade gained 16.14% in 2 days from 2024-11-04 at 124.94, a trade that included the fund's 18.42% day on 2024-11-06, and its worst lost 17.75% in 3 days from 2023-03-06. Its win count fell from 20 of 45 in 2021 and 20 of 48 in 2022 to 8 of 29 in 2026, and the 2026 return was negative 19.3%. At 10 basis points its CAGR fell from 13.73% to 4.02%, the largest fall of any template.
Momentum breakout made 45 round trips and won 42%, with a median trade of negative 2.1%. Only the 55.81% November 2023 trade and a 25.57% trade from 2021-03-22 did most of the work. Two short losses, 10.32% over 2 days from 2023-01-17 and 10.25% over 3 days from 2024-08-02, show a breakout entry on a 3x fund can reverse almost at once. Its CAGR of 3.85% is the sum of those pieces, with most of the winners in 2021 and 2023 and the losses spread across the rest.
How FAS behaved
| Measure | FAS |
|---|---|
| Data in this test | 2021-01-04 to 2026-10-02 (1444 sessions) |
| Total return, buy and hold | 180.7% |
| Annualized volatility | 55.0% |
| Deepest drawdown | −66.9% (2022-01-12 to 2023-10-27) |
| Up days | 52.7% |
| Average daily range | 4.15% |
| Average overnight gap | 1.41% |
| Correlation to SPY | 0.78 |
| Correlation to QQQ | 0.60 |
| Correlation to TLT | -0.04 |
| Sessions above the 200-day average | 59.5% |
| Crossings of the 200-day average | 61 |
| Falls of 10% or more from a 20-day high | 55 |
Calendar years
| Year | Return |
|---|---|
| 2021 | 125.7% |
| 2022 | −43.1% |
| 2023 | 14.8% |
| 2024 | 84.2% |
| 2025 | 21.6% |
| 2026 (part) | −15.0% |
Biggest single days
| Best day | Move |
|---|---|
| 2025-04-09 | 22.5% |
| 2024-11-06 | 18.4% |
| 2022-11-10 | 15.4% |
| 2021-01-06 | 12.8% |
| 2022-10-13 | 12.3% |
| Worst day | Move |
|---|---|
| 2025-04-04 | −21.9% |
| 2025-04-03 | −15.0% |
| 2023-03-09 | −12.2% |
| 2023-03-13 | −11.9% |
| 2022-03-07 | −11.5% |
Average return by calendar month
| Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 7.3% | 4.0% | −4.4% | −1.7% | 3.8% | −0.6% | 11.4% | 3.6% | −10.8% | 9.7% | 14.8% | −0.2% |
Most and least correlated funds
| Most correlated | Least correlated | ||
|---|---|---|---|
| XLF | 1.00 | SH | -0.78 |
| VOOV | 0.89 | SDS | -0.78 |
| VTV | 0.89 | SPDN | -0.78 |
| SSO | 0.78 | UVXY | -0.65 |
| VOO | 0.78 | VXZ | -0.63 |
Leverage and decay against XLF
| Year | FAS | XLF | 3× XLF, reset daily |
|---|---|---|---|
| 2021 | 125.7% | 36.6% | 129.3% |
| 2022 | −43.1% | −10.6% | −40.2% |
| 2023 | 14.8% | 12.1% | 30.1% |
| 2024 | 84.2% | 30.4% | 109.4% |
| 2025 | 21.6% | 14.9% | 35.6% |
| 2026 | −15.0% | −1.1% | −8.2% |
The last column compounds 3× XLF's daily return with no fees or financing: the return a perfect daily-reset fund would have had. It leaves out the fund's fees, its borrowing costs and the interest it earns on cash, which put a real fund below or above this line. A daily-reset fund does not aim for 3× the underlying's return over a year.
Over 2021-01-04 to 2026-10-02, FAS returned 180.7% while XLF returned 102.9% and a perfect daily-reset 3× version would have returned 364.7%. Its measured daily beta to XLF was 2.98.
What 3x leverage did to the path
FAS is a daily 3x fund. Over the whole window its realized beta to XLF was 2.98 with an R-squared of 1.00, so it tracked three times the daily move almost exactly. The underlying returned 102.88%. A frictionless daily-reset 3x version of XLF would have returned 364.74% in the same data, and the fund returned 180.7%, well below that line. The ideal leaves out fees, borrowing costs and interest on cash, which put a real fund below or above it. The table shows the comparison by year.
In 2021 the fund returned 125.73% against 129.32% for the daily-reset ideal, so leverage helped in a steady uptrend. In 2022 it returned -43.13% against -40.24%. In 2023 it returned 14.83% against 30.09% in a year when the underlying made 12.11%. In 2025 the fund returned 21.65% against 35.56%. In 2026 to date it returned -15.01% while XLF returned -1.15%, against -8.17% for the ideal. The fund finished below the ideal in every year, with the widest shortfalls in 2023, 2024 and 2025, while 2021 and 2022 were close. The pattern in this data is that the fund tracked the ideal closely in 2021 and 2022 and fell further behind it in the later years.
The volatility is the other half. Annualized volatility was 54.99% and the average daily range was 4.15%. Up days were 52.67% of sessions with an average gain of 2.53%, and down days averaged -2.55%. The deepest drawdown was 66.86% from 2022-01-12 to 2023-10-27, recovered by 2024-10-16, and the longest drawdown lasted 692 sessions. The overnight share of return was 91.85%, so almost all of the gain came from the gap between the close and next open and very little from intraday movement.
The RSI statistics explain the leaderboard. After RSI(14) fell under 30, which happened on 33 sessions, the median 20-day forward return was 10.06% against a baseline of 3.14%. After RSI(2) fell under 10, on 174 sessions, the median 5-day forward return was 0.57% against a baseline of 0.93%, and the 20-day figure was 3.95% against 3.14%. The slower oversold signal had a clear edge and the faster one had almost none, which fits RSI(14) mean reversion ending ahead of RSI(2) with a smaller drawdown. FAS rewards patience after a deep sell-off. The lag-1 autocorrelation was 0.02, so there was no persistent daily direction to follow.
The 200-day average was crossed 61 times, and the fund spent 59.52% of sessions above it. The trend templates were whipsawed. Every trend rule had a worst trade of at least -14% and lost most of its round trips. Winners such as the 59.05% EMA trade from 2023-11-10 to 2024-04-16 were too few.
FAS has a correlation of 1.00 with XLF, 0.89 with VOOV and VTV, and 0.78 with SSO and VOO. Beta to SPY was 2.62 and to QQQ 1.47. Correlation to TLT was -0.04. Among the other leveraged funds, TQQQ held at 25.35% and SOXL at 33.32%, and both of those also had RSI(2) or trend templates at the top. On QLD and TQQQ the RSI(2) snapback was the best template. On FAS it was third. The two leveraged Treasury funds, TMF at -31.16% and UST at -8.51%, show what the same leverage does without a rising underlying.
Seasonal averages have 5 or 6 observations per month. November averaged 14.82% and July 11.45%, September -10.75% and March -4.42%. They describe six years and are not a pattern. All results are for one window with no fees in the headline run.
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 FAS?
Of the 12 templates tested on FAS over 2021-01-04 to 2026-10-02, the strongest by CAGR was dip buyer at 24.1% (max drawdown 65.3%), versus 18.3% 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 FAS?
3 of 12 templates beat FAS buy-and-hold (18.3% CAGR) on this window; 10 of 12 had a shallower maximum drawdown than holding (66.3%).
What was the best strategy for FAS?
The drawdown dip buyer made 24.15% a year with 18 winning round trips, against 18.34% for buy-and-hold. Its max drawdown was 65.26%, close to the fund's 66.26%. RSI(14) mean reversion made 22.26% with a shallower 50.07% drawdown.
Did any strategy beat buying and holding FAS?
Three of twelve did on CAGR: the dip buyer, RSI(14) mean reversion and the RSI(2) snapback at 18.66%. Ten had a shallower max drawdown than holding and eight had a positive CAGR. With 10 basis points of slippage the RSI(2) snapback fell to 12.30%.
Why do trend rules lose money on FAS?
The fund crossed its 200-day average 61 times, and a daily 3x fund moves hard against a late entry. The golden cross lost 5.49% a year, the 200-day filter lost 2.74% and the trend plus trailing stop lost 11.05%.
How much does leverage decay show up in FAS?
Over the window FAS returned 180.7% while XLF returned 102.88%. A frictionless daily-reset 3x version of XLF would have returned 364.74%, so the fund came in below that line. In 2023 the fund returned 14.83% against 30.09% for the ideal. The ideal leaves out fees, borrowing costs and interest on cash.
How volatile is FAS?
Annualized volatility was 54.99% and the average daily range was 4.15%. The best day was 22.51% on 2025-04-09 and the worst was -21.93% on 2025-04-04. The deepest drawdown was 66.86%.
Is the dip buyer's 100% win rate on FAS safe?
The win rate counts closed trades only. The dip buyer closed 18 trades at a profit and held one for 1031 days through a 65.26% drawdown. A trade opened on 2026-09-17 was still open and 13.64% under water at the end of the data.
Other leveraged 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.