DeployQuantUSDU trading strategies, backtested
WisdomTree Bloomberg U.S. Dollar Bullish Fund: a long-dollar position against a broad currency basket. Every DeployQuant template run on USDU over 5.7 years of minute data, same engine, same window, sorted by return.
USDU is the WisdomTree Bloomberg U.S. Dollar Bullish Fund, a long-dollar position against a broad currency basket. Holding it from 2021-01-04 to 2026-10-02 made 5.2% a year, a total of 33.9%, and turned $10,000 into $13,394. The maximum drawdown was 9.0%, from the peak on 2022-09-27 to the low on 2023-02-01, and the Sharpe ratio was 0.84. The fund reached that with annualized volatility of 6.4%.
All 12 DeployQuant templates ran on USDU. One of them, the monthly cycle at 5.4% a year, finished ahead of holding on return. Ten had a shallower drawdown and ten made money. The templates that traded the most gave up the most, and the dip buyer and momentum breakout never traded, which leaves them at $10,000 and 0%.
The point of the page is the size of the gap. A fund that rises steadily in a narrow band leaves little for an entry or exit rule to improve, and each exit costs the position's drift for the days it is away. The headline runs carry no fees or slippage. The cost runs matter more here than on most funds, because USDU trades about $6,125,205 a day and the median minute bar has 400 shares. Other currency funds in the test are FXE and UDN, which moved in the opposite direction to USDU.
| Strategy | CAGR | max DD | Sharpe | trades | win rate | final value |
|---|---|---|---|---|---|---|
| First-to-Last Day of Month | 5.4% | −7.6% | 0.87 | 69 | 51% (+1 open) | $13,524 |
| Weekly Entry + 7% Target | 5.1% | −11.0% | 0.85 | 49 | 14% (+1 open) | $13,305 |
| SMA-200 Trend + 15% Trailing Stop | 4.7% | −8.9% | 0.81 | 0 | – (+1 open) | $12,992 |
| Golden Cross (SMA 50/200) | 3.7% | −8.3% | 0.71 | 2 | 100% (+1 open) | $12,323 |
| SMA 10/50 Trend | 3.5% | −7.4% | 0.74 | 15 | 53% (+1 open) | $12,203 |
| RSI(14) Mean Reversion | 3.4% | −5.2% | 0.92 | 16 | 88% | $12,124 |
| EMA 12/26 Trend | 3.2% | −4.3% | 0.69 | 23 | 52% (+1 open) | $11,962 |
| RSI(2) Dip Snapback | 3.1% | −3.3% | 0.79 | 172 | 56% | $11,925 |
| 200-Day SMA Regime Filter | 2.7% | −11.2% | 0.54 | 17 | 29% (+1 open) | $11,657 |
| 3-Month Momentum Switch | 0.5% | −6.5% | 0.16 | 3 | 33% | $10,284 |
| Drawdown Dip Buyer + 8% Target | 0.0% | −0.0% | 0.00 | 0 | – | $10,000 |
| 20-Day Momentum + Trailing Stop | 0.0% | −0.0% | 0.00 | 0 | – | $10,000 |
| Buy & hold USDU | 5.2% | −9.0% | 0.84 | – | – | $13,394 |
How the 12 templates ranked on USDU
The monthly cycle made 5.4% with a 7.6% drawdown and 69 round trips. It ended at $13,524, which is $130 above holding. It is invested 95.2% of the time, so the small gain comes from the days it skips rather than from a signal. The win rate was 51% but the average win of 1.87% was almost twice the average loss of 0.99%, giving a profit factor of 1.87. It beat holding in 2021, 2022 and 2025, and trailed in 2023, 2024 and 2026.
The weekly 7% target made 5.1% with a deeper drawdown of 11.0% and 49 round trips. Its win rate was 14%, with an average win of 7.19% and an average loss of 0.54%. A 7% target takes weeks to reach on a fund with a daily range of 0.54%: the best winners held for 79, 160, 92, 351 and 78 days. A longest losing streak of 13 trades is a feature of the rule. Its best year was 2024 at 15.8%.
The trend with a trailing stop made 4.7% with a drawdown of 9.0% and no closed trades. It entered on 2021-10-19 at an adjusted price of 20.72 and was still in the position at the end of the test, with a 30.6% gain on the trade. The rule's result is the fund's own return from that date with no exits. The golden cross made 3.7% on 2 closed trades, both winners, holding 458 and 659 days, and was still in a third position at the end. The SMA 10/50 trend made 3.5% over 15 round trips with a 53% win rate and a profit factor of 4.41.
The RSI(14) mean reversion made 3.4% with a 5.2% drawdown and a 88% win rate across 16 trades. Its profit factor of 21.37 reflects an average win of 1.49% against an average loss of 0.49%, and it made money in every calendar year, from 0.4% in 2025 to 6.8% in 2026. It was invested only 36.8% of the time. The EMA 12/26 trend made 3.2% with a 4.3% drawdown and 23 round trips, and the RSI(2) snapback made 3.1% with a 3.3% drawdown and 172 round trips. The 200-day regime filter made 2.7% with an 11.2% drawdown and a 29% win rate.
The 3-month momentum switch made 0.5% on 3 trades and was invested 18.3% of the time. The dip buyer and momentum breakout made no trades. Both reach 0% and show a drawdown of 0%, but a rule that never enters has not been tested.
The ordering is close to the order of time in the market. The two rules that stayed in the fund for nearly all of the window made the most. The rules that sat in cash for most of it made the least. On a fund with a steady positive drift, every day out of the position costs something, and no rule found enough value in the days it avoided to offset that cost.
Years and drawdowns across the templates
Holding made 4.1% in 2021, 7.4% in 2022, 3.0% in 2023, 14.3% in 2024, lost 3.1% in 2025 and made 4.8% in 2026. The templates that beat holding in a given year are mostly the ones that were out of the fund when it fell. In 2025 the RSI(2) snapback made 7.4% against a loss of 3.1% for holding, the weekly 7% target lost 2.6%, the RSI(14) mean reversion made 0.4% and the SMA 10/50 trend made 0.3%. The golden cross and 200-day regime filter lost 5.3% and 4.9% that year, deeper than holding, because they were still holding when the fund fell.
In 2024, the best year, holding made 14.3% and most rules trailed. The weekly 7% target made 15.8% and the golden cross matched the fund at 14.3%, while the RSI(2) snapback made only 1.1%, the RSI(14) mean reversion 4.6% and the 3-month momentum switch 2.5%. A rule that spends most of a strong year in cash gives up a large part of the move, and 2024 shows that more clearly than any other year.
The drawdown tables show the long recoveries. The weekly 7% target had an 11.0% drawdown beginning 2022-10-14 and recovered on 2024-06-07. The 200-day regime filter had 11.2% from 2022-09-27 and recovered on 2024-11-13, and a second one of 8.6% that had not recovered at the end of the test. The golden cross had 8.3% twice. Holding itself recovered from its first drawdown on 2023-10-31. For these rules, drawdowns were deeper than holding because each held through the decline and then missed part of the rebound.
The lowest drawdowns belong to the rules that were out of the market for most days. The RSI(2) snapback had 3.3% and the RSI(14) mean reversion 5.2%. The EMA 12/26 trend had 4.3%, close to the 4.3% of the third drawdown in holding. All of them are ahead of holding on risk and behind it on return.
How each strategy traded USDU
| Strategy | Time in market | Avg hold (days) | Best trade | Worst trade | Profit factor | With 10 bps slippage |
|---|---|---|---|---|---|---|
| monthly cycle | 95.2% | 28 | 4.3% | −3.8% | 1.87 | 2.9% |
| weekly 7% target | 92.5% | 38 | 7.5% | −2.5% | 2.08 | 2.8% |
| trend + trailing stop | 86.1% | – | – | – | – | 4.7% |
| golden cross | 68.6% | 559 | 10.1% | 7.6% | – | 3.6% |
| SMA 10/50 trend | 62.1% | 86 | 8.7% | −2.1% | 4.41 | 3.0% |
| RSI mean reversion | 36.8% | 49 | 3.3% | −0.7% | 21.37 | 2.9% |
| EMA 12/26 trend | 62.3% | 56 | 5.8% | −1.8% | 3.12 | 2.4% |
| RSI(2) snapback | 37.7% | 5 | 2.5% | −2.5% | 1.61 | −2.8% |
| 200-day regime filter | 68.0% | 71 | 9.3% | −1.9% | 2.16 | 2.1% |
| 3-month momentum | 18.3% | 129 | 4.8% | −1.5% | 2.51 | 0.4% |
| dip buyer | 0.0% | – | – | – | – | 0.0% |
| momentum breakout | 0.0% | – | – | – | – | 0.0% |
Costs, hold times and the RSI(2) result
The RSI(2) snapback is the clearest case for reading the cost columns. It made 3.1% a year in the headline run with a 3.3% drawdown. Add 5 basis points per trade and it made 0.13%. Add 10 basis points and it lost 2.76%, with a drawdown of 15.5%. With 172 round trips and an average win of 0.50%, the trade edge is about the size of the cost. The average hold was 4.7 days and the profit factor was 1.61. Its best trade gained 2.5% in 4 days from 2023-02-02, and its worst lost 2.5% in 10 days from 2022-11-08.
The monthly cycle held up better, going from 5.4% to 4.2% at 5 basis points and 2.9% at 10. The weekly 7% target went from 5.1% to 3.8% and 2.8%. The drawdown on the weekly rule rose from 11.0% to 14.7%. The slow rules were barely affected. The golden cross went from 3.7% to 3.7% and 3.6%, the SMA 10/50 trend from 3.5% to 3.3% and 3.0%, and the RSI(14) mean reversion from 3.4% to 3.1% and 2.9%. At 10 basis points the RSI(14) mean reversion is ahead of the RSI(2) snapback by a wide margin.
The holding times show why. The golden cross held for 558.5 days on average, the weekly 7% target for 37.9 days, the SMA 10/50 trend for 86.1 days, and the RSI(14) mean reversion for 48.6 days. The RSI(2) snapback held for under five days and the monthly cycle for 28.4. A rule that trades every few days pays costs on every swing, and on a fund where the swing is a fraction of a percent, the costs consume it.
The cost assumption is a flat figure per trade, and USDU is a thin fund. The average daily dollar volume is $6,125,205 and the median minute bar has 400 shares, so a test account of $10,000 is fine but a large order would move the price. The cost runs on a flat figure do not capture that. The RSI(2) snapback on USDU page has the full trade list.
The long holds and the months that moved the results
The slow templates each rest on one or two long trades. The golden cross made 10.09% from 2023-08-02 to 2025-05-22, and its other closed trade made 7.63% from 2021-10-19 to 2023-01-20. The 200-day filter made 9.31% from 2023-12-29 to 2025-04-14 and 9.17% from 2021-10-19 to 2022-12-14. The SMA 10/50 made 8.66% from 2022-03-03 to 2022-11-14, and the EMA 12/26 made 5.79% from 2024-10-04 to 2025-02-18, ahead of its 4.12% trade in the autumn of 2022.
The golden cross and the 200-day filter are both holding again at the end of the data. The golden cross entered on 2025-11-13 at 25.9 and was up 4.48%. The 200-day filter entered on 2026-02-19 at 25.68 and was up 5.37%. The trailing-stop template has not exited since its entry on 2021-10-19 at 20.72, which is why its yearly figures match holding's in 2022, 2023, 2024, 2025 and 2026.
The months that moved holding were April 2022 at 3.77% and November 2022 at -4.38%. Most of the slower rules show the same two figures as their best and worst months, because they were invested in both. The monthly cycle made 4.22% in November 2024 and lost 3.73% in April 2025. Its three worst trades were whole months: April 2025 at -3.81%, November 2022 at -3.48% and November 2023 at -2.43%. Two of those three are Novembers, and the calendar-month averages bear that out, with November at -0.7% and December at -0.28%.
The RSI(14) mean reversion rule had the cleanest record of any template that traded. It won 88% of its 16 round trips with a profit factor of 21.37 and a 5.18% drawdown, and its largest loss was 0.74% from 2022-11-14 to 2023-02-23. Its best trade made 3.25% from 2022-06-01 to 2022-06-14. At 3.41% a year it earned less than holding because it was invested 36.8% of the time. The fund's RSI(14) fell below 30 on only 8 sessions, with a median 20-day forward return of 1.14% against a 0.32% baseline, so the sample behind the rule is small.
How USDU behaved
| Measure | USDU |
|---|---|
| Data in this test | 2021-01-04 to 2026-10-02 (1444 sessions) |
| Total return, buy and hold | 34.4% |
| Annualized volatility | 6.4% |
| Deepest drawdown | −9.1% (2022-09-27 to 2023-02-01) |
| Up days | 50.9% |
| Average daily range | 0.54% |
| Average overnight gap | 0.26% |
| Correlation to SPY | -0.34 |
| Correlation to QQQ | -0.30 |
| Correlation to TLT | -0.24 |
| Sessions above the 200-day average | 79.0% |
| Crossings of the 200-day average | 34 |
| Falls of 10% or more from a 20-day high | 0 |
Calendar years
| Year | Return |
|---|---|
| 2021 | 4.2% |
| 2022 | 7.5% |
| 2023 | 3.1% |
| 2024 | 14.6% |
| 2025 | −3.2% |
| 2026 (part) | 4.9% |
Biggest single days
| Best day | Move |
|---|---|
| 2022-09-13 | 1.5% |
| 2022-09-23 | 1.4% |
| 2024-11-06 | 1.3% |
| 2025-05-12 | 1.3% |
| 2023-02-03 | 1.2% |
| Worst day | Move |
|---|---|
| 2022-11-10 | −2.1% |
| 2025-04-10 | −1.8% |
| 2022-11-04 | −1.7% |
| 2025-04-03 | −1.5% |
| 2022-08-10 | −1.4% |
Average return by calendar month
| Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.2% | 0.7% | 0.5% | −0.1% | −0.1% | 1.1% | 0.3% | 0.4% | 1.6% | 1.4% | −0.7% | −0.3% |
Most and least correlated funds
| Most correlated | Least correlated | ||
|---|---|---|---|
| EEV | 0.51 | UDN | -0.90 |
| SDS | 0.34 | FXE | -0.86 |
| SH | 0.34 | EEM | -0.51 |
| SPDN | 0.34 | QAI | -0.45 |
| SQQQ | 0.30 | IAU | -0.44 |
What USDU's own behaviour explains
USDU's daily moves are small and balanced. Up days were 50.9% of sessions, the average up day was 0.32% and the average down day was negative 0.32%. The average daily range was 0.54%. The lag-1 autocorrelation was negative 0.04. Returns arrived at the open and during the session in similar amounts: the overnight share of the log return was 43.7% and the intraday share was 56.3%.
The calendar years were 4.2% in 2021, 7.5% in 2022, 3.1% in 2023, 14.6% in 2024, negative 3.2% in 2025 and 4.9% for the partial year 2026. The fund rose in five of six years. It had no year of loss like the equity funds had in 2022, which means a rule that exits in a decline had less to avoid. The 2022 figure was a gain, and the rules that protected capital elsewhere gave up some of that gain: the 3-month momentum switch made 4.7% in 2022 against 7.4% for holding, and the RSI(2) snapback made 2.5%.
The fund had three drawdowns of note. The first, 9.0%, ran from 2022-09-27 to 2023-02-01 and recovered on 2023-10-31. The second, 7.6%, began 2025-01-13, reached its trough on 2025-07-01 and recovered on 2026-09-29, which is 455 days of recovery. The third was 4.3% in late 2023. The longest drawdown lasted 428 sessions. The fund never fell 10% from a 20-day high.
Mean reversion signals had a small edge. After RSI(2) closed below 10, on 125 sessions, the median 5-day return was 0.19% against a baseline of 0.09%, and the median 20-day return was 0.33% against 0.32%. After the 14-day RSI closed below 30, on 8 sessions, the median 5-day return was 0.90% and the 20-day return was 1.14%. With only 8 such sessions, the RSI(14) mean reversion made 16 trades on a different signal, and won 14 of them. The 5-day edge over the baseline for RSI(2) is real but small, and it is smaller than the cost of trading it.
USDU was above its 200-day average on 79.0% of sessions and crossed it 34 times. That is a high crossing count for a fund with this little volatility, and the 200-day regime filter with a 29% win rate shows it. The regime filter held for 70.9 days on average, and its best trades, 9.31% over 472 days and 9.17% over 421 days, carried the whole result.
The correlation to SPY was negative 0.34 and to QQQ negative 0.30, with a beta to SPY of negative 0.13. The fund is most correlated with EEV at 0.51 and least with UDN at negative 0.90 and FXE at negative 0.86. By calendar month, September averaged 1.63% and October 1.37%, while November averaged negative 0.70%. Each month has 5 or 6 observations. The best day was 2022-09-13 at 1.47% and the worst was 2022-11-10 at negative 2.07%.
Limits of the USDU test
The window is 5.74 years in which the dollar rose in most years. A window with a falling dollar would reverse the picture, and the funds that move opposite to USDU, FXE and UDN, show what that looks like: on both of those pages the best template was the trend with a trailing stop, at 2.1% on FXE and 1.9% on UDN, against holding results of negative 0.9% and negative 1.1%.
Rules that exit have a real chance to help only when holding loses money. On USDU holding made money in five of six years, and one rule out of 12 beat it, by $130. That result is within the noise of a single test. None of the figures is a forecast, and the headline runs carry no costs. The cost runs show that a fast rule on this fund needs a cost near zero to keep its return.
Overnight returns, weekdays and the months of the year
USDU's return split is nearly even between the two halves of the day. The overnight log return was 13.02% and the intraday log return was 16.76%. A rule that trades only inside the session gets a little more than half of the return, and one that holds overnight gets the rest.
The weekday averages are 0.03% on Monday, 0.02% on Tuesday and Wednesday, 0.01% on Thursday and 0.03% on Friday. They are effectively equal. The calendar-month averages are less even, with September at 1.63%, October at 1.37% and June at 1.05%, against November at -0.7%, December at -0.28% and April and May at -0.07%. With 5 or 6 observations per month those figures describe the window. The fund's best days were 2022-09-13 at 1.47% and 2022-09-23 at 1.35%.
The worst days were 2022-11-10 at -2.07%, 2025-04-10 at -1.8% and 2022-11-04 at -1.7%. The first and third fall in the month the fund lost most, November 2022. The fund's correlation to SPY was -0.34. The fund never had a fall of 10% or more from a 20-day high, so the dip buyer, which waits for one, made no trades.
The funds that moved against USDU were UDN at a correlation of -0.9 and FXE at -0.86. On those two funds holding lost 1.12% and 0.9% a year, and the trend plus trailing stop made 1.88% and 2.14%. On USDU the same template made 4.67% against 5.22% for holding. The rule that helped on the falling-dollar funds trailed slightly on the rising-dollar fund.
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 USDU?
Of the 12 templates tested on USDU over 2021-01-04 to 2026-10-02, the strongest by CAGR was monthly cycle at 5.4% (max drawdown 7.6%), versus 5.2% 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 USDU?
1 of 12 templates beat USDU buy-and-hold (5.2% CAGR) on this window; 10 of 12 had a shallower maximum drawdown than holding (9.0%).
Did any strategy beat buy-and-hold on USDU?
One did. The monthly cycle made 5.4% a year against 5.2% for holding from 2021-01-04 to 2026-10-02, a gap of $130 in final value. The other 11 templates finished below holding.
What was USDU's maximum drawdown?
Holding USDU fell 9.0% from 2022-09-27 to 2023-02-01 and recovered on 2023-10-31. A second drawdown of 7.6% ran from 2025-01-13 and took until 2026-09-29 to recover.
Does RSI(2) mean reversion work on USDU?
It made 3.1% a year with 172 trades and a 3.3% drawdown. At 5 basis points of cost the return fell to 0.13%, and at 10 basis points it lost 2.76%. Each trade's edge was close to the size of the cost.
Why did two strategies make no trades on USDU?
The dip buyer and momentum breakout found no entry in the window, so both show 0% and $10,000. USDU's daily range of 0.54% and its steady rise gave neither rule a trigger.
How volatile is USDU?
Annualized volatility was 6.4% and the average daily range was 0.54%. The best day was 1.47% on 2022-09-13 and the worst was negative 2.07% on 2022-11-10. The fund rose in five of six calendar years.
How does USDU relate to stocks?
Its correlation to SPY was negative 0.34 and to QQQ negative 0.30, with a beta to SPY of negative 0.13. It was the most correlated with EEV at 0.51 and least with UDN at negative 0.90.
Other currency 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.