DeployQuantEMA 12/26 Trend on QID
ProShares UltraShort QQQ: -2x daily Nasdaq-100. Backtest 2021-01-04 to 2026-10-02, $10,000 starting capital, computed by the same engine that runs live DeployQuant strategies.
Here the EMA 12/26 trend rule is applied to QID, ProShares UltraShort QQQ, a fund that aims for twice the inverse of each day's Nasdaq-100 move. The rule owns the fund only on days when the 12-day exponential average sits above the 26-day average. Starting with $10,000 on 2021-01-04, it finished on 2026-10-02 at $4,312. That is a CAGR of -13.63% and a total return of -56.88%. Holding QID the whole way left $1,348 (a CAGR of -29.46%). The rule lost money, and the loss was smaller than holding's.
The rule made 24 round trips and 4 of them won, a win rate of 17%. The average win was 9.24% and the average loss was -5.85%. The profit factor was 0.31, well under 1. The account was in the fund 29.2% of sessions, and trades lasted 25.8 days on average. The Sharpe ratio was -0.3 and the max drawdown was 61.64%, against 87.33% for buy-and-hold.
A trend signal needs a price that keeps going one way. QID did not do that over this window. It fell in 2021, rose 66.4% in 2022 as the Nasdaq-100 fell, and then fell again every year from 2023 to 2026. The crossover caught part of the 2022 rise and then took repeated small losses on rallies in a fund that was decaying. The rule ranks 7th of 12 strategies on QID by CAGR, and QID ranks 53rd of 59 funds for this rule. These are hypothetical results for a single 5.74-year window, and the headline run charges no fees.
Year by year
| Year | EMA 12/26 trend | buy & hold |
|---|---|---|
| 2021 | −26.1% | −45.2% |
| 2022 | 24.3% | 63.3% |
| 2023 | −24.5% | −55.6% |
| 2024 | −19.1% | −31.8% |
| 2025 | −10.0% | −31.6% |
| 2026 | −14.7% | −29.5% |
How each year looked
Year by year the rule made -26.1% (2021), 24.3% (2022), -24.5% (2023), -19.1% (2024), -10% (2025) and -14.7% (2026 so far). The fund itself did -45.2%, 63.3%, -55.6%, -31.8%, -31.6% and -29.5%. Holding was behind in five of the six years. The one year it led was 2022, by 39 points, and 2022 was also the only year QID rose.
The five years in which the rule beat holding are all years in which it lost. A rule that is out of the fund for most sessions loses less than the fund when the fund keeps falling. In 2023 the rule lost 24.5% while holding lost 55.6%, a gap of 31.1 points. In 2021 the gap was 19.1 points, in 2025 it was 21.6, and in 2026 it was 14.8. In 2024 the gap was 12.7. Those gaps come from time spent in cash, not from any skill in the signal.
2022 is the one test of whether the signal had value. QID gained 63.3% that year on a buy-and-hold basis, and the rule gained 24.3%, so it caught some of the move and missed most of it. The crossover only turns after the fund has already moved, so the entries came after part of each rise was done. The first 2022 trade was entered on 2022-01-11 at 70.35.
In a fund that decays, a rule can lose less than the fund without making money. In this window it did that in every falling year. In the single rising year it made money but less than holding.
Month by month
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2021 | 0.0% | 0.0% | −12.8% | 0.0% | −7.9% | 0.0% | 0.0% | 0.0% | 2.0% | −9.7% | 0.0% | 0.0% |
| 2022 | 6.9% | 6.7% | −6.2% | 14.1% | −1.0% | 7.7% | −9.5% | 0.0% | 24.4% | −9.4% | −7.5% | 1.3% |
| 2023 | −11.6% | 0.0% | −0.3% | 0.0% | 0.0% | 0.0% | 0.0% | −4.4% | 0.4% | 0.5% | −11.3% | 0.0% |
| 2024 | 0.0% | 0.0% | 0.0% | 2.9% | −6.8% | 0.0% | −3.9% | −4.6% | −8.0% | 0.0% | 0.0% | 0.0% |
| 2025 | −10.1% | 3.8% | 15.8% | −7.8% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% | −7.0% | −2.8% |
| 2026 | 0.0% | −2.6% | 9.7% | −10.1% | 0.0% | 0.0% | 3.9% | −9.5% | −5.5% | 0.0% | – | – |
Months with nothing and months with large moves
The monthly table has many zeros. In 2021 the rule was in cash for January, February, April, June, July, August, November and December. In 2023 it was in cash in February, April, May, June, July and December. In 2025 it was in cash from May to October. A 0 means the position was flat for the whole month. Each zero is a month in which the 12-day average stayed under the 26-day average and the rule did not buy.
The rule's best month was September 2022 at 24.36%, and its worst was March 2021 at -12.83%. For the fund, April 2022 was the best month at 28.51% and April 2026 the worst at -22.08%. The rule's best month belongs to the trade entered on 2022-09-02. In April 2022 the rule returned 14.11%, against 28.51% for holding, because its trade that month only started on 2022-04-19 at 81.25.
The negative months stack up. The rule lost 11.59% in January 2023 and 11.26% in November 2023. It lost 10.11% in January 2025 and 10.14% in April 2026. In each of them the crossover had turned positive after a bounce in QID, and the fund then fell back while the rule was still holding. January 2025 is an example: the trade started on 2025-01-13 at 31.35 and was closed on 2025-01-22 at 28.11, a loss of 10.33% in 9 days.
The March 2025 month was a gain of 15.79% and April 2025 was a loss of 7.79%. Both months sit inside one trade, which entered on 2025-02-27 and closed on 2025-04-29. That trade is the second best in the list at 11.05%. Monthly figures break one trade into pieces, so the sign of a month does not always match the sign of the trade it belongs to.
Every trade
EMA 12/26 trend on QID made 24 closed round trips, an average hold of 26 days, an average winner of 9.24%, an average loser of −5.85%, a profit factor of 0.31, a longest losing streak of 9. It held a position at the close on 29.2% of trading days.
| Entry | Entry price | Exit | Exit price | Return | Days held |
|---|---|---|---|---|---|
| 2021-03-04 | $116.53 | 2021-03-18 | $109.47 | −6.1% | 14 |
| 2021-03-19 | $112.51 | 2021-03-23 | $107.22 | −4.7% | 4 |
| 2021-03-25 | $114.15 | 2021-03-30 | $110.75 | −3.0% | 5 |
| 2021-05-13 | $104.71 | 2021-05-27 | $96.21 | −8.1% | 14 |
| 2021-09-29 | $80.27 | 2021-10-20 | $73.80 | −8.1% | 21 |
| 2022-01-11 | $70.35 | 2022-03-23 | $75.42 | 7.2% | 71 |
| 2022-04-19 | $81.25 | 2022-06-09 | $92.93 | 14.4% | 51 |
| 2022-06-10 | $100.57 | 2022-07-18 | $96.95 | −3.6% | 38 |
| 2022-09-02 | $90.06 | 2022-11-14 | $93.96 | 4.3% | 73 |
| 2022-12-20 | $105.45 | 2023-01-18 | $94.14 | −10.7% | 29 |
| 2023-03-14 | $86.37 | 2023-03-15 | $86.13 | −0.3% | 1 |
| 2023-08-15 | $55.54 | 2023-09-01 | $52.38 | −5.7% | 17 |
| 2023-09-21 | $58.13 | 2023-10-17 | $56.90 | −2.1% | 26 |
| 2023-10-20 | $59.14 | 2023-11-08 | $54.64 | −7.6% | 19 |
| 2024-04-16 | $41.82 | 2024-05-09 | $40.06 | −4.2% | 23 |
| 2024-07-26 | $36.88 | 2024-08-21 | $33.75 | −8.5% | 26 |
| 2024-09-09 | $38.15 | 2024-09-16 | $35.01 | −8.2% | 7 |
| 2025-01-13 | $31.35 | 2025-01-22 | $28.11 | −10.3% | 9 |
| 2025-02-27 | $29.42 | 2025-04-29 | $32.67 | 11.1% | 61 |
| 2025-11-19 | $20.74 | 2025-12-02 | $19.26 | −7.1% | 13 |
| 2025-12-19 | $19.70 | 2025-12-22 | $19.13 | −2.9% | 3 |
| 2026-02-05 | $20.68 | 2026-04-13 | $19.84 | −4.1% | 67 |
| 2026-07-20 | $14.65 | 2026-08-10 | $13.76 | −6.1% | 21 |
| 2026-09-16 | $14.40 | 2026-09-21 | $13.59 | −5.6% | 5 |
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
There are 24 round trips, and 4 were winners. The best was a 14.38% gain from 2022-04-19 at 81.25 to 2022-06-09 at 92.93, a hold of 51 days. The second was 11.05% from 2025-02-27 at 29.42 to 2025-04-29 at 32.67, held 61 days. The third was 7.21% from 2022-01-11 at 70.35 to 2022-03-23 at 75.42, held 71 days. The fourth was 4.33% from 2022-09-02 at 90.06 to 2022-11-14 at 93.96, held 73 days. Prices are adjusted for splits and dividends, so they are lower than the quotes printed at the time.
Three of the four winners were in 2022. The per-year table says the rule had 4 round trips that exited in 2022 and 3 of them won. In every other year the count of winners was 0 or 1. The rule exited 5 trades in 2021 with none winning, 5 in 2023 with none winning, 3 in 2024 with none, 4 in 2025 with 1 winning, and 3 in 2026 with none.
The losses are small and frequent. The worst trade was -10.73%, from 2022-12-20 at 105.45 to 2023-01-18 at 94.14, held 29 days. The second worst was -10.33% in January 2025. Then came -8.49% in July and August 2024, -8.23% in September 2024 and -8.12% in May 2021. The median trade lost 4.7% and the median hold was 21 days. The first trade, from 2021-03-04 at 116.53 to 2021-03-18 at 109.47, lost 6.06%. The last trade, from 2026-09-16 at 14.4 to 2026-09-21 at 13.59, lost 5.63% in 5 days.
The 2022-09-02 trade shows how a trend signal gives back a gain. September 2022 was the rule's best month at 24.36%. The trade was not closed until 2022-11-14, and by then the trade's total was 4.33%. The crossover held through the October and November declines, which cost 9.36% and 7.54% in those months. The signal exits after the average has turned, so a fund that reverses sharply takes back much of what it gave.
The shortest trade held 1 day, from 2023-03-14 to 2023-03-15, for -0.28%. The longest held 73 days. Most entries came during a bounce in a fund that kept sliding.
Largest drawdowns
| Peak | Low point | Depth | Days to low | Recovered | Days to recover |
|---|---|---|---|---|---|
| 2022-10-14 | 2026-09-21 | −61.6% | 1438 | not yet | – |
| 2021-03-08 | 2022-01-12 | −33.1% | 310 | 2022-09-30 | 261 |
| 2022-09-30 | 2022-10-04 | −10.5% | 4 | 2022-10-11 | 7 |
Buy-and-hold's deepest drawdown ran from 2021-03-08 to 2026-10-02 and reached −87.3%.
The drawdowns
The deepest drawdown, 61.64%, began at a peak on 2022-10-14 and reached its trough on 2026-09-21, 1,438 days later. There was no recovery before the window closed. The peak was the high point after the 2022 gains, and the next four years were a slow slide. The second drawdown was 33.08%, from 2021-03-08 to 2022-01-12, which recovered on 2022-09-30 after 261 days. The third was 10.47% over 4 days, from 2022-09-30 to 2022-10-04, and recovered in 7 days.
Buy-and-hold had a single drawdown of 87.33% from 2021-03-08 to 2026-10-02, which was still open on the last day. It never recovered. The rule's deepest drawdown was shallower by 25.69 percentage points. Both curves end below their starting point, so neither of them shows a recovery. The rule's 2022 gain was followed by a long decline that took most of it back.
The drawdown figures show why a rule that loses less than holding is still not a hold-and-wait. After 2022-10-14 there was no recovery, and each entry into a falling fund added to the slide. The 61.64% figure comes from many small losing trades, and no single event produced 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) | −13.6% | −61.6% | $4,312 | -0.30 |
| 5 basis points | −14.0% | −62.2% | $4,206 | -0.31 |
| 10 basis points | −14.3% | −62.8% | $4,116 | -0.32 |
What trading costs did
Adding 5 basis points per side gave a CAGR of -14%, a max drawdown of 62.23% and an end equity of $4,206. At 10 basis points the CAGR was -14.33%, the max drawdown 62.8% and the end equity $4,116. The cost-free run ended at $4,312.
Twenty-four round trips is enough for costs to register and too few for them to dominate. Each cost step moved the CAGR only slightly. The Sharpe ratio stayed near -0.3 across all three runs, at -0.313 and -0.323 in the cost runs. Costs left the verdict alone: the rule lost money before them and slightly more after. QID had an average daily dollar volume of $228,567,798, so the order sizes in a $10,000 account are small against the fund's trading.
Changing the parameters
| Version | CAGR | Max drawdown | Round trips | Win rate | Final value |
|---|---|---|---|---|---|
| Published rules | −13.6% | −61.6% | 24 | 17% | $4,312 |
| EMA 8/21 | −16.9% | −68.5% | 33 | 15% | $3,467 |
| EMA 20/50 | −13.5% | −61.4% | 12 | 8% | $4,361 |
| EMA 12/50 | −12.9% | −63.4% | 15 | 20% | $4,510 |
What other settings did
Three other pairs of averages were run.
- EMA 8/21, a faster pair: CAGR -16.85%, max drawdown 68.47%, end equity $3,467, with 33 trades and 5 wins.
- EMA 20/50, a slower pair: CAGR -13.46%, max drawdown 61.38%, end equity $4,361, with 12 trades and 1 win.
- EMA 12/50, the same fast average with a slower one: CAGR -12.95%, max drawdown 63.4%, end equity $4,510, with 15 trades and 3 wins.
The base rule, EMA 12/26, had a CAGR of -13.63% with 24 trades and 4 wins. The faster pair traded more and lost more, and the slower pairs traded less and lost a little less. All four results are negative, and the CAGRs sit within a few points of each other. Here the settings change how often the rule trades and leave the sign alone.
The 12/50 variant had the best result of the four, at -12.95%, and 20/50 was near it. Neither is evidence that a slower average suits QID. The differences are small, the trade counts are small, and the window is one stretch. The loss does not hinge on the exact pair of 12 and 26 days.
How QID behaved
| Measure | QID |
|---|---|
| Data in this test | 2021-01-04 to 2026-10-02 (1444 sessions) |
| Total return, buy and hold | −89.3% |
| Annualized volatility | 44.7% |
| Deepest drawdown | −89.6% (2021-03-08 to 2026-10-02) |
| Up days | 45.1% |
| Average daily range | 3.15% |
| Average overnight gap | 1.19% |
| Correlation to SPY | -0.94 |
| Correlation to QQQ | -1.00 |
| Correlation to TLT | -0.09 |
| Sessions above the 200-day average | 20.0% |
| Crossings of the 200-day average | 34 |
| Falls of 10% or more from a 20-day high | 78 |
Why QID was a hard fund for a trend rule
QID's realized beta to QQQ was -2, in line with its -2x daily target. QQQ returned 151.07% over the window while QID returned -89.27%. A simple -2x multiple of the QQQ return would be -302.14%, which the fund cannot deliver because a fund cannot lose more than all its value. The daily rebalanced ideal was -93.34%. The fund's own result of -89.27% sits close to that figure.
The fund's calendar returns were -46.44% in 2021, 66.4% in 2022, -57.25% in 2023, -34.03% in 2024, -34.98% in 2025 and -34.35% in 2026 to date. Only 20% of its sessions closed above the 200-day average, and it crossed that line 34 times. The fund's own max drawdown was 89.58%, starting 2021-03-08, with no recovery by 2026-10-02. A long-only trend rule has few good sessions to work with in a series like that.
Volatility makes it harder. Annualized volatility ran at 44.7%, with an average intraday range of 3.15%, an average up day of 2.16% and an average down day of -2.02%. Up days were just 45.11% of the total. The best day was 12.12% on 2025-04-04 and the worst was -23.53% on 2025-04-09. The rule's 2025 return was -10%, a year that held both of those days.
How QID compares with other inverse funds
Of the 11 inverse funds, QID's -13.63% under this rule is seventh best. TBF returned 7.6% with 25 round trips and a max drawdown of 17.96%. SPDN returned -3.78%, SH returned -3.91%, PSQ returned -6.71% and SDS returned -6.89%. EEV returned -12.74%, REW returned -18.47%, SQQQ returned -24.32%, TECS returned -29.4% and SOXS returned -48.38%.
Results across the group run from a gain of 7.6% on TBF to a loss of 48.38% on SOXS, with a category median of -12.74%, close to EEV's result. The rule's median over all 59 funds was 2.76%, and QID sat well below it.
Other rules on the same fund give a spread of results. The RSI(2) snapback lost 1.85% a year on QID and was the best of the 12 strategies. The momentum breakout lost 3.08%, the golden cross lost 6.03%, and the trend with a trailing stop lost 7.6%. The 200-day regime filter lost 10.46% and the SMA 10/50 trend lost 13.07%. The weakest were the monthly cycle at -28.81%, the dip buyer at -27.11% and the weekly 7% target at -26.24%.
The rules
Hold while the 12-day exponential average is above the 26-day, the core of the MACD used as a position switch.
- WHEN the market opens · IF not invested AND EMA(12) > EMA(26) · THEN buy with 98% of the sleeve
- WHEN the market opens · IF invested AND EMA(12) < EMA(26) · THEN sell the whole position
The 12/26 exponential moving average pair is the core of the MACD indicator. Exponential averages weight recent days more than simple averages, so this crossover reacts faster than an SMA pair of the same length. Used as a long/flat switch, it gives you a position you can hold and measure.
Good for: traders who like MACD logic but want it expressed as a simple, testable long/flat rule.
Watch out: faster reaction means more trades and more whipsaws than a 50/200 cross; check the trade count on each backtest page.
How the entry and exit rules behave on this fund
The rule buys at the open with 98% of the sleeve once the 12-day exponential average is above the 26-day, and it sells everything when the 12-day drops below the 26-day. Exponential averaging weights recent days more heavily, so the crossover reacts faster than a simple-average pair of the same lengths, and the strategy page notes that this brings more trades and more whipsaws than a 50/200 cross.
The whipsaw is visible on QID. The median hold was 21 days against an average of 25.8, and six trades lasted 7 days or less. With a 3.15% average daily range, the two averages cross often when the price swings sideways. Each cross costs a small loss, and the 17% win rate shows losses were the usual outcome.
The signal also has a built-in delay. The entry happens after the fast average has already crossed above the slow one, so it follows a rise that has already begun. The exit follows a fall that has already begun. On a fund that makes its moves in a few weeks, the crossover enters late and leaves late. The average win of 9.24% was larger than the average loss of -5.85%, which was not enough with a 17% win rate. The result is a profit factor of 0.31.
The test covers one fund, one rule and one window of 5.74 years, and that window held a single rising year for QID. Neither the rule nor the variants were tuned to this fund, and the figures are hypothetical.
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 EMA 12/26 trend beat buy-and-hold on QID?
Over 2021-01-04 to 2026-10-02, EMA 12/26 trend on QID returned −13.6% annualized vs −29.5% for buy-and-hold: it beat buy-and-hold by 15.8% per year, with a maximum drawdown 25.7 points shallower than holding (61.6% vs 87.3%).
How many trades did it make?
24 completed round trips over 5.7 years (48 fills), with 17% of round trips closing profitably.
Is this the same as trading MACD signals?
It uses the MACD's underlying trend component. Classic MACD trades the signal-line crossover of the 12/26 spread. This template trades the spread's sign directly, which is simpler and can be built fully in blocks.
How did the EMA 12/26 trend rule do on QID?
Between 2021-01-04 and 2026-10-02 it turned $10,000 into $4,312, a CAGR of -13.63%. The max drawdown was 61.64% and the Sharpe ratio was -0.3. It made 24 round trips, and 4 won.
Did the rule beat buy-and-hold on QID?
It lost less than holding. Buy-and-hold on QID ended at $1,348 with a CAGR of -29.46% and a max drawdown of 87.33%. The rule beat holding in 2021, 2023, 2024, 2025 and 2026, and trailed in 2022, when QID rose.
Why did the win rate come out at 17%?
QID drifted down in most of the window, so most upward crosses of the 12-day average faded. The rule won 4 of 24 trades, and three of those wins were in 2022. The average win was 9.24% and the average loss was -5.85%.
Do other EMA settings fix the result?
No setting tested made money. EMA 8/21 returned -16.85%, EMA 20/50 returned -13.46% and EMA 12/50 returned -12.95%. The faster pair traded 33 times and lost the most.
How much do trading costs matter on this rule?
A cost of 5 basis points moved the CAGR from -13.63% to -14%, and 10 basis points moved it to -14.33%. End equity went from $4,312 to $4,206 and then to $4,116. Costs made a loss slightly larger.
Is QID a good fund for a trend rule?
In this window it was not. QID is a -2x daily fund on the Nasdaq-100, and it lost 89.27% while QQQ gained 151.07%. It rose only in 2022, and the rule ranks 53rd of 59 funds on CAGR.
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.