Canary-Signaled Tactical ETF Strategy Turned $10,000 → $15,458 (Full Backtest)
A $10,000 stake run through this rule as a historical simulation would have finished near $15,460, while the same money parked in SPY would have finished near $41,180. The rule pays for that gap with a shallower worst drop and a monthly switch that steps out of leveraged funds when inflation-protected bonds are falling. Whether that trade is worth it depends on how you weigh a smaller hole against a much smaller pile.
📈 Yearly return (CAGR): 4.5%
📉 Worst drop (max drawdown): -16.9%
⚡ Sharpe Ratio: 0.46
💰 Total Return: 54.6%
🆚 S&P 500 (SPY) over the same period: 311.8% — this strategy underperformed by 257.2 points
🎯 Universe: TQQQ, UPRO, SOXL, FNGU, BULZ, TECL, TNA, SPXL, TIP, BIL, IEF, GLD
| Metric | Canary-Signaled Tactical ETF Strategy | S&P 500 (SPY) |
|---|---|---|
| Total return | 54.6% | 311.8% |
| Worst drop (max drawdown) | -16.9% | -33.7% |
| Period measured | 2016-10-04 – 2026-09-15 | |
Both columns come from the same backtest run over the identical period. A higher return with a deeper drawdown is not automatically better.
| 2016 | 2017 | 2018 | 2019 | 2020 | 2021 | 2022 | 2023 | 2024 | 2025 | 2026 |
|---|---|---|---|---|---|---|---|---|---|---|
| +0.0% | +11.8% | +24.1% | -4.3% | -0.9% | -0.1% | +1.4% | +10.0% | +3.1% | -0.7% | +3.1% |
How the Rule Works: Rank, Hold, Check the Canary
On the last trading day of each month, the rule scores every leveraged growth fund (TQQQ, UPRO, SOXL, FNGU, BULZ, TECL, TNA, SPXL) by blending its one, three, six, and twelve month returns, then holds the top few in equal weights until the next check. Before it does any of that, it reads a single canary: TIP, the inflation-protected Treasury fund, scored the same way. A positive TIP blend means risk-on and the leaders get bought; a negative blend means the entire portfolio moves to whichever of IEF, BIL, or GLD scores best, on the theory that falling real-rate bonds usually precede trouble for growth stocks. The adapted version here adds a second gate, requiring SPY to sit below its 200-day moving average before the defensive switch fires, because TIP alone can flinch during short bond-market scares while equities are still fine.
Does It Beat the S&P 500? No, by a Wide Margin
It does not: SPY's total return was 311.8 percent against 54.6 percent for this rule, an underperformance of 257.2 points over the same window, and SPY's worst drop of 33.7 percent was only about twice this rule's 16.9 percent. The shortfall comes from mechanics, not bad luck: the canary and the moving-average filter pull the portfolio into bonds and gold after damage is already done, then keep it there while leveraged funds stage their fastest rebounds, so the rule repeatedly sells low and rebuys higher. Leveraged ETFs also decay when markets chop sideways, and monthly ranking cannot avoid that grind. A higher return with a deeper drawdown is not automatically better, but here the index delivered both a far higher return and a drawdown most long-term holders could tolerate, so the defensive machinery cost more than it saved.
Where It Hurts: Whipsaws and Slow Re-Entry
The losing stretches cluster around regime turns: a sharp sell-off triggers the defensive switch near the low, the market recovers within weeks, and the rule sits in BIL or IEF until the next month-end confirms the turn. With only twelve decisions a year and 3x-leveraged holdings, one badly timed month can erase a year of gains, which is why the test logged four negative years against seven positive ones. The double filter reduces false alarms from TIP but also delays real exits, so the worst drop is not a ceiling; a faster crash than any in the test window would pass straight through a month-end gap. Gold and intermediate Treasuries can fall alongside stocks when rates rise, which means the defensive basket is not a guaranteed shelter.
Who This Fits: A Research Framework, Not a Core Holding
This suits someone who wants a rules-based way to hold leveraged growth funds without watching them daily and who values a bounded drawdown over maximum growth. It does not suit anyone whose benchmark is simply owning the index, because that alternative won both on return and on effort. Position sizing has to assume the 16.9 percent drop recurs and possibly deepens, so the rule belongs in a satellite slice rather than the whole portfolio. Treat these figures as a historical simulation to study, compare against a plain SPY holding and against unleveraged versions of the same signal, and do not expect the past sequence of regimes to repeat.
Frequently Asked Questions
How risky is a canary-signaled leveraged ETF rotation strategy?
In this simulation the worst peak-to-trough drop was 16.9 percent, milder than SPY's 33.7 percent, but the holdings are 3x-leveraged and the exit only checks once a month, so a faster crash could produce a larger loss than the test shows.
Can I trade this strategy today with a normal brokerage account?
All twelve funds are exchange-listed and the rule needs only month-end prices and a 200-day average of SPY, so it is mechanically tradeable, but the results here are historical simulation and say nothing about live returns or future profits.
What does this backtest assume?
It assumes fills at month-end closing prices with no slippage, commissions, taxes, or bid-ask spread, that every fund had continuous trading history over the window, and that the ranking and canary rules were fixed in advance rather than tuned to the data.
| Period | 2016-10-04 – 2026-09-15 |
| Universe | TQQQ, UPRO, SOXL, FNGU, BULZ, TECL, TNA, SPXL, TIP, BIL, IEF, GLD |
| Years with a gain | 7 of 11 |
| Price data | FinanceDataReader daily bars |
What this does not prove. No out-of-sample split and no parameter-sensitivity test were run, so these figures may be flattered by hindsight. Signals are computed only from bars that closed before the session they trade. Full assumptions, including which integrity checks are unverified: avalonquant.com/methodology.
More Backtests
📊 A new strategy backtested every day — the numbers, the drawdown, and what breaks it.
▶ Subscribe to Avalon Quant⚠️ Not financial advice. This is for educational purposes only. These results come from a historical backtest — past performance does not guarantee future results. Always do your own research before investing.
📊 Live strategy performance and daily quant briefings: avalonquant.com
📄 Original paper: Dual and Canary Momentum with Rising Yields/Inflation: Hybrid Asset Allocation (HAA)
Comments
Post a Comment