We don't lead with one number, because one number hides how it was measured.
Below is the same framework measured several ways — from the conservative, reproducible-by-anyone basis to the most favorable. Every figure is a real backtest on real data (TQQQ from 2010, through 2026-06-05). The honesty is in showing both anchors with equal weight and labeling the method on each — never in hiding one.
The two anchors
| Basis | CAGR | Max DD | Calmar |
|---|---|---|---|
| Tracked (weekly 220-day) | 31.2% | −38.9% | 0.80 |
| Conventional (daily 200-day) | 29.1% | −38.6% | 0.75 |
The honesty block — the SMA sweep
The moving-average length is a dial. Here is how it moves the result across the favorable range — so you can see we're not quoting only a best case. 44 weeks is the high point, and its neighbors score lower, which tells you it's the top of a favorable band, not a magic number.
| Macro SMA | CAGR | Max DD | Calmar |
|---|---|---|---|
| 40 weeks | 27.0% | −45.0% | 0.60 |
| 42 weeks | 30.0% | −41.3% | 0.73 |
| 44 weeks (locked) | 31.2% | −38.9% | 0.80 |
| 46 weeks | 27.1% | −38.5% | 0.70 |
| 48 weeks | 23.8% | −41.3% | 0.57 |
The benchmark, for scale
TQQQ held naked returned more on paper — and fell 80%, which almost nobody holds through. The framework's job is to capture leveraged growth while keeping the drawdown survivable enough to actually stay invested.
| Strategy (2010–2026, real data) | CAGR | Max Drawdown |
|---|---|---|
| Alpha Gear (locked) | 31.2% | −38.9% |
| S&P 500 buy & hold | 12.0% | −32.2% |
| TQQQ buy & hold | 39.4% | −80.5% |
The bad years — shown on purpose
A strategy that only shows its good years is selling, not informing. Here is every rough year, the framework versus just holding. 2015 was our worst relative year — the rules went defensive and the market chopped sideways; the caution cost us.
| Year | Framework | S&P 500 | TQQQ |
|---|---|---|---|
| 2011 | −1.4% | −2.9% | −19.6% |
| 2015 | −19.5% | −1.3% | +13.1% |
| 2018 | +6.1% | −7.2% | −21.6% |
| 2020 | +34.2% | +14.3% | +98.8% |
| 2022 | −30.6% | −19.5% | −79.2% |
| 2025 | +11.0% | +16.0% | +30.7% |
The upside anchor. The same rules that lag in choppy years ride the booms when trend, slope and confirmation all align.
Leverage's structural floor: a fast crash from an all-green state hits before any weekly signal can fire. The rules then defend the following week.
Why the test starts in 2010
TQQQ — the 3× fund the framework actually trades — launched in February 2010. We test on the real instrument from the day it existed, and not one day earlier. We could manufacture a longer record by splicing in synthetic pre-2010 leverage; we don't, because modeled leverage is unreliable. So the record is exactly 15 years — honestly bounded by the instrument's real life. The cost of that honesty: this window contains no 2008, and no full-cycle bear market for leveraged tech. We'd rather you know that.
The hardest number — time underwater
The worst drawdown was −38.9%. But the figure that actually tests you is this: the framework spent its longest stretch — about 1.6 years — below a previous high before recovering. Could you hold a leveraged position for 19 months in the red without selling? If the honest answer is no, leveraged trend-following is not your strategy.