Two engines
Two engines, two personalities.
The Guardian aims never to fall with the market. The Growth Engine aims to beat it over time. Here are their curves, their lows and their years, with nothing hidden.
$100,000 invested in October 2007: what each is worth in September 2026
Simulation on real prices from 2007-2026, fees and the cost of leverage included, with interest earned on cash. Logarithmic scale: the same height represents the same percentage.
Engine No. 1
The Guardian: eighteen years without a major fall
Its strength is not rising faster than the market. It is never collapsing with it.
Its declines from its last peak, compared with the market's
Each dip measures the decline from the last record high. The Guardian never spent a single week more than 20% below its record. The market spent 111.
| The Guardian | Market (S&P 500) | |
|---|---|---|
| $100,000 at the end of 2007 becomes | $306,000 | $700,000 |
| Average gain per year | +6.1% | +10.8% |
| Worst drawdown from a peak | −17% (2018) | −55% (2009) |
| Worst year | −11% (2018) | −37% (2008) |
| 5-year periods ending in a gain | 99.8% | — |
| Weeks more than 20% below the record | 0 | 111 |
- What it buys
- About thirty major US markets: indexes, sectors, gold, oil, government bonds and very large companies.
- When it buys
- After a sharp drop lasting a few days, when the price stays above its long-term trend. It sells as soon as calm returns, often within three days.
- Its reserve
- About 40% of the account waits in US Treasury bills, earning the current rate and available the same day.
- Its circuit breaker
- A full stop if the loss exceeds a limit set in advance. Never reached in 18 years of simulation.
The Guardian is replayed with its current list of 36 instruments, which includes 8 large US stocks chosen in 2026. That choice makes the past look better: on a list chosen without knowing what came next, our internal tests show 2 to 3 points less gain per year and a worst drawdown of around −25 to −30%.
Engine No. 2
The Growth Engine: 33 years, capital multiplied by 100
A simple rule applied to the world's largest index: invest more when the market is calm, less when it gets nervous, and nothing at all when the long-term trend turns down.
33 years of history: $100,000 in 1993 becomes $10.3 million
Simulation of the rule on the S&P 500 (1993-2026), fund fees, the cost of leverage and trading frictions included. Logarithmic scale.
In 2000-2002 and in 2008, it got out before most of the fall
Worst drawdown of the Growth Engine: −41% in 2022, over the S&P 500's history since 1993. This is not a floor: a future decline can be deeper.
| The Growth Engine | Market (S&P 500) | |
|---|---|---|
| $100,000 in 1993 becomes | $10,270,000 | $2,930,000 |
| Average gain per year (33 years) | +15.2% | +10.8% |
| Year 2008 | +1% | −37% |
| Year 2020 | +13% | +18% |
| Year 2022 | −32% | −18% |
| Worst drawdown from a peak | −41% (2022) | −55% (2009) |
| 10-year periods ending in a gain | 100% | — |
Year by year since 2008: 14 winning years out of 19
2026: from January 1 to September 9.
- What it holds
- The S&P 500 index itself, and a listed fund that multiplies it by three for the leveraged part. Nothing exotic, and everything can be sold in a second.
- How much it invests
- Between 0 and 2.5 times the capital, depending on how nervous the market is. The calmer the market, the more it is invested.
- When it gets out
- As soon as the index falls clearly below its average of the last 200 days, it sells everything. It only comes back once the trend has recovered.
- Its pledge
- It never sells in reaction to a one-off drop. Its worst historical drawdown is known and accepted in advance.
The Growth Engine uses a leveraged fund (index ×3): a listed product that is easy to buy and sell, and that amplifies moves in both directions. It never holds it below the long-term trend.