The Nugget Illusion

Let’s be honest. Almost everyone has had that thought at least once:

Why didn’t I just buy Amazon in 2000?

Why didn’t I hold Nvidia for the last ten years?

Why didn’t I buy bitcoin back in 2012?

Looking back, the winners feel obvious. The giant gold nugget in the riverbed.
Looking forward, they never were.

For every long-term superstar, there were dozens of companies that looked just as promising. And then they quietly disappeared – Yahoo, MSN, Oatly, Meyond Meat… the overwhelming amount of gravel that must be sifted through. Once praised as the next Wall Street star, now quietly forgotten. What remains in hindsight is a clean success story. What disappears is failure and uncertainty. Historically, stock markets always were driven by very few companies. What once was General Electric or ExxonMobil in the 1990s now are Amazon or Apple or Tesla.

Why stock picking underperforms

That is survivorship bias. The illusion that finding the next gold nugget is easier than it actually is. The truth is: it never was and never will be.

But instead of trying to predict the next outlier, we asked a different question:

What if investing was not about finding the next nugget — but about operating a structured mining process that systematically extracts value instead of hunting for luck?

From Prediction to Measurement

But how? We knew: Markets do not reward stories and promises.
They reward strength.

One of the most extensively documented return factors in financial research is momentum — the tendency of stocks that have outperformed their peers to continue outperforming over defined time horizons. Momentum ETFs apply the same principle but remain broadly diversified, often holding hundreds of stocks with slow turnover.

What is important to mention: Momentum does not forecast the future.
It measures relative strength in the present.

That distinction matters.

So instead of asking,
“Which company will dominate the next decade?”
momentum tells:
“Which stocks are currently strongest compared to all others?”

This shift from narrative to measurement forms the foundation of our system.

Structural Design

How can this now be converted into a significant and convenient system that is as simple as an ETF but as powerful as real trading? First: It required a strong focus on a low amount of stocks but also: a reduction in transaction effort. That’s why we defined clear boundaries:

Five stocks — concentrated, but tradable
Monthly rotation — systematic, but practical
Equal weighting — disciplined capital allocation
Predefined exit thresholds — rule-based risk control

Could we increase turnover? Yes.

Could we hold 10–15 stocks and rebalance weekly or even daily? Also yes.

But models are not traded in spreadsheets.

Higher turnover increases:

Transaction costs
Slippage
Operational complexity
Emotional override risk

We consciously accepted a small theoretical edge reduction in exchange for consistency, cost awareness, and real-world tradability.

A strategy must survive execution — not just simulation. A lesson we learned the hard way from other services. (you want to know how real execution costs affect your performance? Click here to learn more)

How the Model Operates

Each month, all stocks within a defined universe (e.g., S&P 500 or a selected German index) are ranked. The ranking is based on a defined set of parameters that quantify relative strength and momentum characteristics. Stocks with a high momentum are ranked higher, those with a low momentum are ranked lower. The ranking is updated daily to reflect dynamic changes within the universe.

From this ranking:

The top five stocks form the portfolio at the beginning of a month.
Positions are reset to equal weight.
During the month, predefined downside thresholds monitor risk – otherwise, the portfolio remains untouched.
If a stock breaches its threshold, it is removed and held as cash for the remainder of the current month.
The cycle repeats by investing into the new (or old) top five stocks of the following month.

There is no discretion.
No interpretation of earnings calls.
No macro forecasts.
No “gut feeling.”

Only rules. React, not believe, hope or anticipate.

Removing the Human Variable

This became our only rule: to follow and apply the rules. To eliminate one of the greatest weakness: the human emotions. Even experienced investors are prone to:

Holding losers too long and selling winners too early („pruning roses, manuring weed”)
Overreacting to news
Overriding rules “just this once”

A purely rule-based structure removes that layer and in fact allows to read and benefit from these collective emotions that are distilled in the price movement of stocks. These two main drivers: fear and greed.

But a purely rule-based system allows even more than mere elimination of bias. It enables systematic validation.

Because the rules are explicit, they can be:

Backtested
Stress-tested
Cross-validated
Optimized within controlled boundaries

And this is precisely what we did. We used discrete machine learning algorithms, synthetic markets and parameter shake tests to improve and validate our model. The result is not a model that has performed best in the past (a classical bias for backtesting) but a model that showed robustness when the conditions change. This didn’t make our model bullet-proof but fit to react to different situations it hasn’t seen before.

The objective was not historic perfection.
But future consistency.

A System, Not a Story

The model does not attempt to identify the next Amazon in advance.

Instead, it continuously reallocates capital toward stocks that already demonstrate measurable relative strength.

It does not eliminate risk.
It does not guarantee returns.

But it replaces hindsight-driven storytelling with structure.

And over time, structure tends to outperform emotion.

This is how we moved from digging for gold in the riverbed to systematically processing the entire mountain. It may not produce shiny nuggets or heroic stories. But it produces something far more valuable: consistency.