A compass to navigate the markets while keeping the helm: time-tested principles, real numbers, and honesty about what no one can guarantee.
From the invention of the share in Amsterdam (1602) to today's algorithms, the tools have changed everything. The passions, never. Newton, ruined by a bubble, said it all: "I can calculate the motion of the stars, but not the madness of men."
Bachelier showed that tomorrow's price is unpredictable. Nash, that an easy edge, once shared, destroys itself. Kelly, how to size a bet without ruin. And behavioural finance closed the loop: the market is irrational because we are.
The method separates two pockets with radically different behaviour:
The bedrock of serenity. A prudent mix whose components have historically delivered, over the long run, an order of magnitude of 4 to 6 % per year. It's the hull: it takes the swell.
The auxiliary sail. Small, framed, bound by a contract of responsible behaviour. A budget for learning and enjoyment, never an engine of wealth.
None of the figures below is a promise. They are data observed in spring 2026, in ranges, with public sources.
| Instrument | Recent observed | Risk |
|---|---|---|
| Euro funds | ~2.6 % | Capital protected, eroded if inflation > yield |
| Money-market (€STR) | ~1.9 % | Low, tracks short ECB rates |
| IG euro bonds | ~3.6 % | Rate, credit, liquidity |
| World equity ETF | ~8–11 %/yr* | Volatile: 30–50 % drops possible |
*long-run average, with real volatility of 13–16 %/yr. Sources: ACPR, ECB, MSCI, iShares, Cboe.
⚠ No available product guarantees a minimum of 4 % or more. When the risk-free rate sits at 2–3.7 %, guaranteeing more would mean hidden risk. The "4–6 %" is the order of magnitude of a prudent-to-dynamic mix over the long run — never a contractual floor.
Choose a goal and a horizon, or a monthly amount and a goal. The method never shows a single figure: always three scenarios (prudent, median, optimistic) and a ±10 % margin.
Scenarios: prudent 3 % · median 4.5 % · optimistic 6 % (recalibrated to the market). The effect is slow in year one, powerful over 10–20 years.
The 10 pocket follows non-negotiable rules, drawn from mathematics (Kelly) and experience:
Bet a small fraction of capital, proportional to the real edge. After a loss, the bet shrinks automatically — never the opposite.
Splitting the pocket over ~5 positions rather than one reduces the spread of outcomes at no cost to return. The only prudence that doesn't cost.
A maximum loss set in advance, respected. ⚠ A stop slips in reality: on an opening gap, you exit lower than planned. It cuts the expected loss, not the maximum loss.
Never average down. Never buy back the same day to recover (that's a martingale, a path to ruin). Never increase the bet to catch up.
Over 5 years of real prices and thousands of simulations, three robust findings — and one warning.
1. Diversification wins. Spreading over 5 names instead of one clearly narrows the gap between best and worst case.
2. The small pocket optimises risk. A 90/10 offers the best risk-adjusted return; loading the risk pocket raises drawdowns faster than gains.
3. Holding beats tinkering. A rule that enters and exits at the "right moment" did worse than simply holding — while taking more risk.
⚠ What we will never do: present the best past results as an expected return. The past of a basket chosen after the fact, over a favourable window, does not repeat. Backtest figures are not a promise — they help understand a structure, not sell a future.