We ran a real €84,615 portfolio through 2020 and 2022. Here is what happens.

Volatility is abstract; “−€20,815 in March 2020” is not. Four historical episodes replayed on a real multi-asset portfolio — with the numbers.

We took OplynQ's public sample portfolio — 11 positions, €84,615, a realistic mix of stocks, ETFs, crypto, bonds, commodities and cash — and ran it through four market episodes, from the COVID crash of March 2020 to the 2022 rate shock. Not as a thought experiment: with OplynQ's stress test engine, which replays historical shocks on the portfolio's current weights. The worst of the four scenarios costs €20,815.

This article explains what a stress test is, why that number lands harder than any volatility figure, how to read the four results — and what any of this is actually for. It will never tell you what to buy or sell: measuring is not advising.

What a stress test actually is

The idea fits in one sentence: take your portfolio exactly as it stands today — your positions, your weights — and apply the shocks each asset class actually suffered during a given historical episode. Stocks fell by so much during that episode, bonds by so much, crypto by so much: apply each shock to the matching slice, add it up, and you have an answer to the question “if that episode replayed on my portfolio as it is today, how much?”.

The canonical episode is 2008 — the one every professional replays. But your assets need a price history that reaches back that far: crypto, for one, did not exist in 2008. An honest tool says so — OplynQ shows that the scenario cannot be covered — rather than extrapolating a number nothing supports.

Two properties keep the exercise sane. It is deterministic: same weights, same episode, same result — nothing random, nothing opaque. And it is not a prediction: no future shock will look exactly like a past one. A stress test gives you an order of magnitude grounded in what actually happened, not a prophecy.

Why it lands harder than a volatility figure

Volatility, VaR and drawdown — we walk through how to read them in our article on portfolio risk analysis — compress a portfolio's behaviour into annualised percentages. Indispensable, but abstract: nobody feels a standard deviation. A stress test converts the same risk into the unit you will actually live it in: euros, attached to an episode that has a name and memories. “March 2020: −€20,815” triggers a reaction no annualised percentage ever will.

The two families complement each other. Volatility and VaR describe the ordinary regime — the churn of normal days. Stress tests illuminate the tails: those few weeks per decade when everything moves at once, which is where most real risk lives.

The portfolio we tested

The portfolio is the one any visitor can explore on OplynQ without creating an account: 11 positions, €84,615 on the day of the calculation, allocated as follows:

• Stocks — 32%
• ETFs — 28%
• Crypto — 13%
• Cash — 12%
• Bonds — 8%
• Commodities — 7%

It is neither a model portfolio nor a recommended allocation — just a realistic, multi-asset example, the kind our users actually import.

The results

Here is what OplynQ's stress test engine returns on those weights, scenario by scenario, as a percentage of the portfolio and in euros:

How to read these numbers

The most instructive line is not COVID — it is the second one. The 2022 rate shock was far less spectacular than March 2020: no flash panic, no historic sessions. Yet it costs this portfolio almost as much: −€17,769 versus −€20,815. The reason: in 2022, bonds — the usual shock absorber of a diversified portfolio — fell at the same time as stocks. Add up the slices in the line of fire — stocks (32%), ETFs (28%, which themselves hold stocks and bonds), bonds (8%) — and most of the portfolio takes the same episode at the same moment. When correlations rise, apparent diversification evaporates: that is exactly the trap we measure in our article on effective diversification.

The 2022 crypto winter is the mirror image: the most brutal of the four episodes for the asset class involved, yet contained at −8.3% of the portfolio — because crypto only weighs 13% of the total. An extreme shock on a small slice produces a bounded loss. That is the arithmetic of weights, and it is precisely what a stress test makes visible: risk is not the violence of a scenario, it is the product of violence × weight.

One last, mechanical detail: the 12% held in cash takes part in none of these shocks. It earns nothing in the calculation — but it caps the total loss, in every scenario, by simple arithmetic.

What it is for (and what it is not for)

A stress test exists to size your risk before the shock, not during it. The useful question is simple: “if −€20,815 appeared on my screen tomorrow morning, would I hold my strategy — or sell everything at the worst possible moment?”. If the honest answer is no, the stress test has taught you something essential about your allocation. What you do with that is yours alone: OplynQ measures and explains; it never recommends buying or selling anything.

During the shock it is too late to ask: decisions get made on adrenaline, with numbers moving every second — historically the moment when the worst calls are made. A stress test moves that reflection to a moment when you are calm. That is its entire value.

One more habit worth building: re-run the test after your portfolio changes. Weights drift — a position that doubles, a new deposit, a sale — and every drift moves the scenario losses with it. A stress test is not a certificate you earn once; it is a reading you take again whenever the portfolio you actually hold stops resembling the one you tested.

What OplynQ does

On OplynQ these stress tests run on your portfolio in one click: import your broker statements (CSV, PDF, Excel, OFX) or connect accounts read-only, and the engine replays the episodes on your real weights, in percent and in euros. The AI copilot then explains the result: which slice drives which scenario, and why. And when a holding's history is too short to cover a scenario, you see “—” and the reason — never a silent extrapolation.