Almost every retirement projection you'll ever see rests on a single number: an average annual return. You type in 7%, the spreadsheet grows your money by 7% every single year, and it tells you your money lasts forever.
Here's the problem. The market has never once returned its average. It gives you +22%, then −9%, then +4%, then −31%. The average of those four years is real enough, but no individual year looked anything like it.
While you're still saving, that barely matters. Once you start living off the money, it matters more than almost anything else, and it's the part of retirement planning I find gets explained the least.
The average that hides the risk
Picture a portfolio of €500,000. You take out €25,000 at the start of each year, which is 5% of the starting balance. Over the next three years the market returns −30%, −10% and +50%, in some order.
Two people retire with identical portfolios and live through those same three years. The only thing that's different is the order.
| Bad years first | Good years first | |
|---|---|---|
| Start | €500,000 | €500,000 |
| Year 1 return | −30% | +50% |
| End of year 1 | €332,500 | €712,500 |
| Year 2 return | −10% | −10% |
| End of year 2 | €276,750 | €618,750 |
| Year 3 return | +50% | −30% |
| End of year 3 | €377,625 | €415,625 |
Three years in, there's €38,000 between them, close to 8% of the original capital. Nothing created that gap except the order the years arrived in.
Now take the withdrawals away. Leave both portfolios completely alone and they both end at exactly €472,500, because multiplication doesn't care about order: 0.70 × 0.90 × 1.50 gives you the same answer read backwards.
The idea in one line: The order of returns is irrelevant to a portfolio nobody touches, and decisive for one you're drawing from. Selling units in a fallen market turns a temporary loss into a permanent one, because those units aren't there when the recovery finally arrives.
That's sequence-of-returns risk. It's why a rough first decade of retirement isn't something you quietly make back later, and why two people who retire two years apart with the same plan can end up somewhere so different.
Where the 4% rule came from, and what it actually claimed
Back in 1994, a financial planner called William Bengen went looking for the withdrawal rate that would have survived the worst moment in the historical record. He tested rolling 30-year retirements against real US market history and found that 4% of the starting portfolio, rising each year with inflation, had never run out within 30 years. The Trinity study broadened the work a few years later, and the number stuck.
It's worth reading that claim closely, because it's a lot narrower than the way you usually hear it repeated:
- US data. A century of returns from the most successful stock market of that century.
- Thirty years. Not forty-five. If you're retiring at 45, you're asking a different question.
- A fixed, inflation-adjusted withdrawal. The retiree in the model never once reacted to a crash.
- Before costs and taxes. Fees and tax come out of the same portfolio, and they aren't in the number.
- "Success" means one euro left on the last day. A plan that finished with €12 counts as a win.
None of that makes the 4% rule useless. It's a genuinely useful piece of history and a good place to start a conversation. It was just never a law, and it was never a promise.
What a Monte Carlo simulation adds
If the order matters, then a single projection is only telling you about one ordering out of an enormous number of possible ones. A Monte Carlo simulation generates thousands of them instead.
Each run draws a fresh sequence of yearly returns from a distribution you choose (an expected return, a volatility) and plays your plan through it: your withdrawals, your allocation, your horizon. One run is a story. Ten thousand runs are a distribution, and the answer stops being a number and turns into a shape.
What comes out is a success rate, meaning the share of simulated futures in which the money outlasted you. It'll also show you the failures, which are honestly the more instructive half. You get to see when those plans ran dry, and what their first five years looked like on the way there.
How to read a success probability honestly
A 90% success rate doesn't mean "you'll be fine". It means one in ten of the futures the model generated ran out of money. A few things are worth holding on to:
- It's a model of a model. Random draws from a smooth distribution are much tidier than real markets, which have fat tails and long moods. Independent draws also lose the way bad years tend to cluster together.
- The inputs dominate the output. Move the expected return by one percentage point and the success rate shifts more than any clever refinement will.
- Nobody behaves like the model retiree. Real people cut back in a crash. That single behavior, which the simple version leaves out, is worth more than most portfolio changes.
- Chasing 100% has a price. Certainty gets bought with extra years of work and a smaller life. Somewhere in the high eighties, more precision stops being the useful question.
The four levers that genuinely move the outcome
When a plan looks fragile, only a handful of things really change the picture, and they aren't equally hard:
- Spending flexibility. A rule you agree with yourself in advance — hold withdrawals flat after a down year, trim them by 10% after a bad one — lifts survival more than almost anything else, because it stops the forced selling that does the damage.
- The first few years in cash. One to three years of expenses held outside the market means your first bad year doesn't have to be paid for by selling into it.
- Allocation. Enough growth to outrun inflation over thirty years, enough stability to survive the first five. Both failure modes are real, and only one of them is loud.
- Any income at all. Part-time work, a pension that starts later, one property. Small, unglamorous income streams shorten the window your portfolio has to cover on its own, and the effect is much bigger than it feels.
What none of this can tell you
A simulation can't tell you what the market will do, what your health will do, or what you'll want at seventy. It isn't a forecast, and it certainly isn't personal advice. What it does is more modest and a lot more useful: it shows you which assumptions your plan is leaning on, and how much has to go wrong before it breaks.
So run your own numbers through the Monte Carlo FIRE simulator, then change one input at a time and watch which one the answer really cares about. Ten minutes there will teach you more about your own plan than another average return ever could.
Ready to Master Your Money Habits?
The order of returns isn't something you control, but your savings rate, your cash buffer and your reaction to a bad year all are. Take our habit assessment to see where your plan is strong and which of those levers is worth pulling first.