Femi is 58. He has £750,000 invested, wants £33,000 a year out of it, and the Monte Carlo simulator puts the success rate at 58.1%. Which is the same as saying it fails 41.9% of the time. A coin flip, tilted only slightly his way, on the question of whether he spends his eighties with money in the bank.
Nobody stops working on that number. Femi does another year.
Now run the identical simulation through rich, broke or dead, which takes the same 1,000 paths and asks a narrower question: in how many of them is Femi alive, and broke, at the same time? The answer is 10.4%.
Neither number is wrong. The second one is just the one he was trying to ask. The pound signs are a display setting rather than an assumption, by the way: the currency selector at the top of the input panel switches them to dollars, euros, rupees or any of the others on the list, and nothing underneath moves. The run is the tool's own default as well, a thousand paths at the 15% volatility it holds fixed, drawn from a seed that never changes, so your chart will match this one.
Where the 42% actually lands
A conventional simulation counts a path as failed if the money runs out at any point before 95, and treats a failure at 94 exactly like a failure at 68. Those are not remotely the same event, and the difference is not financial.
Femi's failures cluster late. Of his 1,000 paths, 4.2% have run dry by 75, 15.0% by 80, 26.1% by 85 and 41.9% by the end. So roughly two in three of the failures arrive after 80. By that age, on the population tables the tool uses, 41.7% of the people who were 58 alongside him have died.
Multiply the two curves together and you get the chart: at every age, the share of people still going with money, the share still going without it, and the share no longer here.
| Age | Rich | Broke | Dead |
|---|---|---|---|
| 75 | 70.5% | 3.1% | 26.4% |
| 80 | 49.5% | 8.7% | 41.7% |
| 85 | 29.4% | 10.4% | 60.2% |
| 90 | 13.6% | 7.7% | 78.8% |
| 95 | 4.3% | 3.1% | 92.6% |
Read down the broke column and notice it goes up, then comes back down. That shape is the whole idea. The band is thin early because the portfolio has barely been touched, thin at the far end because hardly anyone is there, and fattest at 85, where 39.8% of Femi's cohort are still alive and a quarter of the paths have failed. Eighty-five is the year his plan is most exposed, and it is nowhere near the 95 the headline rate is built around.
The 4.3% in the bottom left is worth a look too. That is the chance Femi reaches 95 with money still in the account, and it is the headline rate with the survival question put back: 7.4% of people his age get to 95 at all, 58.1% of the simulated money does, and the corner of the chart where both happen is 4.3%.
What a year of his life buys
Femi's instinct was to work longer, so price it. Suppose he stays another year and saves £1,900 a month while he does. His pot reaches £810,818 by 59: £22,800 of his own money and £38,018 of growth.
Rerun it. Success goes from 58.1% to 64.4%, and the worst-age broke figure falls from 10.4% to 7.8%.
A whole year of being 58, for 2.6 percentage points.
A second year takes the pot to £874,677 and the risk to 5.8%, so the next 2 points cost roughly what the first ones did. Meanwhile, cutting planned spending from £33,000 to £31,000 gets him to 7.3% on its own, which is slightly better than the first extra year of work and costs him £167 a month in retirement instead of twelve months of his life.
And the largest single move on the page is one he has already paid for. Add the State Pension as an income event from 67, at the current full new rate of £241.30 a week, or £12,548 a year, and the alive-and-broke peak drops to 3.9% while success climbs to 79.7%. Guaranteed income indexed for life is mortality insurance in the only form that matters: it pays precisely as long as you are around to need it. Check when yours starts, because the age depends on your birth year, on GOV.UK. A defined-benefit pension or an annuity bought in your seventies does the same job, which is why the annuity question belongs in this conversation and not in the one about beating the market. The name on it is local and the arithmetic is not: Social Security, the Age Pension, a company scheme, all of them pay while you are alive to be paid, which is exactly the risk the chart is pricing. One practical note, since an income event is a full-plan control rather than a slider: events, a partner and a withdrawal rule all live on the FIRE calculator, which carries Rich, Broke or Dead as one of its chart modes. The standalone page has the six sliders and the currency, which covers everything up to this point.
The version where mortality makes it worse
None of this is a general licence to relax, and there are three places the overlay cuts against Femi rather than for him.
The first is that he might not be planning alone. Give him a partner aged 56 and the money has to last until the second death, not the first. The chance that at least one of them reaches 85 is 63.4% rather than 39.8%, and the alive-and-broke peak rises from 10.4% to 17.8%, landing at 86. Same pot, same spending, same simulation, and seventy per cent more risk. For couples the mortality overlay is not the comforting view; it is the honest one, and the couples calculator is where that plan belongs.
The second is care. The model spends a flat £33,000 a year for life, and residential care, for anyone who ends up needing it, is an expense of a different order from ordinary living costs. It also lands in exactly the decade the chart is busy reassuring you about, and the people who live long enough to need it are the same people the dead band was supposed to excuse from the calculation. Nothing in the tool knows about any of it, so price your own local figure separately and treat the chart as the answer to a smaller question.
The third is that early failures get no protection at all. A path that runs out at 68 is a path where Femi is almost certainly alive, healthy enough to notice, and too far from the workforce to fix it. Mortality discounts the tail. It does nothing for the years that actually frighten people, which is why a plan should still go through the crash test before anyone takes comfort from a low alive-and-broke number.
Then there is the curve itself. The Gompertz-Makeham curve behind the chart is calibrated to contemporary UK and US unisex period tables, the kind the ONS publishes in its national life tables, and period tables assume today's death rates never improve, so they run pessimistic for anyone planning decades ahead. A population average also says nothing about one person. Good health at 58, or a family that routinely makes it into its nineties, and Femi's own dead band is thinner than the chart's, which makes his broke band correspondingly worse. The ONS life expectancy calculator is a reasonable sanity check on where you sit if those tables are yours. If they are not, treat the dead band as borrowed: the UN's World Population Prospects carries survival tables country by country, and where yours runs materially shorter or longer than the British and American average, every figure in that table shifts with it.
The cheapest fix is not money
Switch Femi's plan from fixed withdrawals to guardrails, where spending drops 10% in any year the withdrawal rate has drifted 20% above where it started, and the alive-and-broke band all but disappears. Success reaches 99.5%.
That is not free, and the free-looking version of it is a lie. Across the paths, his average spending settles at £29,622 rather than £33,000, about 10% less, and the model assumes he cuts every single time the rule says to, with no floor under how far spending can fall. A retiree who agrees to the cut in principle and not in practice is running the fixed plan with extra steps. But as a trade, a tenth of the spending in bad years against two years of work and a 10.4% tail, it is not close, and the withdrawal strategies tool is where to see what the rule would have demanded of you historically. It demanded a lot: the post on guardrails versus fixed withdrawals prices the cuts out year by year.
So the order goes: flexibility first, then guaranteed income later in life, then a lower starting rate, and only then more years at work. That is roughly the reverse of the order most people try.
Run your own plan both ways and compare the two numbers. If the failure rate and the alive-and-broke rate come out close together, your risk is landing in years you will very likely be alive for, and it needs fixing. If they are far apart, the thing keeping you at work is not the plan. It is the chart you were shown. The walkthrough of the controls covers the reading; what to do about it is the part only you can answer.