What does an 85% retirement success probability mean?

It means that in 2,000 simulated futures, each with a different randomly drawn sequence of market returns and inflation, the money lasted to age 90 in about 1,700 of them and ran short in about 300. It is a count of simulated outcomes, not a forecast of yours.

Is a 95% success probability better than 85%?

Not automatically. Pushing a plan from 85% to 95% almost always means spending less every year for decades to insure against futures that mostly do not arrive. That is a real trade, and for many people the 85% plan with more spending in it is the better life.

Why does my success probability change so much when I change my spending?

Because spending is the input the plan is most sensitive to. The same portfolio at the same retirement age can move from a comfortable probability to a poor one across a $40,000 range of annual spending — often a swing of more than 50 percentage points.

What does a success probability not account for?

Long-term care costs, your specific tax situation, a market that behaves unlike any historical pattern, and your own behaviour in a downturn. A simulated retiree keeps drawing calmly through a 40% crash. Real ones often do not.

What a retirement "success probability" actually means

An 85% success probability is not a forecast, a grade, or a promise. It is a count of how many simulated futures your money outlasted — and knowing what it counts changes what you should do about it.

Every retirement calculator eventually shows you a percentage. Ours does too. It is the single most misread number in retirement planning, and most of the misreading comes from one assumption: that it is a prediction about you. It is not. It is a count.

What the number actually counts

To produce it, the model builds 2,000 separate futures. In each one it draws a different random sequence of annual market returns and inflation rates, applies them year by year to your balance while subtracting what you spend, and checks whether the money is still there at 90. The percentage is simply how many of those 2,000 futures ended with money left.

So an 85% is not “you are 85% ready” or “you have a B grade.” It is: in roughly 1,700 of 2,000 simulated futures this plan survived, and in roughly 300 it did not. The 300 are not hypothetical in any meaningful sense — they are ordinary sequences of bad luck, mostly involving poor returns early on.

This is also why the number is not a forecast. You will live through exactly one future, and it will not be the average of the 2,000. The percentage tells you how much room for bad luck the plan has, which is a different and more useful thing than a prediction.

Why 95% is not obviously better than 85%

This is where the number does the most damage. Chasing a higher percentage feels like prudence, and sometimes it is. But there are only so many levers, and the usual one is spending less — every year, for thirty years.

A plan that never fails in simulation is often a plan that underspends for three decades to insure against futures that never arrive. The percentage cannot see that cost, because a simulated retiree does not mind a smaller life. You will. If your number is already high, the more interesting question is usually whether you could safely spend more, not less.

There is a floor to this reasoning, though. Below roughly 60%, more spending is not a live option and the trade is a different one altogether: a later date, a lower spending floor, or part-time income early on.

Why the number moves so much with spending

Of all the inputs, spending is the one the answer is most sensitive to — more than the retirement age, and much more than a percentage point of assumed return. This is easy to see rather than argue about. The can-I-retire pages run every combination at three spending levels, and in some cells the probability falls from near-certain to near-hopeless across a $40,000 range of annual spending.

Which is the real reason any single retirement number deserves suspicion. “Can I retire at 62 with $500,000?” has no one answer, and anyone who gives you one has quietly picked a spending level on your behalf.

What the number cannot see

A success probability is only as honest as the assumptions under it, and ours are stated in full on the methodology page. The important omissions are worth naming here:

Long-term care. Not modelled at all. It is the single largest uninsured risk most retirees carry.
Your tax situation. The model works in pre-tax balances and does not know your bracket, your state, or the order you will draw from your accounts.
Your own behaviour. The simulated retiree keeps drawing calmly through a 40% crash. Real ones sell, or stop spending in the year that mattered most.
A market unlike any we have seen. The return and inflation assumptions are drawn from long-run history. History is evidence, not a guarantee about the next thirty years.

None of that makes the number useless. It makes it a measurement with stated limits, which is the only kind worth having. Use it to compare one plan against another and to see which levers move it — not as a verdict on whether you are going to be fine.