How to read a Monte Carlo retirement result

Median, percentiles, success rate, and what a cone of outcomes is not. How to compare two plans on the same engine without treating a single percentage as a grade.

The parts of the output

A Monte Carlo retirement result is a pile of paths. The success rate is the share that never hit zero before the horizon. The median ending is the middle path’s leftover, not the path you will live. The 10th percentile ending is a bad-but-not-worst leftover. A cone chart is those percentiles through time. None of these is a forecast.

Independent lognormal draws around a mean — what the free Monte Carlo retirement calculator runs — are not the same as Killion’s five-regime engine on the live demo. Different engines, different clustering of bad years, different success rates for the same household. Compare plans inside one engine.

How to compare two plans

Same engine, same seed, same household, change one thing. A 10% spending cut that lifts success from 78% to 91% is a statement. A 91% on a slide with no sibling is a slogan. The probability of success guide is why 100% is usually an expensive lifestyle, not a safer model. The methodology page is how the product engine is built.

Not a grade

90% does not mean a 10% chance of dying broke. It means 10% of modelled paths depleted under those assumptions with no mid-course change. Households cut, work, or reallocate. Read the 10th percentile as “what a bad path still had left,” not as a promise.

Frequently asked questions

What should I look at first in a Monte Carlo result?
The success rate and the 10th-percentile ending, on the same engine and seed as the plan you are comparing it with. The median is the middle path, not the path you will live.
Is a 90% success rate a 10% chance of dying broke?
No. It is 10% of modelled paths depleting before the horizon, with no mid-course change. Households cut spending, work longer or change the mix.