A simple retirement projection takes an average return, say 6% a year, and applies it every year until the plan ends. It's easy to follow, and it's wrong in a way that matters: markets don't return their average every year, and the order in which good and bad years arrive can make or break a retirement.

A Monte Carlo simulation deals with that by running your plan many times, each through a different possible future.

What the simulation does

Picture 1,000 copies of your plan. Each one lives through its own sequence of market years: some start with a boom, some with a crash, most with something in between. In every year of every copy, your contributions, withdrawals, taxes and Social Security are worked out exactly as they would be in the plain projection. Only the returns differ.

The returns are drawn at random, but not wildly: each year's stock, bond and cash returns come from a distribution with the average and the volatility you assume, and they move together the way those assets tend to. The RetirementPace planner runs 1,000 of these markets every time you change a number.

Reading the success rate

The success rate is the share of those 1,000 futures in which the money lasted the whole plan. A plan that succeeds in 850 of them shows 85%.

That isn't the probability that you will be fine. It's how often this plan, unchanged, made it through the futures the model can imagine. Real people don't leave their plans unchanged: they spend a little less after a bad year, work a year longer, or claim Social Security later. A success rate is best read as a measure of how much room your plan has, and how much it depends on luck.

Reading the fan chart

The fan chart shows your total savings over time across all the simulated futures:

  • the middle line is the median, the future where half did better and half did worse;
  • the inner band holds the middle half of the futures (the 25th to 75th percentiles);
  • the outer band holds the middle 80% (the 10th to 90th percentiles).

A fan that widens quickly means the plan is exposed to markets. The bottom edge is the one to watch: it shows how the unlucky futures go.

You'll notice the median sits below the plain projection's line. That isn't an error. Ups and downs eat into compound growth (a 50% loss needs a 100% gain to recover), so the typical outcome of a volatile portfolio is less than its average return suggests.

Why 100% isn't the goal

Pushing a plan to a 100% success rate usually means saving far more or spending far less than you'll turn out to need, to protect against the very worst futures. How high is high enough depends on how flexible you can be: a plan you'd adjust after a bad stretch can live with a lower rate than one with no room to change. A lower rate isn't a verdict; it's a sign the plan has less room for bad luck.

What the model assumes

Every simulation is only as good as its assumptions. This one:

  • treats each year's returns as independent of the last, so it has no booms that build on themselves and no recoveries that follow crashes;
  • holds inflation at the rate you choose, rather than letting it vary;
  • uses the average returns and volatility you set, which no one knows for certain.

Historical studies take the opposite approach: they replay real sequences of past returns, as William Bengen did in the research behind the 4% rule. History offers a limited number of 30-year stretches, they overlap, and the future won't repeat them exactly. Simulations can test many more futures, but only the kind the model allows. Neither is a forecast; both are ways to stress-test a plan.

Pair it with pace

A success rate is a snapshot of a plan. Once you're living the plan, the useful question becomes whether you're keeping up with it. Comparing your real balances each month with the path the plan expected, and seeing the success rate move as you do, turns the simulation from a one-time verdict into a running check.

Try it

The retirement calculator runs your plan through 1,000 simulated markets and draws the fan chart. The safe withdrawal rate calculator shows how often a given withdrawal rate lasts 30 years. The methodology has every formula.