What "chance of success" really means

A retirement plan's "chance of success" is the share of simulated futures in which your money lasts as long as the plan does. If a plan succeeds in 830 of 1,000 simulated futures, its chance of success is 83%.

That sounds precise. It isn't a forecast, though, and it's easy to read too much into it. This guide explains what the number measures, how the futures are made, and what to look at besides the headline.

What counts as success

In the app, a simulated future counts as a success when your plan can pay every expense and tax bill, every month, until the end of the plan. The plan ends at the life expectancy you set for the last surviving adult.

A future fails the first month your cash and investments can't cover what's due. That includes selling anything you've marked as "sell if needed," such as a second home.

You can change the definition on the Outlook screen. Two stricter options are available:

• Spending stays above a floor. A future also fails if spending ever drops below a share of your planned spending.
• You leave a minimum estate. A future also fails if what's left at the end, in today's dollars, is below an amount you set.

Whatever definition you pick, success is a yes or no for each future. It doesn't say how close a failure was or how much was left over.

Where the simulated futures come from

Markets don't repeat, but they rhyme. The app's default method builds each future from actual historical market returns and inflation, about a century of annual US data.

It works like this:

1. Pick a random year in history.
2. Take that year and the four that followed it, a five-year block, with each year's stock returns, bond returns and inflation kept together.
3. Pick another random block and attach it.
4. Repeat until the future is as long as your plan.

This is called a block bootstrap. Keeping five years together preserves things that random draws often miss. Bad years tend to cluster. High inflation tends to persist. Stocks and bonds sometimes fall together. A future built this way contains real sequences like those, in new combinations.

You can switch methods in the simulation settings on the Outlook screen. The options include random returns from a bell curve, a version with fatter tails for more extreme years, and replaying every historical starting year in order. Each method answers a slightly different question. None is the truth.

A worked example

Lena is 62, single and retiring this year. Her plan spends $70,000 a year in today's dollars. She'll claim Social Security at 67. Her plan runs to 95.

Outlook runs her plan against 1,000 simulated futures:

The figures are invented for illustration. The shape is common, though. Most failures arrive late, after years of below-average returns early in retirement. The 17% isn't 17% chance of being broke at 70.

Two runs with different random futures give slightly different answers. With 1,000 futures, the result can move by a point or two between runs. Small differences between two versions of a plan may be noise. The app compares plans on the same futures, so a change you test is measured against the same markets.

Five ways to read it well

1. It's a rate of shortfall, not a rate of disaster. In a failed future, Lena's money runs short years after a bad start. In real life she'd notice and adjust long before. That's what guardrails are for. See spending guardrails explained.

2. Look at the bad cases as well as the count. The range chart on Outlook shows outcomes by year at several percentiles. The 10th percentile line is the future where only 1 in 10 did worse. If that line stays above zero until late in life, a "failure" usually means a modest late shortfall.

3. Higher isn't always better. A plan with a 99% chance of success may be spending far less than it could. For a household willing to adjust, a result between 75% and 90% can be reasonable. Your own range depends on how much you could cut if needed.

4. The assumptions matter more than the decimal. Returns, inflation, how long you live and what you spend all drive the result. Changing life expectancy by five years can move it more than any investment choice.

5. History is not a forecast. The past century of US markets was, on the whole, a good one. Future returns could be lower. You can test that by lowering return assumptions or by picking a different simulation method and comparing.

Why not use average returns?

A plan that assumes the same return every year has one outcome. It's useful for seeing how money flows, and it's what the Cash flow screen shows. But it hides the main risk in retirement: the order of returns.

Two retirees with the same average return can end up in very different places if one gets the bad years first. Averages can't show that. A simulation can. See sequence-of-returns risk for an example.

How many simulated futures is enough?

More futures make the estimate steadier. The app runs 1,000 by default, and you can choose more. Beyond a few thousand, extra runs rarely change the answer by more than a point.

What matters more is that the futures are realistic, and that you compare plans on the same set. Every comparison in the app, from scenarios to improvements, uses the same simulated markets as your plan.

Frequently asked questions

Why did my chance of success change when I didn't edit anything?

Balances you update, a new calendar year or a change in tax figures all feed the plan. If you pressed "Run again" with a different random seed, the futures themselves change too.

What does "money lasts" include?

Cash, taxable investments, retirement accounts, Roth accounts and health savings accounts all count. Home equity doesn't, unless you planned a sale.

Is 80% good enough?

There's no universal answer. It depends on how flexible your spending is and how you'd respond to a bad decade. Look at the size and timing of the shortfalls, then decide what you're comfortable with.