Almost every FIRE spreadsheet I have ever seen makes the same assumption: markets return 7 percent, every year, in a smooth line, forever. Then the person building the spreadsheet quits their job at 43, hits a two-year drawdown in year one, and spends the next decade watching their withdrawal rate rise into the danger zone they never planned for.
The problem is not the math. The problem is that a single average return number is not a plan. A plan has to survive the years that look nothing like the average, especially the years right after you stop earning a paycheck. This piece walks through how to pressure-test a FIRE number the way an insurance actuary would, using free tools you already have access to, and how to translate what you see into concrete adjustments to your savings rate, your target number, or your quit date.
Photo by Vito Goričan on Pexels
Why the average return is the least useful number in your plan
If you assume 7 percent annual growth and a 4 percent withdrawal rate, your money technically lasts forever. That is the calculation that got the 4 percent rule famous in the first place, based on the Trinity Study done at Trinity University in 1998. The study looked at historical rolling 30-year windows and found that a 4 percent initial withdrawal rate, adjusted for inflation, survived almost all of them.
The catch is right there in the phrase "almost all of them." The windows that failed had one thing in common: a bad first decade. When you start retirement with a big drawdown, you sell shares while they are cheap to cover living expenses, and there are fewer shares left to compound when the market eventually recovers. This is called sequence-of-returns risk, and it is the single largest failure mode for early retirees.
An average tells you what happens if the future looks like the mean of the past. Sequence risk tells you what happens if the future has bad news first and good news later. Those two scenarios can produce identical average returns and completely different portfolio outcomes. If you plan for the first and get the second, you find out at age 55.
What a stress test actually is
A stress test replaces the single "7 percent average" input with a range of plausible sequences. Some sequences are historical: what would have happened if you had retired in 1966, or 2000, or right before the 1929 crash. Some are randomly generated inside plausible bounds, which is what a Monte Carlo simulation does. Either way, the output is not one number. It is a probability: given this savings rate, this portfolio, and this withdrawal plan, what percentage of simulated futures leave you with money at age 90.
A well-built FIRE calculator by EvvyTools runs the Monte Carlo pass by default and shows you the distribution instead of hiding it. The number you actually want to look at is not the median outcome. It is the 10th percentile, sometimes called the "bad decile" outcome. That is the tail scenario your plan needs to survive, because you only get one retirement.
The three FIRE variants and why the stress test matters more for the leaner ones
The FIRE community usually distinguishes three flavors:
- Lean FIRE targets a portfolio that supports around 25,000 to 40,000 dollars a year of spending. Roughly a 625,000 to one million dollar target using the 4 percent rule.
- Regular FIRE targets middle-class spending, roughly 50,000 to 80,000 a year, so a 1.25 to two million dollar target.
- Fat FIRE targets 100,000 or more per year, well above two and a half million.
Stress testing matters most for lean FIRE and least for fat FIRE, and the reason is not intuitive. Lean plans have less slack. If a bad decade cuts your portfolio by 30 percent, a fat FIRE retiree can cut discretionary spending back to lean FIRE numbers and ride it out. A lean FIRE retiree is already at the floor. Their only lever is going back to work, which usually happens in the worst possible labor market, because those tend to arrive with bad markets.
The stress test does not tell you to abandon lean FIRE. It tells you that lean FIRE with zero margin for error is a different plan than lean FIRE with 15 percent margin, and the second plan is what actually gets you to age 90 without a phone call to a former employer.
Photo by Anna Tarazevich on Pexels
The four inputs that move the needle
When you run the stress test, you will discover that four inputs matter far more than the rest. Everything else is decoration.
Savings rate. Not the dollar amount, the percentage of take-home pay. A 25 percent savings rate gets you to FIRE in roughly 32 years. A 50 percent savings rate gets you there in roughly 17. The relationship is nonlinear and steep, which is why Mr. Money Mustache built an entire blog around that single insight. Bumping your savings rate from 30 to 40 percent buys you almost a decade off your working life.
Withdrawal rate. Every full percentage point matters. 4 percent has around a 95 percent success rate over 30-year windows. 5 percent drops to around 75 percent. 3.5 percent is closer to 99 percent. If you plan to retire before 55 and expect a 40 or 50-year horizon rather than 30, most modern research suggests starting at 3.25 to 3.5 percent, not 4. See the ongoing work at Early Retirement Now for the deep math.
Asset allocation. A 100 percent stock portfolio has the highest expected return and the worst sequence risk. A 60/40 stock-bond portfolio gives up around a percentage point of average return in exchange for a much softer bad-decile outcome. For early retirees specifically, the "rising equity glidepath" idea, explored in depth by Michael Kitces, suggests starting retirement with more bonds and shifting toward stocks over time, which counterintuitively reduces sequence risk. The Bogleheads community has thousands of threads working through the tradeoffs in practical detail.
Inflation assumption. Most people plug in 3 percent because it feels normal. Real US inflation has ranged from -1 to 14 percent over the last century. Your stress test should include at least one scenario that assumes inflation runs at 5 to 6 percent for a decade, because that scenario has historical precedent and it eats a lean FIRE plan alive.
Running the actual test: five sequences that matter
You do not need to run a thousand simulations to get useful information. Five specific scenarios cover most of the tail risk.
- The 1966 sequence. Retire the year the market peaks before a decade of high inflation and mediocre real returns. Almost every early retirement calculator failure case looks like this scenario replayed.
- The 2000 sequence. Retire into the dot-com bust, followed by 2008. Two major drawdowns in the first decade with a weak recovery in between.
- The 2022 sequence. Retire into a year where both stocks and bonds fall together, breaking the classic 60/40 hedge. This one is recent enough that most calculators do not weight it heavily yet.
- A Monte Carlo bad-decile pull. Whatever your calculator produces at the 10th percentile. If your plan works at the 10th percentile, it works.
- A flat-real-returns decade. Assume equities return exactly inflation for 10 years, then resume normal growth. This one is unusually punishing and it is what actually happened from 2000 to 2010 on a real basis.
If your plan survives at least four of these five, you have a plan. If it survives all five with slack, you can consider pulling the quit date earlier or downgrading the "one more year" instinct that plagues everyone in the last 24 months.
Translating results into decisions
Here is where most stress tests fail to become useful. Someone runs the calculator, sees "78 percent success rate at the 10th percentile," and then does nothing with the number.
If your plan hits between 60 and 80 percent at the bad decile, you have three usable levers before "work longer." First, drop your baseline withdrawal rate by 0.5 percent. That is often enough to move the needle 10 to 15 points. Second, plan for a "guardrails" spending rule, where discretionary spending automatically drops 10 percent if the portfolio falls below a threshold. Guyton and Klinger's guardrails work is the reference here, and the math is stronger than the 4 percent rule for early retirees. Third, plan for one to three years of part-time work early in retirement, specifically to avoid selling shares during a drawdown. Even 15,000 dollars a year of income in years one through three of a bad sequence dramatically improves outcomes.
If your plan hits above 90 percent at the 10th percentile, you are probably overtaxing yourself with additional savings. Consider whether one more year of work is actually buying you anything, or whether it is just anxiety wearing the costume of prudence.
If your plan hits below 60 percent at the 10th percentile, the honest answer is that you do not have a plan yet. You have a target. That is fine. Adjust the savings rate, adjust the target, or adjust the timeline. Pick two.
Photo by Patrick Gamelkoorn on Pexels
What to actually do this week
The calendar test for whether this article was useful is simple. Sometime in the next week, open your favorite retirement planning tools and run these numbers:
- Enter your current portfolio balance, your monthly savings contribution, and your target retirement spending in today's dollars.
- Turn on the Monte Carlo simulation if the tool has one.
- Look specifically at the 10th percentile outcome, not the median.
- Then try the same run with a 0.5 percent lower withdrawal rate and see what happens to the bad-decile number.
If you have never done this exercise, the first pass usually surprises people in one direction or the other. Some people discover they are much further along than they thought and their real problem is calibrating what they want to do with an early retirement. Others discover their timeline is aspirational and needs an honest adjustment.
Either answer is more useful than continuing to plan around a 7 percent average that no ten-year window in history has actually delivered exactly.
The one habit that separates working plans from wishful ones
The people who actually reach FIRE and stay retired share one habit: they re-run their numbers annually and update their assumptions when the world changes. When bond yields go from 1.5 percent to 5 percent, that changes the math. When healthcare cost inflation runs at 6 percent instead of 4, that changes the math. When you have a real spending year and it turns out to be 20 percent higher than you projected because you underestimated travel, that changes the math.
A stress test is not something you do once. It is a checkup you do every 12 months, using the same free FIRE calculator from EvvyTools you started with, comparing this year's inputs to last year's. The people who quit at 45 and stay retired are the ones who noticed at 47 that something was drifting and made a 3 percent adjustment. The ones who go back to work are usually the ones who assumed the plan they built at 40 still worked at 50 without checking.
The best time to stress-test your plan was five years ago. The second-best time is right now, before the next bad decade tells you the hard way.
