How to run a FIRE projection on your actual portfolio, not a generic calculator

Most FIRE calculators ask for three numbers: current savings, monthly contribution, and an expected annual return. They multiply it out and give you a year. It feels precise. It is not, and the reason matters if you are making real decisions about when you can stop working.

A better projection does not give you one year. It gives you a range, and it tells you how wide that range is. Here is why, and how to run one on your own numbers.

A calculator that assumes 7% every year is modeling a world that does not exist. Real returns arrive as a sequence of good and bad years in an order you cannot predict. Two portfolios with the same average return can end up far apart depending on when the bad years land, especially once you start withdrawing. This is sequence-of-returns risk, and a single-number calculator is blind to it by construction.

The result is false confidence. The calculator says "age 52" with the authority of a decimal point, when the honest answer is "somewhere in a range, and here is how likely each part of the range is."

A Monte Carlo projection runs your plan thousands of times, each with a different randomized sequence of returns drawn from an expected return and a volatility you specify. Instead of one endpoint you get a distribution:

Now the answer is useful. "If markets behave like my assumptions, I hit my number around year 12 in the median case, but a bad sequence pushes it to year 16, and I should plan for that" is a decision you can actually act on. A single "year 12" is not.

The other honest part: the projection is only as good as the assumptions you feed it. Expected return and volatility are inputs you supply, not truths handed down by the tool. A good projection makes you state them out loud and shows how sensitive the answer is to them, rather than hiding a 7% behind a clean number.

Generic calculators use a generic return because they do not know what you own. But your actual portfolio has its own history. A projection tied to your real holdings can start from your measured return and volatility instead of a round guess, and it can run several assumption sets side by side, a conservative one, a base case, an optimistic one, so you see the fan of outcomes rather than a single line.

To do this you need two things: your holdings and savings rate in one place, and a projection engine that takes your assumptions rather than baking in its own.

Opula is a connector for Claude or ChatGPT. Once your holdings, balances, and cash flow are in, you can ask it to project your net worth at any horizon, from one month to several decades, and it runs a Monte Carlo simulation, 1000 iterations per scenario by default, over the assumptions you give it.

Two details keep it honest. First, it does not invent return or volatility numbers, you supply them, and you can pull your own measured return and volatility from its risk summary to use as one scenario. Second, every scenario in a single call shares the same randomized shock sequence, so when you compare a 6% assumption against a 14% one, the difference reflects the assumptions, not random noise between runs.

It is a thought-experiment tool, clearly separated from the diagnostic reads. It tells you how a plan plays out under stated assumptions, discloses those assumptions, and leaves the judgment to you. That is the right shape for a FIRE projection: not a false single year, but an honest range you can plan around.