cFIREsim

What it is

cFIREsim — Crowdsourced Financial Independence and Retire Early Simulator — answers one question: if you retired into the worst sequence of returns on record, would the money have lasted? You give it a retirement year, an end year, a portfolio value and a yearly spend, and it replays that plan against every overlapping historical window of that length. A 30-year plan entered today runs 127 cycles, 1871–1900 through 1997–2026, and "success" means the portfolio never touched zero. The default inputs — $1,000,000, $40,000 a year, 75/25 — came back at 93.70%, with the failures clustered on 1966 and 1969 starts.

The modelling underneath is deeper than the front page suggests. Allocation is split across equities, bonds, gold and cash, with an annual fee percentage, optional annual rebalancing and a glide path that walks you to a different target allocation between two years you pick. Eight spending plans are offered: inflation-adjusted, not inflation-adjusted, percent of portfolio, variable spending with a z-value, VPW, Hebeler Autopilot, Guyton-Klinger and variable CAPE — each with an optional floor and ceiling on top. Inflation runs on historical CPI or a flat rate you set. Two Social Security streams, pensions and up to eighty one-off or recurring income and expense adjustments sit alongside, each with its own start year, end year and inflation treatment.

Pricing

Free, with nothing to buy. There is no account, no login, no tier and no donation link anywhere on the site. The only commercial presence is a banner for the author's newer product, FIREproof.

Data & coverage

It ships no market data and makes no forecast. The historical series are Robert Shiller's monthly S&P 500 dataset — prices, dividends, earnings, US CPI and the 10-year Treasury yield — running from 1871, plus gold prices from Kitco added in 2021. Cash is the exception: the author found no usable historical series, so you supply a growth rate, defaulting to 0.25%. Everything is US.

Besides replaying all of history, the Data Method dropdown will restrict the run to a year range, substitute a constant growth rate, or trace a single named cycle. An Investigation option solves the other direction — the largest initial spend that still clears a success rate you choose.

Integrations

None, and that is the design. There is no public API, no login and no saved scenarios. What leaves the page is a CSV of every year of every cycle and a permalink that reloads your inputs, which is the intended way to keep a scenario.

Limitations

• No taxes and no account types. One pot of money, no Roth, no RMDs, no brackets. You are expected to bake tax into your spending number yourself.
• The cycles are not independent trials. 127 overlapping 30-year windows share most of their years; a 93.7% success rate is not a 93.7% probability. There is no Monte Carlo mode.
• US data only. Shiller's S&P 500, US CPI, US Treasuries. No international equity or bond series at any setting, and gold was flat for decades of the 1800s, which flatters gold-heavy allocations.
• Nothing persists. Lose the bookmark and you retype the form.
• Closed source and feature-frozen. The site's changelog ends in September 2022; the open repository is the 2.x JavaScript version whose last code commit was March 2016.

Alternatives

FIRECalc is the older tool this one was built to improve on, with the same historical-cycle method and a plainer interface. Portfolio Visualizer covers similar ground with a Monte Carlo mode and more asset classes. FIREproof, from the same author, is where the missing half lives — federal and state tax, account types, RMDs and Roth conversions — free to try with a $10-a-month Pro tier, and it is the reason cFIREsim itself is unlikely to gain those features.