Calculator methodology

Model scope and units

The FIRE calculator is an annual, deterministic projection. Spending, balances, contributions, and every displayed balance are expressed in today’s dollars. Recurring cash flows marked to keep pace with inflation remain constant in today’s dollars; other recurring cash flows lose purchasing power over time. Ages are integer years. The build phase runs to retirement and the drawdown phase runs to the entered planning age, which defaults to 90.

Inputs are normalised before calculation. Invalid or out-of-range ages, rates, balances, probabilities, and allocations are clamped to the ranges accepted by the calculator. In couple mode, each person’s spending, balances, and contributions are combined for the household calculation.

Australian FIRE, bridge, and detailed plan

The calculator reports three separate milestones. Classic FI is the first age whose projected drawdown reaches the expected super access age with the FIRE number still invested. Bridge-ready additionally requires accessible outside-super assets to pay every modelled cash need before super access. Australian FIRE is the first age when both conditions hold at the same time.

FIRE number = net ongoing portfolio spending ÷ selected safe withdrawal rate

Australian FIRE age = first age with classic FI + an accessible bridge

Only certain, inflation-adjusted recurring income and withdrawals that are active by the later of target retirement or super access and last through the planning horizon change the FIRE number. Non-indexed and temporary cash flows remain in the detailed projection instead of being treated as permanent real offsets. The FIRE number remains a fixed-real-spending benchmark even when a variable drawdown strategy is selected. The separate plan-funded age is the first age whose deterministic cash-flow projection avoids every shortfall through the selected planning age. It is a detailed planning check, not an extra hidden condition in the definition of Australian FIRE.

The entered return is the expected nominal return for the whole portfolio before fees. The entered portfolio allocation is used to weight the stock, bond, and cash fees at every age; an optional glide path can change that weighting. Glide paths are off by default, and allocations are normalised to 100% at each age. A zero-total allocation falls back to 100% stocks.

net nominal return(age) = (1 + entered portfolio return) × (1 − allocation-weighted fees) − 1

real return(age) = (1 + nominal return(age)) ÷ (1 + inflation) − 1

The model does not add hidden asset-class premiums to the entered return. Change the expected portfolio return directly when modelling a different allocation or return outlook.

Accumulation sequence

For each projection year, the model grows existing outside-super and super balances at that year’s real return, adds end-of-year contributions grown by the wage-growth assumption after removing inflation, and then applies expected cash flows. The entered contribution is the first year’s amount in today’s dollars. Entered contributions are treated as savings after living costs, so pre-retirement spending is not deducted again.

Super contributions mean the net amount credited to the account after contributions tax, including employer and personal contributions you expect to receive. The calculator does not deduct contributions tax, apply Division 293, or check contribution caps; adjust the entered amount yourself.

An expense first uses outside-super assets, then accessible super. Any unpaid amount remains visible as unfunded spending instead of being discarded. In couple mode, both people retire together at the shared target FIRE age. The model uses one household age and shared super-access, wage-growth, return, and tax settings.

Bridge and super access

Before the expected super access age, outside-super assets must fund spending, tax, and expenses. Super remains invested but unavailable. The bridge qualifies only when that drawdown reaches the access age with the full FIRE number, based on the selected safe withdrawal rate, still invested across outside-super and super balances. The rate defaults to 3.5%. From the access age the two balances can fund the remaining plan together. Age 60 assumes the user will meet a condition of release; age 65 represents unrestricted age-based access. The calculator stores one shared household access assumption and does not determine legal eligibility.

Drawdown order

The deterministic drawdown starts from the balance immediately before retirement-age cash flows. It applies the event on that birthday, then sets adaptive spending from the opening portfolio for each retirement year. Active recurring income offsets that spending, the real return and tax proxies are applied, withdrawals use outside super before accessible super, and the next birthday’s one-off event closes the year.

Fixed-dollar, percentage, VPW, guardrail, and bounded endowment strategies use the same helper in qualification, deterministic drawdown, and stress testing. VPW divides the current balance across the remaining withdrawal years. The endowment option applies the selected percentage to the current balance and clamps it to the configured real-dollar floor and ceiling; it is not a moving-average strategy.

The tax model is a simplified scenario adjustment and defaults to zero. One proxy taxes positive real outside-super investment growth. The CGT proxy applies the configured effective rate to outside-super sale proceeds and grosses up the sale so the requested net cash and its proxy tax are both funded. The model does not track cost base or transaction lots.

For example, a 20% CGT proxy with a 50% discount gives an effective 10% charge on the whole sale: funding $9,000 of spending requires selling $10,000. It does not mean the whole sale is a taxable capital gain under tax law. Applying both growth tax and the sale proxy can overstate tax, especially where withdrawals mostly return original capital.

Retirement stress test

The stress test uses a separate path-generation and drawdown loop. The default is 3,000 trials for each of three scenarios, using the same planning age as the deterministic projection:

• Random sequence, 60%: return blocks are resampled throughout retirement.
• Early downturn, 20%: the weakest block of up to five years in that trial is moved to retirement.
• Late downturn, 20%: the same block is moved to the end of the horizon.

The three scenarios for a trial use the same return multiset and cash-flow realisation. Moving the same weak block changes sequence without also changing lifetime compound returns. The 60/20/20 result is a planning-weighted stress score, not an empirical probability that either downturn timing will occur.

Market paths are generated before the entered sleeve fees. Each path is copied, then the allocation-weighted fee for each retirement year is compounded exactly once.

net real return(age) = (1 + gross real return) × (1 − allocation-weighted fee(age)) − 1

Each trial evaluates the retirement-year sequence in this order:

1. Apply a realised event on the retirement birthday before the first year starts.
2. Calculate adaptive spending from the opening portfolio and offset it with active recurring income.
3. Apply that year’s return and the configured growth-tax proxy.
4. Withdraw the remaining cash need from outside-super assets, including the CGT proxy, then use super only after the expected access age.
5. Reinvest recurring income above planned spending and apply the next birthday’s one-off event.
6. Mark the trial as failed if any pre-retirement or retirement cash need cannot be funded.

A failed stress path is assigned zero balances. As in deterministic qualification, percentage and VPW paths must meet the entered opening spending target, while endowment paths must meet their configured floor. Success then means that every planned withdrawal is funded through the entered planning age. The reported success rate is the weighted average of the three scenario success rates, not the proportion of paths from one pooled sample.

The 60/20/20 weights are chosen stress scenarios, not estimated frequencies of real market conditions. The result is a weighted scenario score, not a calibrated probability of retirement success. More simulation paths reduce sampling noise but do not remove uncertainty in returns, tax, spending, or the scenario weights.

Historical analogue mode

When at least 12 consecutive aligned monthly returns are available, the stress test uses dividend-adjusted SPY and AGG prices from Financial Modeling Prep. The mixes are conservative (40% SPY / 60% AGG), balanced (60% / 40%), and growth (80% / 20%). Calendar gaps break a run rather than being treated as one month.

Overlapping return windows share observations and are not independent years of evidence. A long resampled simulation does not create additional market history; inspect the reported source period before interpreting the score. These US-dollar proxies do not model an Australian investor's currency exposure.

The source series remains nominal. The stress test converts each month using the selected planning-inflation assumption; it does not claim to use observed historical CPI:

scenario real return = (1 + historical nominal return) ÷ (1 + planning inflation) − 1

Annual windows use 12 consecutive months and never wrap the newest observation into the oldest. Trials resample valid blocks from the longest consecutive aligned run. The provider, symbols, period, observation count, nominal return basis, and warnings are included in the result.

Assumption fallback mode

If fewer than 12 monthly returns are available, no historical series is used. The model draws annual returns from:

gross annual real return = ((1 + entered nominal return) ÷ (1 + inflation) − 1) + normal random draw × annual volatility

The fallback paths start from the gross Fisher real return. The model then compounds each year’s age-specific fee exactly once. Annual volatility defaults to 15% and is bounded by the model. Early and late scenarios move the same path’s weakest block rather than adding a separate return penalty. Annual returns are capped at −70% and +70%. The fallback is labelled in the result even if the caller supplied an unusable historical-data report.

The random generator uses a fixed default seed of 42. A supplied seed changes the paths; the same inputs and seed reproduce the same simulation paths.

Cash-flow treatment

The displayed deterministic projection always includes expenses and withdrawals. Optional windfalls and recurring income appear at their probability-weighted expected value in charts, but Australian FIRE, bridge-ready, and plan-funded qualification counts positive cash flows only when their probability is 100%. Recurring cash flows are included over their active age range, with non-inflation-adjusted amounts deflated from the current age into today’s dollars.

The stress test includes expenses and withdrawals in every path. It samples optional windfalls and recurring income once per trial, including events before retirement when calculating that trial’s starting balance. The same realisation is used for its random, early, and late return sequences.

Success metrics and revised SWR

Scenario results include success rate, negative-return rate, and ending-balance quantiles. The distribution shown in the interface uses the scenario weights when assigning ending balances to buckets.

The revised withdrawal rate is a grid search. The calculator tests rates from 0.25% to 8.25% in 0.25 percentage-point increments against each trial’s realised retirement starting balance and returns the highest rate whose weighted success rate meets the selected target. It is not a continuous optimisation and returns no rate when there is no positive starting capital.

Longevity and Die With Zero view

The longevity view copies the explicit target retirement age, balances, portfolio assumptions, fees, tax proxies, withdrawal strategy, and cash-flow schedule from the FIRE plan. That local copy stays fixed while upstream inputs change and updates only when you choose to sync it. Outside-super and super balances remain separate, and super cannot fund cash needs before the expected access age.

Each simulation samples nominal returns around the entered portfolio assumption, compounds fees, converts the result to today’s dollars, and uses the same annual cash-settlement order and opening-spending target as the FIRE plan. Optional pre-retirement and retirement cash flows are sampled once per trial. Income above planned spending remains invested.

The chart horizon controls the displayed age-by-age distribution and the funded-at-end metric. “Run out before death” and “die with more than you started” are lifetime-weighted measures through age 130. Turning on mortality adds the probability of death to the stacked chart; it does not change the underlying portfolio paths.

Mortality uses the national male and female tables in ABS Life expectancy, 2022-2024, Table 1, released 11 November 2025. Both survival curves start at 100% at the selected retirement age. We average their subsequent survival probabilities equally: combined survival is 0.5 times male conditional survival plus 0.5 times female conditional survival. We do not average annual death rates, which would produce a different result.

This equal-weight average is derived by the app, not published by the ABS as a combined population table. It is not weighted to the population composition at retirement age and is not a personal forecast. These are period life tables: observed death rates stay fixed, with no assumed future mortality improvements or personal health adjustments. The data is bundled locally; no demographic information is requested or sent to a server.

The published workbook provides annual death probabilities through exact age 100. For later ages, each table's age-100 annual probability is held constant. This tail is an app extrapolation, not ABS data. The model stops at age 130; remaining survivors contribute their terminal portfolio value to the ending-wealth lifetime measure, and failures after 130 are not modeled. The chart does not force survival to zero at either age 100 or 130.

Funded outcomes allow the selected strategy and flexibility rules to reduce spending. The separate no-cut share counts funded paths that never selected less than the entered annual budget; it is not a second simulation with cuts disabled. The spending-cut share includes reductions selected by adaptive strategies as well as the flexibility setting. These metrics describe outcomes through the chart horizon.

Capital gains tax projection

The CGT calculator models eligible shares or ETF units held personally by an Australian-resident individual from 1 July 2027 and sold on or after each holding anniversary. Price growth and unfranked cash dividends are calculated independently from the opening balance. Dividend income tax is paid annually at the selected flat, levy-exclusive rate, and after-tax dividends are manually reinvested at year-end as separate cost-base lots.

Old taxable gain = max(0, (sale value - nominal cost base) x 50%)
Old estimated sale tax = old taxable gain x flat tax rate

New real gain = max(0, sale value - indexed cost base)
Base new sale tax = new real gain x flat tax rate
Minimum-tax proxy top-up = floor(max(0, (new real gain x 30%) - base new sale tax))
New estimated sale tax = base new sale tax + minimum-tax proxy top-up

Total tax = sale tax + cumulative dividend tax
Effective tax rate = total tax / total pre-tax investment profit

For the new-rules comparison, each cost-base lot is compounded by the entered constant CPI assumption from its investment year to the sale year. A year-end reinvestment receives no indexation in that same year. The indexed base is limited so indexation can eliminate a nominal gain but cannot create or increase a capital loss.

Total pre-tax investment profit is the sale value plus cumulative dividend tax already paid, less the initial investment. This reconciles the dividend tax removed before reinvestment. Effective tax is shown as zero when total pre-tax profit is not positive.

This is an educational scenario based on Acts No. 49 and 50 of 2026 and the official Budget explainer. Actual calculations use quarterly ABS CPI indices and personal tax-return information. The model uses a flat-rate proxy for the statutory minimum-tax comparison and lets users turn that proxy off for listed income-support exemptions. It excludes pre-1 July 2027 holdings, capital losses, franking credits, foreign tax offsets, ETF distribution components, progressive brackets, Medicare levy, tax offsets, brokerage, superannuation and specialist CGT concessions.

Rent versus buy

The housing comparison uses the same starting savings and monthly housing budget for both paths. Mortgage repayments reduce the buyer's loan balance; whichever path has lower housing costs invests the difference. The buying result deducts selling costs, and investment balances allow for the selected distribution and sale-tax proxies. Results can be shown in future dollars or today's dollars.

Stamp duty after concessions, grants, LMI, and FHSS funds are entered assumptions. Buying uses FHSS funds once after the entered release tax, while renting retains them in super. Insufficient upfront funds block the projection. The calculator does not assess borrowing capacity, lender eligibility, personal tax, or future policy changes. Use current official figures for your location and circumstances.

The headline rounds the estimated difference; detailed tables retain calculated values. The growth comparison changes only yearly home price growth, usually two percentage points below and above the entered assumption, within model limits. These are illustrative scenarios, not probabilities or a forecast range. Monthly costs show first-year averages, separating loan interest from the amount that reduces the debt.

Scenarios are saved on this device. Shared links use the same compressed URL-fragment approach as the FIRE calculator and include all inputs and the dollar-display setting. Reddit summaries are prepared in your browser for you to copy and post manually.

Important limitations

• Historical analogue paths are resampled scenarios, not forecasts or a replay of the assets you personally hold.
• The SPY/AGG analogue is US-dollar based. It does not capture Australian tax treatment, franking credits, currency hedging, sequence behaviour for every asset class, or the exact portfolio held by the user.
• Portfolio fees are represented as simple annual sleeve deductions. Drawdown tax is represented by the simplified rates described above. The model does not calculate personal income tax, account fees, brokerage, tax offsets, or realised gains using transaction lots.
• The model does not include Age Pension eligibility, contribution caps, insurance, healthcare costs, or future changes in legislation.
• Expected super access age is a user-entered planning assumption, not an eligibility determination.
• Success means the model funded the selected spending rules through the planning age; adaptive rules can reduce the budget. It does not measure lifestyle suitability or advice appropriateness.
• Short market histories can be constrained by the data provider. The fallback model is assumption-driven and should not be read as historical evidence.
• All results are educational estimates and can differ materially from real outcomes.