Glossary

Different accounts have different tax treatment on withdrawal: savings is post-tax cash (no tax), brokerage has long-term capital gains rates, 401(k) and Traditional IRA are ordinary income, HSA is tax-free for medical (or ordinary income post-65 for non-medical), Roth is tax-free for qualified distributions. The conventional order — and the one the engine uses — draws from the lowest-tax accounts first and keeps tax-free growth (Roth) for last. It is common guidance, but it is not always the lowest-tax order over a lifetime: leaving pre-tax money untouched can mean larger required withdrawals at higher rates later. The Withdrawal Order card runs your plan the other ways and shows the result. RMDs (starting at 73 or 75, by birth year) force minimum 401(k) withdrawals regardless of order.

The single biggest predictor of retirement tax outcomes isn't how much you save — it's WHERE you save it. A portfolio that's 100% pre-tax (e.g., all 401k) means every dollar you withdraw is taxed as ordinary income; RMDs (starting at 73 or 75, by birth year) can push you into higher brackets; and you have no flexibility year-to-year. A diversified mix (some pre-tax, some Roth, some taxable brokerage) lets you choose how much taxable income to recognize each year — qualify for ACA subsidies, fund Roth conversions, harvest gains at 0% LTCG, manage IRMAA. The flexibility is worth real money over a 30-year retirement.

During your bridge years (between retirement and Social Security), your taxable income drops significantly. This creates a window to convert pre-tax 401(k) or Traditional IRA money to Roth at a lower marginal tax rate than you'd pay later. The 'sweet spot' is a CEILING rather than a recommendation: the largest amount you could convert without spilling into the next federal bracket, crossing the IRMAA Medicare surcharge thresholds, or losing an ACA premium subsidy. It is measured on the engine's own projected ordinary income for each bridge year and reported for the year with the least room, because one flat annual amount has to fit in every year. Being able to convert that much does not mean you should — on some plans converting nothing leaves the most behind, which is what the Roth Conversion Planner tests. Each year of bridge income shrinks the window. The cheapest years end when Social Security starts — but conversions don't have to: the 'Convert until age' field (Inputs → Tax Assumptions & Strategy, beside the conversion amount) keeps them running past your SS claim toward RMD age, the standard planning horizon for pre-tax-heavy plans. Post-SS conversions cost more per dollar (SS fills part of the bracket, and past 65 they count toward IRMAA thresholds — the engine prices both). The window always closes at RMD age: once required distributions start, converting alongside them is a different maneuver we don't model. The sweet-spot recommendation itself is sized on your bridge years; the Roth Strategies card's year-by-year table prices any extended years honestly.

One conservative simplification through 2028: the temporary $6,000-per-person senior deduction (tax years 2025–2028, for people 65+) is charged in the year-by-year projection but deliberately not counted when sizing conversion room. A 65+ converter in those years has slightly more room than shown — and inside the deduction's income phase-out range, each converted dollar also claws back six to twelve cents of deduction, a real cost the room estimate doesn't price. The window is sized conservative rather than pretend precision on a deduction that expires after 2028.

The IRS requires you to start withdrawing from pre-tax accounts at age 73 (or 75 if you were born in 1960 or later), whether you need the money or not. The amount is calculated each year from a uniform lifetime table — divide your account balance by a factor that decreases as you age. RMDs are taxed as ordinary income, and if your pre-tax balance is large by 73, they can push you into a higher bracket and trigger Medicare IRMAA surcharges. This is why tax diversification matters now — a mix of pre-tax, Roth, and taxable gives you flexibility to manage your taxable income in retirement.

A QCD lets you give directly from a Traditional IRA to a qualified charity once you’re 70½. What makes it powerful is the tax treatment: the gift is EXCLUDED from your income entirely — it’s not a deduction, it simply never shows up as taxable income. It also COUNTS toward your Required Minimum Distribution, so you can satisfy your RMD without the forced taxable withdrawal. Because your reported income is lower, less of your Social Security is taxed and you can stay under the Medicare IRMAA surcharge cliffs — the savings the wealthier, charitably-inclined retiree cares most about. The trade-off is honest: the donated money leaves your estate, so it lowers your projected legacy. A QCD must come from an IRA (not a 401(k)) — if your pre-tax savings are mostly in a 401(k), roll to an IRA first. The annual limit is about $111,000 (indexed). The engine models it both ways: enter a fixed annual amount, or check “give my entire RMD to charity” to neutralize your RMD’s tax + IRMAA impact each year.

Long-term capital gains are taxed at 0% federally while your total taxable income stays under the 0% bracket's top. In low-income years — typically between retirement and Social Security — you can sell appreciated holdings in a taxable brokerage account and rebuy them right away: the gain is realized at 0%, your cost basis steps up, and that gain is never taxed later. When you turn on harvesting, the engine checks each year how much room is left under the 0% bracket and how much unrealized gain you hold, and harvests whatever fits. Two costs are real and are charged: most states tax capital gains with no 0% bracket, and harvested gains count toward the income that sets your marketplace health-insurance subsidy (and, once you claim, how much of your Social Security is taxable). The Roth Conversion Planner tests harvesting alongside conversions, because both use the same low-income room.

Lump sums each get different tax treatment based on real-world IRS rules. Home sale: Section 121 exclusion ($250K single / $500K married) applies to taxable gain on a primary residence; the excess is taxed at 15% federal LTCG plus your state's rate. Inheritance: generally not subject to income tax under stepped-up basis rules — the recipient's basis is reset to fair market value at death, so liquidating immediately produces zero capital gain. Other lump sums (severance, deferred comp, lottery, business sale): treated as a generic windfall routed at face value, so enter your expected after-tax amount. What lands in your brokerage is the after-tax net, not the gross.

Most calculators ask for a single state tax rate and apply it to every dollar of retirement income. State tax codes are far more varied than that. Pennsylvania doesn't tax retirement income at all. Hawaii excludes employer pensions but still taxes 401(k) withdrawals. Five states tax Social Security at typical retirement incomes. Twenty-one states exclude part of retirement income once you reach a set age, usually up to a dollar cap and often only below an income limit — Georgia, for example, excludes up to $65,000 per person from 65, and New Jersey up to $100,000 for a couple under its income limits. Nine states apply preferential rates to long-term capital gains. The engine models 51 jurisdictions individually using their actual structure — pick your state in Inputs and the engine applies the right rules year-by-year through your entire projection. Same plan can land $1M+ different at life expectancy purely from where you live.

Tax-advantaged retirement accounts (401(k), Traditional IRA) come with a string attached: withdraw before age 59½ and the IRS adds a 10% penalty on top of the ordinary income tax you already owe. It exists to discourage tapping retirement money early — which is exactly what early retirees (the FIRE crowd) need to do.

WHY IT MATTERS — and where it bites. The engine draws your accounts in tax-smart order: bridge reserve → savings → taxable brokerage → 401(k)/IRA → Roth. Taxable accounts come first and carry no penalty, so a well-funded early-retirement plan with a real taxable/brokerage bridge to 59½ never triggers it. It only fires when those buckets run out before 59½ and you're forced into the 401(k) early — i.e., on MARGINAL early-retirement plans. That's the dangerous case: without modeling the penalty, the plans that shouldn't feel feasible look feasible.

THE LEGAL WORKAROUNDS — serious early retirees avoid the penalty through one of three IRS-sanctioned paths: (1) 72(t) / SEPP — commit to 'substantially equal periodic payments' for 5 years or until 59½, whichever is longer; (2) Rule of 55 — penalty-free 401(k) withdrawals if you separate from that employer at age 55 or later (current-employer plan only, not IRAs or old 401(k)s); (3) Roth conversion ladder — convert 401(k) to Roth, wait the 5-year seasoning period, then withdraw the converted principal penalty-free. And one account type never needs a workaround: a governmental 457(b) — common for state and local government workers — has no pre-59½ penalty on money contributed to it once you separate from that employer, at any age (amounts rolled INTO a 457(b) from a 401(k) or IRA keep their penalty character, and rolling 457(b) money OUT to an IRA gives up the exemption).

IN THIS CALCULATOR — by default the 10% penalty applies to any pre-59½ 401(k)/IRA draw (the honest floor). For plans retiring before 60, a 'Penalty-free access before 59½' toggle appears under Inputs → Tax Assumptions & Strategy: turn it on to assert you'll execute one of the three strategies (or that your pre-tax balance is governmental 457(b) money), and the penalty is removed. When the penalty applies, it's charged inside the year's Estimated Tax and itemized as an 'of which: early-withdrawal penalty' note in the projection row's detail, so the math reconciles.

WHAT THIS DOESN'T MODEL — the specific mechanism (72(t) payment schedule, Rule-of-55 employer scoping, the ladder's 5-year seasoning); the toggle abstracts them all. Roth withdrawals aren't penalized (Roth is drawn last, and real Roth contributions are penalty-free at any age). State-level early-withdrawal penalties (a few states add their own) aren't modeled — federal only.

The SECURE 2.0 Act (passed in 2022) created a special "super catch-up" contribution window for people aged 60-63. During those four years, you can contribute up to $11,250/yr to your 401(k) instead of the usual $8,000 catch-up — an extra $3,250/yr. If you're in or approaching that window, this is one of the highest-impact moves available to you. It's a use-it-or-lose-it opportunity: once you turn 64, the limit drops back to the standard catch-up amount. Worth coordinating with your HR/benefits team to make sure your plan administrator is supporting the increased limit.

Social Security's spousal benefit tops up a lower earner: once the higher earner has filed, the lower earner receives their own reduced retirement benefit PLUS the excess of half the higher earner's full-retirement-age benefit over their own — each part reduced on its own schedule. This is automatic for everyone in the 45–58 demographic (deemed filing has been mandatory since the 2015 rules), so the engine applies it for households without a toggle. Modeling it matters: without it, a plan keeps the lower earner permanently at their small own benefit, over-pricing an early claim and biasing the Couples SS optimizer toward both-delay. With it, the optimizer can find the canonical 'higher earner delays, lower earner claims early and steps up' split the specialist tools recommend. See methodology §14.

If you claim Social Security before your Full Retirement Age (FRA = 67 for anyone born in 1960 or later) and you're still earning a paycheck, the SSA applies an 'earnings test': it withholds $1 of benefits for every $2 you earn above an annual limit (~$24,480 in 2026). It's the classic surprise for Barista FIRE — claim at 62 to bridge, keep working part-time, and discover much of your expected SS is held back.

IT'S NOT LOST — at your FRA the SSA recomputes your benefit upward to credit back the withheld months, so over a normal lifespan it's roughly actuarially neutral. The real cost is timing: less SS during the early working years (when your portfolio is largest and sequence risk is highest), more later.

IN THIS CALCULATOR — we model the net effect as an EFFECTIVE CLAIM AGE. Working above the limit while claiming early ≈ delaying your claim: the withheld months push your effective claim later, which (a) lengthens the bridge you fund from your portfolio in the working years and (b) raises your benefit once it starts (less early-claim reduction). Both effects fall out of one adjustment, so the headline numbers, the year-by-year projection, and every Monte Carlo path stay consistent. When this applies, a note under your SS claim age shows the estimated withholding and your effective claim age.

WHAT THIS DOESN'T MODEL — partial withholding is modeled as deferral (a slight conservative bias toward showing the cash-flow hit); the higher special limit in the year you reach FRA isn't modeled (we use the standard under-FRA $1-per-$2 limit throughout); the test is individual, so it applies to your own SS based on your part-time income only (there's no spouse part-time income field). Only earned income counts — pensions, account withdrawals, and investment income don't trigger it.

Couples rarely die in the same year, and the years one spouse lives alone are some of the most important — and most overlooked — in a retirement plan. When the first spouse dies, the household loses the smaller Social Security check, moves from married to single tax brackets, and (often) loses part of a pension — while still needing to fund the survivor's life for potentially many more years.

IN THIS CALCULATOR — your household projection runs to whichever of you has the longer life expectancy (enter each spouse's life expectancy separately under Spouse Portfolio → Age & Timeline). At the FIRST death it switches to survivor economics for the remaining years:
• Spending drops to your Survivor Income Goal (default 75% of your monthly goal).
• Social Security consolidates to the LARGER of the two checks — the SSA survivor rule (the smaller benefit stops).
• Tax filing becomes single (narrower brackets, lower standard deduction).
• The deceased spouse's pension continues at the survivor election you set — the joint-and-survivor choice made at retirement, from 0% (single-life) to 100% (full), defaulting to 50% (a common election). Each pension has its own election (set next to that pension — yours under Other Income, your spouse's under Spouse Portfolio). The surviving spouse's own pension continues in full.
• Required minimum distributions continue on the combined pre-tax pool.
This runs in both directions — whichever spouse's life expectancy is reached first triggers the transition.

WHY IT MATTERS — before this, the plan stopped at your life expectancy and assumed both of you were alive the whole time, so your spouse's longevity didn't change the answer at all. Now it does: for couples with different life expectancies, projected legacy moves — up for well-funded plans (the survivor spends less and the portfolio compounds through the extra years) and down for tighter plans (more years of drawdown). It also makes the couples Social Security question sharper: delaying the higher earner buys a bigger survivor benefit, which is worth more the longer the survivor lives.

WHAT THIS DOESN'T MODEL — life expectancies are single point estimates, not distributions, so this models the EXPECTED first death, not the full range of timings. The Survivor Scenario card complements this by stress-testing an EARLIER, unexpected death.

When one spouse dies, a pension's survivor benefit continues to the other at whatever joint-and-survivor percentage was elected — and that election is made per pension, not household-wide. The plan lets you set that carryover percentage separately for your pension and your spouse's (0–100%, default 50%): if you die, your pension carries over at your election; if your spouse dies, theirs carries over at theirs. It threads through both the base household projection and the Survivor Scenario card, so the survivor's income reflects the actual elections rather than a hardcoded assumption. Set it next to each pension (yours in Other Income, your spouse's in Spouse Portfolio); it appears only when that pension is greater than zero.

Pensions come in two flavors. Flat-nominal pensions (most private and corporate plans) pay the same dollar amount every year — so real purchasing power declines steadily with inflation. After 30 years of 3% inflation, a $5,000/month flat pension has the buying power of about $2,000/month in today's dollars. COLA pensions (most government, military, and some union plans) increase payments yearly to keep pace with CPI, preserving real purchasing power. Check your plan summary or ask your plan administrator if you're not sure which type yours is. If unsure, leave the COLA toggle OFF in the inputs — that's the more conservative assumption and keeps your plan from overstating future income.

SOME PENSIONS INCREASE AT THEIR OWN FIXED RATE rather than tracking CPI — CalPERS-style plans with a 2% annual adjustment are the classic case. For those, turn the COLA toggle ON and enter the plan's stated rate in the 'Yearly increase' field that appears: the pension then compounds at exactly that contractual rate, every year, regardless of what inflation does in the projection. A fixed 2% increase in a 3%-inflation world still loses ground slowly — the projection shows that honestly. Leave the rate blank for a full-CPI pension.

AND SOME PENSIONS ADD THE SAME DOLLAR AMOUNT EVERY YEAR — set once as a percentage of the first year's check and never recomputed (simple escalation, not compounding). For those, use the dollar field instead: enter the monthly bump (e.g. $60/mo more each year) and the engine adds exactly that amount annually, starting from the payment you entered. Fill in the percent OR the dollar amount, never both — entering one clears the other.

The couples claim-age optimizer (Deep Search) searches all 81 combinations of when each spouse claims. Its default answer maximizes median legacy — which structurally leans toward delaying — but that isn't everyone's goal. The Goal-Scorecard lets you re-color the 81-cell map by four goals (legacy, safety, income now, and the survivor's floor) and hover any cell for its live four-goal scorecard, with an answer-forward headline for each goal. It gives the survivor question the weight it deserves: because couples rarely die in the same year, the higher earner's claim age sets the income floor the survivor lives on, and a portfolio-aware survivor-safety lens says whether that floor is already self-insured or genuinely load-bearing — the SS-and-portfolio intersection no Social-Security-only tool can answer. See methodology §29.

Every dollar figure in the app is quoted one of two ways, and which one depends on what the figure is FOR.

TODAY'S DOLLARS — what the money would buy right now. Everything you enter is in today's dollars: your spending goal, your Social Security benefit, your pension, your healthcare estimate. The engine inflates them for you, so you never have to do that arithmetic yourself. Answers meant to be judged against your life are given back the same way — what you can sustainably spend, the account draw you'll need, the legacy your plan leaves.

FUTURE (NOMINAL) DOLLARS — the number that will actually appear on the statement in that year. Balances are quoted this way, because that is what a balance is: your nest egg at retirement, the year-by-year projection, the Monte Carlo bands. Anywhere the app shows one, it now also shows roughly what it is worth in today's money.

WHY BOTH — a single yardstick would make one half of the app misleading. Quoting your spending goal in future dollars would mean re-doing the inflation maths in your head every time you changed it. Quoting a balance in today's dollars would mean the number never matches the statement you'll actually be looking at. So the app uses whichever fits the question and labels which one it used.

WHY THE GAP IS SO LARGE — at 3% inflation, prices roughly double every 24 years. A retirement can easily run 30. So a balance at age 90 can read two to three times larger than its purchasing power, and that is not the plan being optimistic — it is the same money on a different ruler. This is the single most common source of 'wait, which number is right?', and the answer is usually that both are, measured differently.

ONE CASE THAT IS NEITHER — a lifetime total (lifetime tax, for instance) adds up a different year's dollars in every row, undiscounted. It has no single date, so there is no meaningful 'today's equivalent' to quote. Those totals say so where they appear rather than inventing one.

When you retire before Social Security claim age (which can be as early as 62 or as late as 70), you have to fund all your spending from your portfolio. This is the highest-risk withdrawal window for two reasons: your portfolio is largest, so a market downturn hits the biggest dollar amount; and you're drawing the maximum because no SS is replacing income yet. It's also the calmest tax window in your lifetime — low taxable income opens up Roth conversions, capital-gain harvesting at 0% LTCG, and other tax-optimization moves that disappear once SS and RMDs start. Bridge length = ssClaimAge − retirementAge.

The single highest-leverage defense against sequence-of-returns risk is a dedicated cash reserve for your bridge years. When you retire before Social Security starts, your portfolio funds 100% of spending — and a bad market in those first years is far more damaging than the same market 15 years in (you're drawing the maximum from the largest balance). A bridge reserve parks cash in a high-yield savings account specifically to cover bridge-year spending. The Monte Carlo simulation draws from this bucket FIRST during bridge years, leaving your investment portfolio untouched through the highest-risk window. You can source the reserve from explicit cash savings, OR earmark expected windfalls: home sale proceeds, inheritances, severance, or other lump sums. Each source has its own checkbox in Inputs; toggling it on routes that money to the reserve instead of the general portfolio.

Most retirement calculators assume you spend the same amount every year of retirement. Research (notably David Blanchett's work) shows that's not how retirement actually unfolds: spending peaks in the early go-go years (active travel, hobbies, entertainment), drops about 20% during slow-go years (70-80), and drops another 10% during no-go years (80+). Healthcare costs eventually rise but rarely as much as the leisure spending fell. Modeling spending realistically — phased rather than flat — often improves projected success rates 5-15pp. The 100/80/70 percentages are research-backed defaults; you can tune them in the inputs.

The 4% rule assumes you spend the same amount every year, adjusted only for inflation. Real retirees don't behave that way — they cut back when markets fall and spend more when their portfolio is up. Guyton-Klinger (Guyton + Klinger, 2006) codifies that intuition into a set of mechanical rules around your initial withdrawal rate.

The two rules the engine models:

1. CAPITAL PRESERVATION: If your current withdrawal rate climbs more than 20% above your initial rate (e.g., from 4.5% to above 5.4%), cut spending by 10% AND skip inflation adjustment that year. This protects the portfolio when sequence risk is biting.

2. PROSPERITY BOOST: If your withdrawal rate drops more than 20% below your initial rate (e.g., to below 3.6%), bump spending by 10% AND apply inflation adjustment. You earned some lifestyle upside.

The academic research finds that with these rules, retirees can historically sustain initial withdrawal rates of 5–5.5% — meaningfully higher than the static 4% rule. The reason is mechanical: you're systematically cutting before the portfolio gets too low to recover.

IN THIS CALCULATOR — Guyton-Klinger is opt-in (Inputs → Spending → Dynamic Spending Strategy, OR via the Smart Move workshop in any of the four lenses). When active, phased spending (Go-Go / Slow-Go / No-Go) is paused — your phase values stay saved but the engine uses your monthly goal × 12 as the year-1 baseline and applies the rules from year 2 onward.

SMART MOVES CAP — when applied as a Smart Move (Spend More / Retire Earlier lens), the engine caps the recommended headline at a 5.0% initial withdrawal rate. This is the conservative end of the paper's 4.5–5.5% empirical safe-start range; subsequent research (Pfau, Kitces) has questioned whether the upper end remains defensible in current market regimes. The cap means the lens won't recommend $X/mo with GK if achieving $X/mo requires starting above 5.0% WR — even though the engine math could technically sustain higher rates via aggressive preservation cuts. For Retire Earlier specifically: the lens additionally requires that the strategy's average lived spending still meets your stated monthly goal at the recommended age — so 'retire earlier with GK' actually means earlier retirement at your stated lifestyle, not earlier retirement with the strategy ratcheting your spending down.

WHAT THE NUMBERS MEAN — with GK active, the spend you see is a year-1 starting rate, not a fixed forever-spend. Bad years cut 10% (skipping inflation), good years bump 10% (with inflation). Across surviving Monte Carlo simulations, the average lived spending tends to land lower than the year-1 starting rate due to the asymmetry of cuts firing more often than boosts.

RESEARCH-COMMUNITY CRITIQUE — the original 2006 paper's safe-start range was calibrated against historical market data through ~2005. The post-2008 valuation environment (Shiller's CAPE, bond yields) differs materially, and research from Pfau (2010+) and Karsten Jeske's 'Big ERN' SWR series (2017+) raises concerns that the historical 4–5% rules-of-thumb may overstate safety for high-valuation start years and longer (40–60 year) early-retirement windows. The engine's 5.0% cap respects this critique; users with sub-3.5% effective withdrawal rates have buffer regardless, users at 4–5% should know they're operating in the contested window where research consensus is genuinely debated.

WHAT THE ENGINE DOES AND DOESN'T DO WITH CAPE — your specific start-year CAPE (Shiller's cyclically-adjusted P/E ratio) is the empirical variable the Pfau + Big ERN bodies of work argue determines whether the 2006 paper's range is conservative or aggressive at your retirement year. The engine does not read your start-year CAPE and does not fold it into any recommendation — the 2006 paper's empirical 4.5–5.5% range is the calibration; the 5.0% cap is the conservative-end anchor that hedges against this exact uncertainty. We deliberately don't take a position on whether today is a high-CAPE environment because that reading changes year-over-year, requires picking + citing a specific CAPE source, and the conservative-end anchor's defensibility doesn't depend on the answer. If you want to interpret the cap against today's valuation environment, the cited Pfau and Big ERN work is the place to start that reading.

THE RESEARCH SUCCESSOR — risk-based guardrails key the trigger to your plan's success odds instead of a withdrawal rate, and only fire when the whole plan is genuinely in trouble; the Spending Guardrails tool (Optimize → Deep Search) computes them for your plan. See the Risk-Based Guardrails entry. The two are mutually exclusive by design: this strategy adjusts automatically inside the simulation, that one is the manual playbook.

Withdrawal-rate guardrails (Guyton-Klinger, 2006) made adjust-as-you-go retirement mainstream, but they trigger off one number — your current withdrawal rate versus where it started. That number can't see that Social Security hasn't kicked in yet, that your later phases cost less, or that the same market fall is far less dangerous at 80 than at 65. The result, shown by later research (Kitces + Tharp, 2021+): the classic rules over-correct, demanding cuts of 28–54% through history's worst eras when a plan-aware trigger would have asked for 0–32%.

RISK-BASED GUARDRAILS flip the trigger: instead of the withdrawal rate, watch the plan's PROBABILITY OF SUCCESS — the same Monte Carlo number this app headlines. When it falls to a floor worth acting on, trim spending just enough to restore the plan's sustainable line; when it rises past a ceiling, a raise has genuinely been earned. Because the trigger is computed from your ENTIRE plan — every income stream, every phase, your actual age — it only fires when the plan is actually in trouble, which is why the trims are smaller and rarer. In the advisor world this approach ships in specialist software (Income Lab); almost nothing consumer-facing offers it.

IN THIS CALCULATOR — the Spending Guardrails tool (Optimize → Deep Search, part of Navigator) finds your two portfolio lines by re-running your whole plan at falling and rising balance levels until your success odds cross the acting lines: the TRIM LINE sits where your odds reach 70%, the RAISE LINE where they reach 99%, and both pre-sized actions restore the plan to the same sustainable-at-85% line the rest of the app uses. Only the accounts markets can move are scaled — savings, brokerage, retirement accounts, HSA — never home equity or income streams. The answer comes back in dollars ('if your portfolio closes below $X, trim to $Y/mo'), so a market drop arrives with a plan instead of a panic. Between the lines, the advice is explicitly: do nothing.

OUR THRESHOLDS ARE DELIBERATELY MORE CONSERVATIVE than the published advisor versions, which let a plan ride to 25% odds before cutting — defensible with an advisor monitoring monthly and holding the client's hand; not how we'd treat someone checking in themselves each quarter. Acting at 70% means trims fire a little earlier and a lot smaller.

WHAT THIS IS AND ISN'T — this is a standing CHECK-IN instrument, not a simulation setting: it never changes your plan, your projections, or your saved numbers, and there is deliberately nothing to apply — you use it by writing the two lines down and re-running after big market moves, birthdays, or plan changes. It is mutually exclusive with the Guyton-Klinger toggle (Inputs → Spending): GK is an AUTOMATIC strategy modeled inside the simulation, and with it active your odds already contain adjustments — guardrail lines drawn on top would mean nothing a user could act on. One honest caveat the tool itself states: a deep, multi-year bear market can ask for more than one trim. No adjust-as-you-go strategy avoids that; this one keeps each step small and tells you the next line before you reach it.

Your planning horizon is the age your money actually needs to last to. It is deliberately NOT your life expectancy. Life expectancy is a coin flip — by construction, half of people outlive it — so planning to it means a 50% chance of running short in your final years. Your planning horizon is instead a survival-percentile: an age only a small, chosen fraction of people like you reach, so you're unlikely to outrun it.

We read it from the SSA's 2022 Period Life Table (the same actuarial data Social Security uses). Given you've already reached your current age, we compute the probability you survive to each later age and pick the age where only your chosen percentage remains — e.g. an 85th-percentile horizon is the age just 15% of people like you outlive.

For couples it's the LAST-survivor age: the money must last until the second death, and the joint 'at least one still alive' curve sits above either spouse's own, so a household's honest horizon is longer than either individual's.

IN THIS CALCULATOR — the Planning Horizon Workshop (the 'Not sure?' link under Life Expectancy) lets you pick a confidence level, shows the mean alongside as the coin-flip for contrast, and writes the suggested age to your plan. It's free. Caveats: the period table slightly understates future longevity (a conservative direction), and it's age + sex only — affluent, healthy, or long-lived-family profiles run longer, so lean cautious.

Rather than asking "can I retire at 60?" and getting a yes/no, the engine asks "what's the earliest year you could retire and still hit 85% success?" It does this by scanning candidate retirement ages from 50 to 75, re-running your full plan at each, and finding the youngest age that meets the target. The result is more actionable than a binary verdict: if your earliest age is 58 but you'd planned to retire at 62, you have 4 years of optionality. If it's 67 and you wanted to retire at 60, that gap is what the Smart Moves panel is for — it identifies the specific levers that could close it.

Coast FIRE is the version of financial independence where you hit your savings number EARLY and let compound growth carry you to traditional retirement. The math: if you have $X today and your portfolio compounds at rate r for n years, your future balance is X × (1 + r)^n. If that future balance is enough to fund retirement, you don't need to contribute another dollar — you just need to keep working long enough for compounding to finish the job.

The appeal is workload, not timeline. Coast FIRE doesn't make you retire earlier (you typically still retire at 60–65); it makes the back half of your career meaningfully different. Once you hit your Coast number, you can downshift to lower-paying work you actually enjoy, take more time off, switch industries, or just remove savings pressure entirely. Your income covers your expenses — that's all it has to do.

IN THIS CALCULATOR — Set 'Stop contributing at age' under Inputs → My Portfolio → Age & Timeline (or the parallel field for your spouse). When this age is reached, ALL retirement contributions stop: 401(k), Roth IRA, taxable brokerage, HSA, and the employer match (employer match is tied to your contributions, so it ends when contributions end). Existing balances compound at your pre-retirement return through retirement age. Catch-up contributions still apply during the contribution window if you're 50+.

LEAVE AT 0 (the default) to contribute right up to retirement — the conventional plan for most people. Set this only if you're modeling a real downshift.

ALREADY RETIRED? This field does nothing for you — contributions already stop at your retirement age, and the engine caps this field there. Whether you leave it at 0 or set it to the age you retired, the projection is identical. It exists for people who stop contributing BEFORE they stop working.

SMART MOVES INTERACTION — when Coast FIRE is active, the Smart Moves workshop hides contribution-boost moves (max your 401k, max your Roth, employer match, backdoor Roth, etc.) for the coasting party — they don't make sense if you've explicitly opted to stop contributing. SS-timing, spending discipline, Roth conversions, and other non-contribution moves still apply.

WHAT THIS DOESN'T MODEL — Coast doesn't change your salary, your income, or your tax bracket during the coast years. The engine assumes you're still working (your wages cover expenses); it just doesn't direct any portion of those wages into retirement accounts. If your real-world downshift involves a meaningful salary cut, edit yourSalary directly to reflect it.

Most calculators assume you save the same dollar amount every year until retirement. Real savers raise their contributions as their income grows — contribution escalation models that. For any contribution (401(k), Roth, savings, brokerage, HSA — for you or your spouse), the 'Increase this each year' option lets it climb: a PERCENTAGE that compounds like a raise, or a flat DOLLAR step. Tax-advantaged accounts (401k, Roth) ramp up to the IRS limit and then hold — they never assume you'd contribute above the legal maximum.

IN THIS APP — set it under Inputs → My Portfolio → Annual Contributions. The Projection tab then shows a 'Building your nest egg' table: year by year, how your contributions and growth compound into the nest egg your retirement projection starts from. Every row reads Start + Contributions + Growth = End, and the final row reconciles to that nest egg to the dollar, so you can validate the accumulation yourself. There's also a free Smart Move — Ramp your 401(k) — that recommends a gradual 5%/yr increase for savers who can't max out today but have the runway to grow into it. Plans with no escalation set compute exactly as before. See methodology §30.

Once you start Medicare (age 65), your premiums for Part B and Part D can jump significantly if your income that year exceeds IRMAA thresholds. For 2026: $109K MAGI for single filers, $218K for married filing jointly is the first cliff; higher tiers exist above. Crossing a threshold by even $1 bumps you to the next tier. Roth conversions, large brokerage draws, and RMDs can all push MAGI into IRMAA territory. The Roth conversion ceiling respects these thresholds in the years that count — from age 63, because of the two-year lookback — stopping at the next surcharge line; earlier years are never limited by IRMAA. IRMAA is computed off MAGI from 2 years prior, so planning ahead matters.

Before Medicare at 65, health coverage is one of the biggest early-retirement costs — but where it comes from matters. If a spouse is still working, their employer plan can cover the household until they retire, so the engine charges no marketplace premium and the ACA subsidy is naturally inert for those years (Medicare's base premium still starts at each person's 65th birthday — a documented conservatism, since a working 65+ spouse often delays Part B). You can also tell the app you're on a high-deductible health plan, which makes you HSA-eligible and unlocks the HSA contribution surfaces with the right self-only or family limit. Both are set on the Household Coverage card in Inputs; the AI Advisor reads them so it never warns about premiums for years you're actually covered, and it nudges an HSA when you're eligible and under the limit. When you do buy marketplace coverage, the subsidy is sized on the income the law uses (IRC §36B): adjusted gross income plus the untaxed part of your Social Security — so ALL of your Social Security counts — against the undefined IRS contribution table, with no subsidy above 400% of the poverty line.

Long-term care (LTC) is the daily-living help most people need at some point in late retirement: assistance with bathing, dressing, eating, meal prep, mobility, or full nursing facility care. Importantly, regular Medicare does NOT cover this — only short-term skilled nursing after a hospital stay. The need is common and expensive: a semi-private nursing home room runs a national median of about $9,580 a month — roughly $230K–$340K for a two-to-three-year stay (CareScout/Genworth 2025 Cost of Care Survey). Costs vary widely by state and by care type; a private room runs about $10,800 a month, and assisted living closer to $6,200.

HOW LIKELY, AND FOR HOW LONG — federal research (HHS/ASPE, Long-Term Services and Supports for Older Americans: Risks and Financing, 2022) estimates that 56% of Americans turning 65 today will develop a disability serious enough to require long-term services and supports, though many will need help for less than three years and about 22% will need it for more than five. PAID care is rarer and shorter than the need itself: 55% of older adults will use no paid care at all, the average is 0.8 years, and only about 4% will use five years or more — family caregivers cover much of the rest. Average lifetime long-term-care costs come to about $120,900 in today's dollars, of which families pay roughly 37% out of pocket. So the three-year PAID stay this app stress-tests is a tail case, not the typical one. That is deliberate — a stress test is for the bad outcome — but it is the bad outcome, and worth reading as one. The frequently quoted “70% will need care” counts any assistance at all, including unpaid help from family, on a broader definition than the one above.

WHAT THIS APP MODELS — the premium you enter under Inputs → Healthcare is charged as a yearly cost, and a self-insurance reserve is held back from the income the plan can draw. You can also describe what a policy PAYS (monthly benefit in today's terms, inflation rider, benefit period, elimination period) and we net it against the care cost. Leave the payout blank and the care cost is charged in full — your exposure before any insurance. The inflation rider is the field that decides most of the answer: the same $6,000/mo policy claimed decades later covers roughly 22% of a care bill with no rider and 67% with a compound one.

The deterministic projection (the line on your chart) assumes the same return every year. Real markets don't behave that way. Monte Carlo simulation runs 1,000 alternate versions of your retirement, each with a different random sequence of annual returns drawn from a log-normal distribution centered at your expected return with your stated volatility — and each with its own inflation path: persistent year to year, centered on your stated rate as the typical year, and leaning against the same year's market return so the joint bad year (poor returns while prices rise) is genuinely in the sample. The success rate is the percentage of those 1,000 runs where your portfolio survived to life expectancy. 85% is a common benchmark for 30-year retirements; longer horizons typically target 95%+. A high deterministic projection that fails Monte Carlo is the simulation telling you your plan is unusually sensitive to sequence — when bad years happen matters more than the average return.

Stress tests model specific, named risks rather than abstract volatility. Each scenario applies one stated change (a 5-year bear market at retirement, a 23% SS cut, a healthcare inflation shock) and re-runs Monte Carlo to see how your plan responds. Sequence-of-returns risk is the most important variant: a bad market in your first 5 years of retirement, when your portfolio is largest AND you're drawing it down, does far more damage than the same market 15 years in. Plans that look healthy on average-case Monte Carlo can show dramatically lower survival rates under sequence stress. Common responses if your stress band is much worse than your baseline: hold more cash at retirement, delay retirement by 1-2 years, reduce early-retirement spending, or build a larger bridge reserve.

Historical back-testing replays your plan against the ACTUAL year-by-year market returns and inflation from a specific past retirement-start year (or rolling window of starts), drawn from Robert Shiller's canonical 1928-2022 dataset (S&P 500 nominal total return, 10-year Treasury nominal total return, and CPI-U inflation). It's deterministic — given the same plan and start year, you always get the same outcome. This is the Bengen (1994) / Trinity Study (1998) methodology that originated modern withdrawal-rate research, decades before stochastic Monte Carlo was widely available. Each year's spending comes out at the start of the year and what is left earns that year's return, the convention used by Bengen, the Trinity Study, FIcalc and cFIREsim. (The Monte Carlo and the projection take spending at the end of the year, after that year's growth.)

COHORT SR vs MC SR — Both express probability on the same 0-100 scale, but they measure different sample spaces:
• Monte Carlo SR = 'fraction of 1,000 stochastic-sample futures where the plan succeeds' (random draws from a normal distribution)
• Cohort SR = 'fraction of historical start-year cohorts where the plan would have survived to life expectancy' (e.g., 6 of 8 starts in the 1969-1976 stagflation window)

Neither is 'more correct.' They're complementary. Stochastic MC explores tail outcomes that never historically occurred (its left tail is wider than history). Historical replay captures mean reversion that the MC model doesn't have access to (recoveries always followed crashes). How close the two land depends mostly on the return you assume: if your assumption is lower than what history actually delivered, the replay will look better than the simulation, and the reverse if it is higher. For thin-margin plans (4%+ WR, 30+ yr retirement — the FIRE community sweet spot), historical replay is where dramatic divergences appear: Bengen famously showed that the 1966-1973 cohort broke at 5% WR while later cohorts survived at 6%.

IN THIS CALCULATOR — Two surfaces. (1) Historical Back-Test card on the Stress Test tab — shows cohort SR for one of five named eras (Great Depression, 1966 bear, stagflation, dot-com bust, financial crisis). Picker selects the era; ring shows the cohort survival rate; detail block names the canonical year's specific outcome. (2) Historical Robustness Workshop on the same tab — the exploration surface. Two-knob (start year + stock allocation) running your plan against every eligible historical start year. Outcome strip visualizes survival per year; selected-year detail card surfaces verdict + end balance + lowest balance year; collapsible balance trajectory chart shows nest egg over time.

METHODOLOGY COMPLETENESS — Investment buckets (401k, Roth, Brokerage, HSA) grow at the historical year's actual market return. Spending goal, SS, pensions with COLA, and other inflation-adjusted income scale via historical CPI-U. Tax brackets (IRMAA, LTCG) inflate via the historical sequence × pre-retirement constant rate. HYSA + Bridge Cash buckets grow at the historical year's inflation rate (real return ~0% — matches Shiller's long-run T-bill empirical regularity; floored at 0% nominal to prevent shrinkage during deflation). Slightly conservative in eras where T-bills genuinely paid above inflation (e.g., 1980s) but unbiased on average — fixed in slice 47 (May 26, 2026), previously these buckets grew at the user's modern HYSA assumption regardless of era.

WHAT THIS DOESN'T MODEL — International stocks / REITs / alts (Shiller is US-only). Fixed 70/30 stock-bond split is the default; the workshop exposes a slider for 0-100% but it's still a two-asset model.

Plan Robustness is a holistic read on how resilient your plan is to the things that actually go wrong in retirement: bad market sequences, healthcare cost shocks, longer lifespans than expected, Social Security cuts, inflation surprises. A plan with 95% Monte Carlo success but only 60% under stress testing is less robust than one with 88%/85% — even though the headline number is lower. The robustness view weighs these together with the AI Plan Score's identification of structural risks (account concentration, bridge sequence exposure, spending sustainability). High robustness means your plan has multiple safety margins, not just one big one.

The deterministic projection assumes your expected return happens every single year — no good years, no bad years, just the average. It's the right tool for seeing the mechanics of your plan: when Social Security kicks in, when RMDs start, when accounts get drawn down, when the bridge ends. It's NOT a forecast of what your portfolio will actually do, because real markets are volatile. Use the deterministic projection to understand the SHAPE of your retirement; use Monte Carlo and stress testing to understand the RANGE of outcomes. Many plans look fine in deterministic view but fail under sequence-of-returns stress — that's the gap the simulation surfaces.

For most retirees the home is the largest asset, and what you do with it — keep it, downsize to something smaller, or sell and rent — is one of the biggest levers on the plan. The 'Your home' card (Inputs → Life Events) takes what you know (value, cost basis, mortgage balance and rate, appreciation) and does the equity math for you: it appreciates the value to your sale age, pays off the remaining mortgage, subtracts about 6% selling costs, applies the IRS Section 121 exclusion ($500,000 married / $250,000 single) and long-term capital-gains tax on any excess, and deposits the NET proceeds into your plan at the age you sell — buying a smaller place at a later age if you downsize, or picking up rent if you sell and exit. Because the card computes the exact tax, the engine deposits the net directly rather than re-estimating it. A paid-off home isn't free, either — property tax, insurance, and upkeep stay in your monthly spending. See methodology §32.

A reverse mortgage turns home equity into cash without moving. From age 62 you can borrow against a home you keep living in and take the money as a monthly payment. You make no payments while you live there — the balance grows instead, and is settled when the home is sold or you pass. Because it is borrowed money rather than income, it is not taxed, so in your projection it covers spending directly without touching your tax bill, your Medicare surcharge or your health-insurance subsidy. The costs are real: roughly 4% of the home's value up front plus insurance of about 0.5% a year on the balance, all compounding because nothing is repaid until the end — and the payment is a FIXED amount for life that never rises with inflation, so its buying power falls every year. As the balance grows the equity shrinks, and there may be little or nothing left in the home to leave or to sell into. You can never owe more than the home is worth. You must keep paying property tax and insurance — falling behind on those is the real-world route to losing the home — and HUD-approved counselling is required before any lender can proceed. The research supports narrow uses (a standby credit line drawn on only in a poor market, or bridging a few years to delay Social Security) more than it supports using one as ordinary income. The app models one honestly and compares it beside keeping, downsizing and selling — and recommends none of them. See methodology §33.

Your monthly goal handles ongoing living costs. Future Expenses cover the lumpier outflows that come and go at specific ages: a daughter's wedding at 68, a new car at 72, a kitchen remodel at 75, a grandkid's tuition help at 70. You enter the amount in today's dollars and the engine inflates it to the year it hits via CPI (or mark it Fixed and the figure is charged exactly — a contract price; or give it its own yearly rate when a contract escalates at a known percent, like a Prop 13-style 2% property-tax cap). When that age arrives during the projection, the engine deducts it from your portfolio using the same waterfall as a home purchase — savings first, then brokerage, then Roth, then 401(k). An entry can also repeat: every year through a chosen age (four years of tuition), or every few years (a new car every 7 years) — one entry, each occurrence priced in its own year. Up to 15 entries. Ongoing lifestyle costs (annual vacations, ongoing parental support) still belong inside your monthly goal. The AI Advisor can add, edit, or remove entries for you — it must send your complete list each time, every entry is validated before anything lands, and you review an item-by-item diff before applying; the list stays yours.

Life includes one-time windfalls: an inheritance, the sale of a business, a legal settlement. Future Inflows lets you name each one, set its amount and the age it arrives, and optionally earmark it as a bridge-years reserve. The engine handles the timing honestly — a windfall that lands before retirement is grown to your retirement date and added to the nest egg; one that lands during retirement is deposited in that specific year. It's the money-IN mirror of the Future Expenses list, and it replaced the old single-slot inheritance fields, so a couple with two separate windfalls at different ages no longer has to fold them into one. Windfalls are managed on the Inputs timeline; the AI Advisor can propose changes that you review and apply. See methodology §17 and §24.

A career break is a stretch of years mid-career where your income stops — an involuntary job loss, or a voluntary sabbatical. Two things happen at once: you stop adding to your accounts, and your living costs have to come from somewhere, so the break draws down the savings that were meant to compound to retirement. The cost isn't just the money spent — it's the growth that money would have earned over the decades that follow, plus, if the gap forces an early withdrawal from a 401(k)/IRA before 59½, a 10% penalty on top of the tax.

IN THIS APP — the free Career-Break Workshop (Stress Test tab) models it as an isolated derivative of the real Monte Carlo: during the gap it pauses contributions and draws your accounts in the real cascade order (savings → brokerage → Roth → 401(k), with the pre-59½ penalty), then runs your actual plan forward from the dented balances. It reads the cost four ways (retirement age, retirement income, legacy, downturn resilience) and prescribes the smallest extra saving that recovers the plan — or says honestly when the lever is timing or spending instead. For couples, a toggle models your break, your spouse's, or both together (the worst case — no one earning). Nothing it shows touches your saved plan unless you apply a change. See methodology §31.

Gray divorce — divorce after 50 — has roughly doubled in a generation (Brown & Lin, 2012), and its retirement math is unforgiving: the household's savings divide, but neither new household costs half of what the shared one did, tax filing becomes single (compressed brackets, an earlier Medicare-surcharge threshold), and health coverage that rode on a spouse's employer ends with the marriage.

THE MOST UNDER-KNOWN FACT in this topic is divorced-spouse Social Security: if the marriage lasted 10 or more years and you are unmarried when you claim, you can receive up to half of your ex's full-retirement-age benefit where that beats your own — it does not reduce their benefit, does not affect their new spouse, and they are never notified. The workshop models it rather than mentioning it: your own benefit plus the excess, on the SSA's actual reduction schedule.

IN THIS CALCULATOR — open the workshop from the Gray Divorce card on the Stress Test tab ('Model a split in detail — privately'), on household plans. You set the split age, how each account divides (50/50 by default, every share editable — HSAs default to each side keeping their own), what happens to the house (sell and split, buy them out, they keep it, or sell later — the dollar facts derived from your own home card), support paid or received (flat dollars, per how orders actually work), and your single-life spending (default 70% of the household goal). Both futures run through the same Monte Carlo engine as your headline number. Three free gap-closers then search for what restores the household plan's security — save more by the split, retire later, or spend less — and the Insulate search (part of Navigator) finds the change TODAY that makes BOTH futures sustainable at the app's 85% bar.

PRIVATE BY CONSTRUCTION — nothing in the workshop is saved: no browser storage, no analytics, no trace in your plan. Closing it erases it. The entry screen states the one honest caveat: your browser's own history and autofill are outside our reach, so use a private window on a shared device.

WHAT THIS DOESN'T MODEL — your state's property-division law (we ask who keeps what rather than assuming), legal costs, your ex's resulting plan, or remarriage. See methodology §37 for the account-division identity and the Social Security math.

Deep Search is where the app stops answering 'what happens if' and starts answering 'what's best.' Each of its ten optimizers runs hundreds of full Monte Carlo scenarios against your actual plan and returns the strongest choice for one decision — the kind of searched optimization the expensive advisor tools charge for. Most are judged on your real legacy outcome; Spending Shape is judged on lifestyle (the most you can spend within a success floor and leave-behind target you choose); Employment Pause, Spending Guardrails and the Combined Risk Scan are judged on success odds — the Scan stacks validated pairs of the Stress Test tab's one-at-a-time risks and reports honestly whether any combination is worse than the individual cards predict. It changes nothing about how your plan is computed (it re-applies inputs the engine already reads), so it never affects your saved AI Plan Score; it's opt-in behind a Run button and part of Navigator. Deep Scan is the capstone: it runs the legacy-seeking optimizers together and reports the true combined payoff, which isn't the sum of the parts because the moves interact — and it deliberately excludes Spending Shape, because spending more is a values trade, never 'optimized' away on your behalf. See methodology §29.

A raw Monte Carlo percentage tells you how often your portfolio survives random market sequences, but it doesn't capture every kind of risk. The AI Plan Score weighs the MC result alongside factors the simulation doesn't isolate: sequence-of-returns vulnerability during your bridge period, Social Security timing risk, account concentration (e.g., 90% pre-tax), and the sustainability of your spending assumptions. A plan with 95% Monte Carlo success but significant sequence risk during a long bridge may score lower — because the AI identified a real risk the simulation alone masked. The score uses the same health labels as the AI deep analysis: Excellent, Strong, Solid, At Risk, Critical. Without a deep analysis run, the ring just shows the raw Monte Carlo success rate.

The AI Advisor doesn't nag. Its edge tab stays quiet and pulses only when a deterministic trigger is armed — a cached deep analysis that's gone stale, a success-rate band you've crossed since you last looked, or a first-time topic worth a cold-open. The rule it can never break is that the glow can't bluff: clicking it always lands on the exact surface that armed it. A newer layer, the Observation Engine, can surface the insight BEHIND a change you just made — for example, that retiring a couple of years earlier stretches the self-funded stretch before Medicare — but only deterministic, directionally-unambiguous observations computed from the real engine output, never a chatbot guess. Every trigger is free and costs no AI quota; opening the panel acknowledges it, and it won't re-surface the same thing.

The Compare tab includes a Median Saver in Your Cohort benchmark automatically matched to your age band, income group, and household type. The benchmark uses cohort-appropriate balance, contribution rate, and Social Security benefit, but mirrors YOUR retirement goal and structural assumptions — so the comparison answers "if I had what a median saver has, would my plan still work?" rather than asking "can a median saver fund a synthetic 75% replacement target?" (the latter produces 0% success for most cohorts and isn't useful). Data sources: Federal Reserve 2022 Survey of Consumer Finances (mean balances by income percentile), SSA 2025 Actuarial Note (hypothetical scaled worker PIAs), Center for Retirement Research at Boston College (median calibration). Nothing synthesized at the individual level — all anchored to published household data.

For most people entering retirement the house is the largest thing they own and the only one producing no income. The Home Equity Workbench (Results tab, under Income — the "Your Home" card) takes the three things people actually do with it — keep it, downsize to something smaller, or sell and rent — and runs each one through your real plan on its own isolated copy, so you are comparing what the house does to your RETIREMENT rather than what a sale nets. You can judge them through five lenses: your odds, income now, retiring earlier, what you leave behind, and how a survivor would fare. It opens on the priority you set in your plan and says so. Sale ages, replacement price and rent are yours to move, and the freed equity shown is always NET — proceeds minus the replacement — because a move sideways can free nothing at all. A reverse mortgage sits separately, opened rather than offered, and is never ranked against the three. Nothing changes your plan until you press Apply. See methodology §32.

Because the app keeps your data local (no account, nothing on our servers), you own it — and you can move it. Three tiers: your check-in progress history exports and imports FREE (it's a personal record with no competitive value, and losing it to a cleared browser would be the worst outcome); the current plan and a full backup (every plan, scenario, and your progress) export on the paid tier; and importing any of them is free everywhere, because restoring your own data should never be gated. A backup deliberately excludes identity and entitlement (no account id, email, or paid-unlock) — the same device-bound convention the share links follow — so a backup file can't clone paid access. In the Inputs toolbar and the Progress tab.