How this works

Getting started

You don’t need your full financial life to get a useful answer. There are four steps. The first takes about two minutes and you only do it once.

1. Quick Start wizard: seven inputs, onceOn your first visit a Quick Start wizard opens over the app and asks for seven things: birth years, retirement age(s), total pre-tax savings, other savings (Roth + taxable + cash), household income, annual savings, and planned retirement spending. It turns them into a complete profile by filling everything else with documented defaults: an 80/20 portfolio gliding to 65/35, average market assumptions, and Social Security estimated from your income. You can also skip it and load the demo profile, or skip straight to a blank profile and fill in the detailed panels yourself. Quick Start is not a separate mode: it just pre-fills the same profile the panels edit. You can bring it back any time with the “Re-run Quick Start” button, which tells you exactly which fields it would overwrite before it changes anything.
2. Refine in the detailed panelsEvery value the wizard set can be edited in the panels (Household, Accounts, Income & Savings, Spending, Strategy, Market Assumptions and more), and the simulation re-runs automatically as you type.
3. Work down the fidelity checklistThe estimate fidelity meter shows how much of your result rests on defaults. Its “Improve your estimate” checklist lists each high-impact field you haven’t set, with why it matters (for example: taxable cost basis affects capital-gains tax when you sell). The Assumptions footnote under the results always lists every default in use. Until the score reaches 50, the success rate is shown as a range (±3 points) instead of a single number.
4. Compare scenarios A/BClick “Duplicate → Scenario B”, change one thing (a Roth conversion strategy, a retirement age, a claiming age) and turn on side-by-side compare. You get paired stat tiles, overlaid net-worth medians, and the difference in lifetime taxes, so you can see what a single decision is worth.

Question-first entry points. If you arrive with one question (“Can I retire at 60?”, when to claim Social Security, whether to convert to Roth, IRMAA cliffs, how much you can spend), the cards on the home page answer it from at most five inputs. They use the same simulation on fewer market paths (1,000; 400 per spending level tried) and list the defaults they assumed under the answer. “Open this in the full simulator” makes those inputs your plan only if you have no saved plan of your own; otherwise it opens as scenario B for comparison, or not at all. With a saved plan, each question can also be answered on that plan, unchanged.

How the simulation works

One simulated year at a time

The simulator moves your household forward one calendar year at a time, from today until the youngest household member reaches the horizon age (default 95). Each year it does the following, in this order:

1. Grow. Apply that year’s market returns to every account, weighted by your stock/bond allocation (which can glide toward retirement). Dividends and interest are recorded as income and reinvested (taxable dividends are part of the total return, not growth on top of it). An optional post-retirement “bond tent” glide (Market Assumptions panel, under Allocation) can arrive at retirement conservative and then drift the stock share back up over a chosen number of years, so a crash in your first retired years hits a smaller stock share.
2. Contribute. While anyone in the household works: wages come in, and 401(k), Roth, HSA, and brokerage contributions flow to their accounts, capped at inflation-indexed IRS limits. The 401(k) limit includes the extra catch-up from 50, and the larger catch-up at ages 60 to 63, which the simulator assumes your plan offers. Enter your gross salary: pre-tax 401(k) contributions are taken out of taxable wages, just as they are on a W-2, so they lower that year’s income tax and MAGI. Roth 401(k) and after-tax contributions do not. Every state in the simulator follows this federal rule, although a few states tax 401(k) contributions when you earn them.
3. Required distributions. Once a person reaches RMD age (73 or 75 depending on birth year, per SECURE 2.0), the required minimum distribution is forced out of pre-tax accounts as ordinary income. The simulator treats all pre-tax money like IRA money here, so it starts these withdrawals on time even for someone still working. A current employer’s 401(k) can usually wait until you retire (unless you own more than 5% of the company), so for a late retiree the simulator can take RMDs, and their tax, too early. Inherited retirement accounts also make their required 10-year-rule distributions here, and Social Security benefits are paid if claimed. Each benefit is raised or cut for your claiming age against your full retirement age for your birth year (67 if you were born in 1960 or later, a few months earlier for 1955 to 1959).
4. Convert. If a Roth conversion strategy is active, pre-tax money is converted to Roth, for example up to the top of a target tax bracket. Conversions happen before withdrawals so the bracket math sees the year’s final ordinary income.
5. Withdraw. Spending, lump expenses, healthcare premiums, and taxes are covered from accounts following your withdrawal order (typically cash, then taxable, then tax-advantaged). Because withdrawing itself changes the tax bill, the simulator recalculates withdrawals and then taxes, over and over, until the two agree to within $50. While anyone still works full-time, everyday spending is assumed to come from wages, but lump expenses and gifts scheduled in those years are still paid from the portfolio.
6. Tax. Federal tax with proper long-term-capital-gains stacking, NIIT, an approximate state tax, ACA premiums and subsidies before 65, and Medicare IRMAA surcharges after 65 (based on income from two years earlier) are all computed and paid.

What “Monte Carlo” means

Nobody knows next year’s market return, so Deorbit Plan runs that year-by-year loop thousands of times (5,000 by default, adjustable from 1,000 to 10,000), each time with a different randomly generated market history. The success rate is the share of those simulated lifetimes in which your money never ran out before the horizon. A separate run that uses average returns every year feeds the detail charts (account balances, taxes, RMDs year by year).

Two ways of generating markets

• Parametric (default): each year’s stock and bond returns and inflation are drawn from bell curves. By default stocks average 7% per year (nominal) with 16% volatility, bonds 4% with 7%, cash 2.5% and inflation 2.5%, with a slight negative stock–bond correlation. Every one of these numbers is editable in the Market Assumptions panel.
• Historical bootstrap: each simulated year is a real calendar year drawn at random from 1928–2025. Because a whole year is sampled at once, the historical relationships between stocks, bonds, cash, and inflation in that year are kept. The order of years is shuffled, though, so multi-year patterns (long bear markets, stretches of high inflation) are not.

Nominal dollars vs. today’s dollars

The simulator works in nominal dollars: the actual (inflated) amounts of each future year. The Display: Nominal $ / Today’s $ toggle above the results applies everywhere. It switches every dollar result (charts, stat tiles, the year table, CSV exports and scenario-compare differences) between the two units. Today’s $ means start-year (2026) dollars: each value is divided by its own simulated path’s cumulative inflation up to that year, so a “$1M in 2055” outcome is shown as what it would buy in 2026. The simulation itself doesn’t change, only how the results are displayed.

The 2008 stress test

As an optional worst-case check, the stress test replaces one simulated year with the actual 2008: stocks −36.5%, bonds +20.1%, cash +1.4%, inflation +3.8%. It can also follow that with the real 2009 recovery (+26% stocks). You choose when it hits: the first year of retirement (the classic sequence-of-returns risk), a random year, or a fixed calendar year. It works with either market model.

The event timeline under the charts

Every chart carries a slim marker lane under its x-axis showing your household’s milestones on the same time scale: ● life events (retirements, Social Security claims, Medicare at 65, RMDs starting), ▲ strategy events (Roth conversions starting/stopping, the first ladder rung maturing), ◆ money events (gifts, lump expenses, inheritance receipts and their 10-year deadlines), and ✕ the stress-test crash. Hover or keyboard-focus a marker for the details, with real dollar figures read off the simulated path; the Events toggle above the results hides the lane entirely.

A probability, not a prediction

An 85% success rate does not mean your plan is 85% finished or that anything is guaranteed. It means that across thousands of plausible market histories, the plan held up in 85% of them. The tool is most useful for comparisons: how a decision moves the odds and the lifetime tax bill. Don’t read much into the decimal places of any single number.

Key concepts glossary

Roth conversion ladder & the 5-year rule

Moving money from a pre-tax account to a Roth account on purpose, paying ordinary income tax now so it grows tax-free forever after. Each converted amount must “season” for 5 tax years (or until age 59½) before it can be withdrawn penalty-free. So early retirees convert a slice every year, building a ladder of money that becomes available year after year. The simulator tracks every conversion’s year and applies the 10% penalty only to unseasoned amounts.

RMD (Required Minimum Distribution)

Starting at age 73 or 75 (depending on birth year), the IRS forces you to withdraw a share of pre-tax accounts each year (the balance divided by a life-expectancy factor) and pay ordinary income tax on it, whether you need the money or not. Large pre-tax balances late in life can force large taxable RMDs. Shrinking them is the main reason people make Roth conversions. RMDs themselves can never be converted to Roth.

MAGI (Modified Adjusted Gross Income)

The income measure that health-related programs use: your adjusted gross income with certain untaxed items added back. Each program picks its own items. For ACA subsidies that includes the untaxed portion of Social Security; for Medicare IRMAA surcharges it is only tax-exempt interest, so untaxed Social Security does not count. Roth conversions raise both versions of MAGI, so they need to be planned around both programs.

ACA 400% cliff

Under post-2025 rules, marketplace health-insurance premium subsidies vanish entirely once household MAGI exceeds 400% of the federal poverty level. One dollar of extra income at that boundary can cost thousands in lost subsidy. At the other end, there is no subsidy below 100% of the poverty level, and Medicaid is not modeled, so very-low-MAGI years show the full, unsubsidized premium. Married filing separately receives no subsidy at all (the credit requires a joint return), so that status shows the full premium. Deorbit Plan models the cliff exactly, and the MAGI chart draws the cliff line so you can see how close each year runs.

IRMAA lookback

Medicare premium surcharges (Income-Related Monthly Adjustment Amount) are tiered on your MAGI from two years earlier. A big Roth conversion at 63 raises your Medicare premiums at 65. The simulator keeps two years of income history and charges the surcharge for each person on Medicare.

NIIT (Net Investment Income Tax)

An extra 3.8% federal tax on investment income (interest, dividends, capital gains) to the extent MAGI exceeds $250,000 for joint filers. The law fixes that threshold and never adjusts it for inflation, so more retirees cross it over time. Conversions raise MAGI but are not themselves investment income.

LTCG stacking

Long-term capital gains and qualified dividends are taxed at 0%, 15%, or 20%. Which rate applies depends on gains being “stacked” on top of your ordinary income. Extra ordinary income (a conversion, an RMD) can push gains that would have been tax-free into the 15% bracket. The simulator models this stacking exactly.

QCD (Qualified Charitable Distribution)

From age 70½, giving to charity directly from an IRA counts toward your RMD but never shows up as taxable income, which usually beats donating cash. In Deorbit Plan, an annual QCD amount goes from your pre-tax accounts to charity tax-free, up to the annual QCD limit per eligible person ($111,000 in 2026, indexed). It counts toward your RMD first, so the taxable RMD falls (never below zero), and any amount above the RMD, or given before RMDs start, simply leaves the account without being taxed. The 70½ rule is modeled from the calendar year you turn 70.

Guardrails spending

A flexible-spending rule (simplified Guyton-Klinger): start at a chosen withdrawal rate and cut spending modestly when markets push your withdrawal rate above a guardrail band, and raise it when the rate falls below. A few small adjustments give a noticeably higher success rate than spending a fixed amount every year.

Withdrawal order

Which accounts you tap first. The default before 59½: cash, then taxable, then Roth contributions, then seasoned conversions, with penalized pre-tax and Roth-earnings withdrawals as last resorts. After 59½ the penalties no longer apply and pre-tax moves up the list. You can change the order, and it can change lifetime taxes a lot.

Giving & inheritance

The Giving & Legacy panel models gifts from the giver’s side and shows an “Estate value to heirs (after their taxes)” number. A few tax rules make giving during your life and leaving money at death very different, and they are worth understanding before you compare scenarios. This is educational only, not advice.

Step-up at death vs. carryover basis on lifetime gifts

Appreciated taxable assets you still hold at death get a stepped-up basis: your heirs’ cost basis resets to the value on the date of death, and the unrealized capital gain is never taxed. The same assets gifted during your lifetime carry a carryover basis instead: the recipient takes over your original cost basis and will owe capital-gains tax on your whole gain when they sell. That is why the estate metric counts taxable accounts at full value, and why gifting cash (or funding gifts by selling and paying the gains tax yourself) is often cleaner than gifting appreciated shares.

The annual gift-tax exclusion

In 2026 each person can give up to $19,000 per recipient with no gift-tax return. A married couple can give $38,000 to one recipient with no return only when each spouse gives $19,000 of their own money. If one spouse gives the whole $38,000 and the couple elects to split it, a Form 709 is required whatever the amount, generally from each spouse. Above the exclusion, you file Form 709 and the excess uses up part of your lifetime estate-and-gift exemption ($15 million per person under 2026 law). Deorbit Plan shows an informational note when a year’s gifts exceed the exclusion, and never computes a gift tax.

Why Roth is the best asset to bequeath, and an HSA the worst

A Roth account passes to heirs completely income-tax-free (they still must drain it within 10 years, but no distribution is ever taxed). A pre-tax account passes with its embedded tax bill: heirs pay ordinary income tax at their marginal rate as they drain it. An HSA is worse still: a non-spouse heir must include the entire balance as taxable income in the year of death, with no 10-year stretch at all. The estate tile’s formula follows this ranking. That is why Roth conversions can raise the after-tax estate even when they lower the ending balance in nominal dollars.

The 10-year rule, seen from the giver’s side

Since the SECURE Act, most non-spouse heirs must empty inherited retirement accounts within 10 years. For a pre-tax account that compresses what might have been decades of small taxable withdrawals into ten years stacked on top of your heirs’ own peak earnings, often at a higher marginal rate than you would ever have paid. The heirs’ marginal-rate field in the Giving & Legacy panel is that assumption. The estate value reduces pre-tax and inherited-traditional balances by that rate.

QCDs and charitable bequests

If some of your money is meant for charity, a Qualified Charitable Distribution is usually the most tax-efficient route. From age 70½ it goes directly from the IRA to the charity, counts toward your RMD, and never appears in anyone’s taxable income, yours or your heirs’. Giving pre-tax dollars to charity through QCDs and leaving Roth and stepped-up taxable assets to people gives each recipient the asset that is worth the most to them.

Filing statuses

Deorbit Plan models four federal filing statuses. The status you pick decides how many people are simulated and which tax tables apply. Brackets, standard deduction, capital-gains thresholds, NIIT, Social Security taxability and IRMAA all depend on filing status.

• Married filing jointly: two people, one joint return. The household retires when the last person retires, and each person claims Social Security and takes RMDs on their own schedule.
• Single: one person, single-filer tables throughout.
• Head of household: one filer with a qualifying dependent. HoH brackets and deduction apply. The dependent is otherwise not modeled, except for an ACA household size of 2 (you can override it).
• Married filing separately: models this filer’s return only; the spouse is not simulated at all. The tax law treats MFS harshly, and the model applies those rules in full: Social Security benefits are taxable from the first dollar (the $0/$0 thresholds for couples living together), and Medicare IRMAA uses the separate two-tier MFS table, so high-MAGI years jump straight to the top surcharges.

State taxes for the one-person statuses use an approximation: 50% of the state’s married-filing-jointly standard deduction and exemption, with the MFJ rates and brackets unchanged, unless we have that state’s own figures for the status. The Assumptions footnote under your results says when this approximation is active.

Married-filing-jointly households can also test the Survivor what-if (Household panel): one spouse passes at a chosen age and the survivor files single from the following year (the “widow’s penalty”). See the filing-status entry under Assumptions & limitations for exactly what it models and approximates.

Already retired?

Deorbit Plan also models a household that is already living on its portfolio. The Quick Start wizard asks “Are you already retired?” right after your filing status. If you answer yes, it skips the income and savings questions and asks for the things that matter now.

• Your actual Social Security checks. If you already receive benefits, enter the real monthly amount (per person). The model pays it from the first simulated year with an inflation COLA and ignores the at-FRA estimate and claiming age entirely. In the Household panel this is the “Already receiving Social Security” toggle on each person’s card.
• Your last two years of MAGI. Medicare IRMAA surcharges look back two years, so your first two simulated years of premiums are set by tax returns the simulation cannot know. Enter the actual values in the wizard or under Household → Medicare IRMAA lookback. Use the IRMAA figure (adjusted gross income plus tax-exempt interest) as shown on Social Security’s IRMAA notice. Near a tier boundary this matters: with the 2026 tiers, crossing one boundary costs about $1,148 to $1,736 a year per person ($581 at the top boundary), and for married filing separately crossing $109,000 costs about $6,355 a year.
• Pensions & annuities. Add each stream in the Household panel: the nominal monthly check at its start age, the start age, and its COLA (0% = fixed, typical for private pensions). Pension income is ordinary income. It counts toward Social Security taxability, MAGI (ACA/IRMAA) and state retirement-income exclusions, and it is not subject to NIIT. That is right for employer pensions and annuities paid from an IRA or retirement plan. An annuity you bought yourself outside a retirement plan works differently: part of each payment is a tax-free return of what you paid, and the taxable part counts as investment income for NIIT. The simulator treats it like a pension, which overstates its income tax and leaves out any NIIT on it. Each pension carries a survivor share (default 50%) that the Survivor what-if uses if the owner is the spouse who passes.
• Existing inherited IRAs. An inherited IRA received in a past year keeps its original 10-year clock: enter the receipt year and the account’s current balance in the Inheritances & gifts received panel, and the required withdrawals continue from where they are now. If the deadline has already passed, the account is emptied right away, with a warning. If the person you inherited from had already started required distributions, the law also requires a yearly minimum withdrawal in years 1 to 9, whichever withdrawal plan you pick. The simulator takes that minimum every year, sized with a rough life-expectancy estimate that runs a little high for older heirs, so it can take slightly more out early than the law requires.
• Recurring gifts you receive. The Inheritances & gifts received panel also handles regular cash gifts you expect to receive, such as “$72,000 cash every year” from a parent. Tick “Repeats,” set how often, and either a fixed count or “until the end of the plan.” Gifts you receive are not taxable income to you. Cash gifts can be deposited into the cash bucket (default) or invested straight into the taxable account. Each repeat of an inherited account gets its own 10-year clock, just like a separate inheritance.
• Part-time work. Wages after your retirement age (a window of ages, today’s $) reduce what the portfolio must cover without delaying “retirement” in the model: retirement spending starts once nobody works full-time, and no 401(k)/IRA/HSA contributions are made from part-time income. The event lane marks the year part-time work ends.
• Working past 65 with employer coverage. The “Defer Medicare while working” toggle skips Medicare premiums and IRMAA (and ACA) for a person 65 or older until they stop working, which matches the penalty-free special enrollment period. It also decides HSA contributions: the law stops HSA contributions once you are on Medicare, so without the toggle the simulator makes no HSA contributions for a worker 65 or older. HSA contributions continue while at least one full-time worker in the household is not on Medicare. The year you turn 65 counts as a full Medicare year, which slightly understates what you could put in that year.

The retirement paycheck tile

When the household is fully retired, the dashboard leads with “Your <year> retirement paycheck”: the spending money in the expected-return path’s first fully retired year, with that year’s healthcare (premiums, Medicare base + IRMAA, out-of-pocket) and taxes (federal + state + NIIT + penalties) shown alongside. Under the guardrails spending strategy the tile’s tooltip also quotes the portfolio levels at which spending would be cut or raised. It summarizes one typical year. For the risk, look at the success gauge and the fan chart.

Strategy Lab

The Strategy Lab (the “Strategy Lab” tab, or /lab) runs full simulations of your current profile across a set range of choices and lays the results out for you to read. It explores; it doesn’t pick an answer for you.

What it does

• Max sustainable spending: finds the highest fixed annual spending (in today’s dollars) whose success rate still meets your target (70–99%, default 90%). It narrows the range by halves, to $1k precision in at most 10 runs. It also shows the range of ending estates at that spending level: even a plan aimed at “die with ~zero” usually leaves something behind. It only applies to the fixed real spending strategy.
• Guardrails floor solve: when your plan uses the guardrails spending strategy, this replaces Tool 1. With guardrails, the strategy itself cuts spending when the portfolio struggles, so the success rate sits near 100% and stops telling plans apart. It tells you whether the money ran out, but not how much you had to cut back. With guardrails the real risk is cuts to your lifestyle, so this tool finds the highest starting spending whose real spending never falls below a floor (default 70% of that starting amount) in enough of the simulated futures (the same 70–99% slider, default 90%). For each spending level it tries, it sets the guardrails starting withdrawal rate from your net worth at retirement on the average-returns path, so the strategy doesn’t immediately work against the number being tested. The result shows the maximum spending, the chance of staying above the floor, the floor in dollars, the ordinary success rate (for reference only), the range of ending estates, and, most prominently, the median LIFETIME real spending, which is what guardrails actually maximizes. Applying it sets both the spending and the matching withdrawal rate together (the confirm dialog lists both), so the strategy stays consistent with the number you just solved for. It does not work with %-of-portfolio spending, whose floor and ceiling already limit spending directly; the tool is disabled there, with an explanation.
• Strategy sweep & frontier: every combination of Roth conversion strategy (none, fill 12/22/24/32%) and Social Security claim age (62/67/70 per person) is simulated with the same market draws (common random numbers), so points differ only by strategy and never by market luck. The scatter plots success rate against a median measure you choose (lifetime tax, estate to heirs, or lifetime spending, all in today’s dollars). Your current plan is marked ⭐, and points beaten on both axes are greyed out as “dominated”.
• Cliff margin: every Lab result shows how close MAGI comes to the next ACA/IRMAA cliff above it on the average-returns path, in warning color when the gap is under $5k. The Strategy panel’s cliff-margin guard can cap fill-bracket conversions to keep a safety margin below the nearest cliff. Because Medicare looks back two years, the IRMAA tiers are watched from age 63, the first year whose income sets a Medicare premium. The guard uses that year’s tier amounts, which are usually a little lower than the ones that will apply two years later, so it errs on the safe side.

What it won’t do, and why

The Lab never recommends a strategy, ranks the frontier, or applies a “winner” for you. Picking the best result from a noisy simulation rewards whatever got lucky under one set of assumptions, so the Lab shows you the trade-offs and leaves the choice to you. Sweeps run on fewer paths (default 1,500) to stay fast, so small differences between points are within simulation noise. “Load as Scenario B” and “Apply to my plan” (which always confirms exactly the fields it changes) re-run the chosen strategy at your full path count.

Sensitivity caveat

Winners can flip with market assumptions. Re-run under historical bootstrap before trusting a choice. A strategy that wins under parametric returns can lose under bootstrapped history (and vice versa); treat any close call as a tie.

Top levers (on the results page)

After each run the simulator tries a short list of single changes in the background: retiring one or two years later or earlier, spending 10% or 5% less or 5% more, claiming Social Security at 62, 67 or 70 (each person separately, only claim ages still ahead), no Roth conversions or filling the 12%, 22% or 24% bracket, and drawing pre-tax accounts before or after the taxable brokerage. Each change runs on the same 2,000 market paths as a baseline copy of your plan, so the comparison is path by path and the difference comes from the change, not from market luck. Using 2,000 paths keeps the whole set to a few seconds. The trade-off is that the baseline figure can differ slightly from your headline success rate, and small effects are hard to tell from noise.

A change is reported as moving the success rate only when the shift is at least 0.5 points and at least 2.5 standard errors of the paired difference (the ± shown is about two standard errors). Ending balance and lifetime tax are medians in today’s dollars, with a 95% bootstrap range; they count only when that range excludes zero and the change is at least 1% of the median. Anything smaller is shown as “no meaningful change”. The list is for exploring, and it is not advice. It does not combine changes or try part-time work, other allocations or a move to another state, and a change can rank differently under other market assumptions.

Nothing changes your plan until you click. Apply changes only the listed fields on the scenario you are viewing and re-runs it with your full number of paths; Open as scenario B puts the change in scenario B for side-by-side comparison (asking first if a scenario B exists). Either one can be undone once, until you edit the plan again. The what-ifs can be cancelled while they run, and results are remembered for the rest of the visit.

Review my plan

A fixed set of checks reads the median path (the simulated path whose ending balance is in the middle) and lists what it saw: years with Medicare IRMAA surcharges, marketplace premium credit lost to the subsidy cliff before 65, early-withdrawal penalties, a long bridge funded from taxable savings, years where the maximum share of Social Security is taxable, required distributions larger than spending, surplus income parked in cash, the survivor what-if’s filing change, unused room under a state retirement-income exclusion, the net investment income tax, taxable income above the level where the model’s missing AMT and itemized deductions matter, and a median path that runs out of money. Each line links the Learn article that explains it and, where one bears on it, the matching Top lever. The checks describe the simulation. They are not recommendations, and one path is not a forecast. Each check is a fixed rule written in advance, and they all run in your browser.

Assumptions & limitations

These are the simplifications in this model. Know them before you rely on a result.

• State tax is an approximation. 33 states (plus a “no state tax” option) are modeled as simplified flat-rate or bracket-table approximations with each state’s Social Security and retirement-income exclusions. Each state’s age-based break is applied to the kinds of income that state allows: most cover only IRA, 401(k) and pension income; Georgia’s also covers interest, dividends and capital gains; South Carolina’s age-65 deduction covers any income, wages included. After a death, South Carolina also lets the survivor deduct some retirement income that came from the spouse who died; the simulator leaves that out, so it can overstate a South Carolina survivor’s tax by up to about $520 a year. Credits, local taxes, and phase-outs are ignored. The Household panel shows each selected state’s own modeling notes (what it keeps and what it drops) right under the state selector. Known simplifications beyond those notes: Minnesota’s Social Security subtraction is not modeled, so Minnesota tax is overstated for Social Security recipients; and states that treat public pensions differently from private pensions are modeled with a single pension treatment.
• No Social Security earnings test. The retirement earnings test (benefits withheld when you claim before full retirement age while still earning above the annual limit) is not modeled. Part-time work combined with early claiming is supported, so in that situation the simulated benefits can be higher than you would actually receive while working.
• No AMT, no itemized deductions. The standard deduction is always used, plus, for each person 65 or older, the extra age-65 standard deduction and, for tax years 2025–2028 only, the $6,000 senior deduction (phased out above $75,000 of income, $150,000 joint; not for married filing separately). The alternative minimum tax and SALT itemization are not modeled. Any year whose federal taxable income exceeds $1,000,000 (in that year’s dollars) is flagged “AMT/itemization possible: consult a professional” in the year table and on the results page. Read the results for those years with extra caution.
• One filing status for the whole simulation, unless the Survivor what-if is on. The status you pick (MFJ, single, head of household, or MFS; see Filing statuses) applies throughout. The one exception is the MFJ Survivor what-if (Household panel), a fixed scenario in which one spouse dies at an age you choose. The death year still files jointly; from the next year the survivor files single (single brackets and deduction, lower Social Security taxability thresholds, ACA household of 1). Medicare IRMAA looks at the tax return from two years earlier, so for the first two years after the death the survivor’s surcharge is figured from the joint return against the joint tiers, and the single tiers apply from the third year. The survivor also rolls the retirement accounts into their own (jointly held taxable gets a half basis step-up, as in a common-law state), keeps the larger of their own Social Security benefit and the survivor benefit. The survivor benefit starts from the full benefit of the spouse who died. If that spouse had not yet claimed, it includes only the delay credits they had earned by the time of death, not the ones they planned to earn. If they claimed early, the survivor benefit is capped at the larger of their reduced check or 82.5% of their full benefit. If you enter the spouse’s current check instead, that check is used as the full benefit, which can understate the survivor benefit after an early claim. The survivor’s own age then sets the share (nothing before 60; 71.5% at 60, rising in a straight line to 100% at the full retirement age for survivor benefits, which is 67 for survivors born in 1962 or later and a little earlier for those born before; the share follows the survivor’s age each year rather than being fixed when the benefit starts). The survivor also receives each pension’s survivor share (default 50%) with COLA on the reduced amount, and spends the survivor spending factor (default 75%) of the couple’s base. The two-year “qualifying surviving spouse” status is not modeled, and mortality odds still appear only as a chart overlay. They are never part of the simulation itself.
• Household penalty age. With two people, the 10% early-withdrawal penalty is keyed to the older person’s age household-wide, so the younger person’s dollars become penalty-free slightly early in the model.
• Two MAGIs. IRMAA tiers use Medicare’s definition (adjusted gross income plus tax-exempt interest, with no Social Security added back); ACA premium credits use the ACA definition, which also adds back the untaxed part of Social Security. Tax-exempt interest is not modeled, so the IRMAA figure is your adjusted gross income. The two years of MAGI history you enter seed both figures for the first simulated years.
• Social Security taxability. The share of benefits that is taxable is figured from provisional income: all other income in adjusted gross income (wages, pensions, IRA withdrawals and Roth conversions, interest, and realized capital gains and qualified dividends) plus half of the benefits. Tax-exempt interest, which also counts, is not modeled.
• Bootstrap years are shuffled independently. The historical model keeps each year’s stock/bond/inflation relationships but not multi-year sequences, so there is no momentum and no long stretches of one kind of market.
• Aggregate cost basis. Taxable-account sales use a proportional (average-basis) approximation with no per-lot accounting. Tax-loss harvesting is not modeled.
• Smaller ones: marketplace premiums use your age on January 1, when the plan year starts; Medicare’s top IRMAA income line stays at $500,000 ($750,000 joint) through 2027 and only then rises with inflation, while the other lines rise every year; dividends and interest are always reinvested; in years with Social Security, conversion room is estimated before the conversion itself slightly raises taxable SS, so a fill-the-bracket conversion can overshoot a little; employer match continues at the household amount while anyone works; the pre-tax ownership split between two people is fixed over time.

Privacy

Everything runs in your browser. The simulation runs in a Web Worker on your machine. There is no backend and no account, and no financial data leaves the page unless you copy a share link yourself (see below). Your inputs are saved to your browser’s localStorage so they survive a refresh, and you can download them as a JSON file (Export JSON) to back up or move to another device, then load them back with Import JSON. Clearing your browser’s site data erases everything. Like most websites, the page loads a Google advertising/verification script, but your inputs and results are never sent to it or to anyone else. The privacy policy has the details.

Sharing a scenario by link

Share this scenario (in the simulator toolbar and on the printable report) builds a link to the scenario on the tab you are viewing. The dialog lists exactly what the link contains: birth years and ages, balances, income, savings, spending, strategy and market settings, and the state. People always appear as “Person 1” and “Person 2”; the scenario name and the labels of expenses, gifts, inheritances and pensions are replaced by generic ones unless you tick “Include my labels”. Your other scenario and your display settings are never included. The question pages have a similar Share this answer link that carries only that question’s inputs.

The scenario is written into the link after the “#”, the part browsers do not send to servers, so the link is built and opened without sending the scenario to our server. Anyone who has the link can see the numbers, though, and so can scripts running on the page when it is opened. Opening a shared link never changes your plan: you first see a summary, and then choose. If you have no plan of your own yet, “Explore it” makes the shared scenario your starting plan; if you do, it can only be opened beside yours as scenario B to compare (an existing scenario B is replaced only after you confirm), or dismissed.

Disclaimer

This is an educational model, not financial, tax, or investment advice. Tax rules are simplified, future tax law is assumed to stay as it is today, and the market settings are assumptions, not forecasts. Use it to understand trade-offs and compare strategies, then verify any important decision with a qualified professional.