"When can I retire — and will the money last?"
The question the whole app exists for. RangefinderInvest runs a seeded Monte-Carlo projection of your entire household in today's dollars — starting from your real balances, split by tax treatment — and models the things that actually decide the answer: income and contributions while you work, spending in retirement, Social Security, per-owner RMDs, and real federal taxes, year by simulated year.
From your first paycheck to your last withdrawal
Four sample households ship with the app — one per life stage — so you can explore every scenario before touching your own data. Every figure below is real output from a deterministic run on these samples.
Fictional demo Four sample households (≈$57k / $352k / $882k / $1.09M) ship with the app. Every number here is read from a deterministic run on those samples, and every chart is a real capture. Reproduce any of it: Files → load a sample → open Projection.
One plan, three verdicts — on purpose
Most planners quietly pick one return model and hand you a single confident number. RangefinderInvest scores the same plan three ways, side by side — because the spread between them tells you how much your plan leans on one reading of history.
- ✓Forward · Normal — a classic forward-looking draw from expected returns and volatility.
- ✓Forward · Student-t — the same, with fat tails, because markets crash harder than a bell curve admits.
- ✓Historical — your plan replayed against the actual 1928–2025 U.S. record, with four sampling methods; the block-bootstrap default keeps the sequence-of-returns risk that sinks real retirees while breaking the one-time 1980–2020 bond tailwind.
- ✓Every knob visible — returns, inflation, spending, ages, brackets: each assumption is a setting you can see and turn. Save any set as a named scenario and switch instantly. No black box.
Real output from a sample run: 86.6% under Forward · Normal, 85.9% under fat-tailed Student-t, 99.7% replaying 1928–2025. The 13-point spread is the finding — that plan was leaning on history's bond bull.
A real tax engine, not a flat haircut
Every simulated year is taxed properly, in today's dollars — the projection knows a dollar in your IRA is not a dollar in your Roth. On top of that sits the optimizer retirees actually need: fill-to-bracket Roth conversions in the lean years between retiring and RMDs.
- ✓Federal brackets + standard deduction, a flat state rate, the Social-Security provisional-income worksheet (0 / 50 / 85% taxable), and ACA / IRMAA income ceilings.
- ✓Per-owner RMDs on each spouse's own SECURE-2.0 start age and the IRS Uniform Lifetime Table — charted with and without your conversion plan, so you see the balloon deflate.
- ✓Fill-to-bracket conversions — top off the 12%, 22% or 24% bracket every year it's cheap, automatically zero once Social Security and RMDs fill it themselves.
- ✓Documented simplifications — ordinary-income only, no LTCG/NIIT stack, stated right on the reference panel. Honest about what it doesn't model.
Built for the years the money goes out
Retirement isn't one decision — it's a decade of them. The projection models the levers that decide whether the plan holds, including the one nobody likes to look at:
- ✓Survivorship — the widow's penalty, faced squarely. At the first death the plan files MFJ → Single: brackets and standard deduction roughly halve, Social Security drops to the larger benefit, spending steps down. You see the survivor's income and tax, not just a success rate.
- ✓A spending guardrail you could actually follow. Guyton-Klinger-style triggers, applied honestly: instead of a blanket cut to all spending, it flexes only the share you've marked flexible or discretionary — a trim you could really make — and it reports the cost in cut years, so a higher success rate never hides its price.
- ✓"When can I retire?" — an optimizer sweeps your retirement age and reports the earliest year your plan clears the success bar you set.
- ✓Home affordability — the priciest home your plan still supports at your chosen success target, the smartest way to pay for it, and what that answer becomes if you wait a year or two (just below).
Every chart in this section is a real capture from a bundled sample — load it and you'll see the same thing.
"How much house — and what if I wait?"
Ask for the priciest home your plan still supports and most planners give you one number. This one shows what the number leans on — then re-solves it against the market you'd have lived through if you waited.
Real capture, near-retirement sample. Wait one year and the answer spans $180k–$290k; wait three and it spans $160k–$320k — the downside falls, the upside rises, and the widening is the answer to "should I wait?"
- ✓The columns are your own return model, not a mood. Weak · 10th, Soft · 25th, Expected · median, Firm · 75th, Strong · 90th are that model's growth over exactly the waiting years — the same five-band ladder the probability fan draws. Soft to Firm is the ordinary course: half of all outcomes land between them, while Weak and Strong mark the 1-in-10 envelope either way.
- ✓Every cell re-solves the house, financing held fixed to the one the main table picked, and prints the portfolio you'd be sitting on at that buy year — so when the wait is actually over, you read your real balance against the column you turned out to be living in.
- ✓One success target, named on the grid. Every cell is solved at the first target in your list, and the heading says which — no quiet mixing of standards between rows.
- ✓It says where it's more confident than the table above. Pinning the waiting years to one outcome removes uncertainty the main table keeps, so those prices can read slightly high — the panel tells you that rather than letting you find out. Under Historical replay the grid switches itself off, because pinning the first years would erase the realized-sequence signal that lens exists to show.
- ✓And the recommendation is scored three ways. Each home the finder lands on is run under Forward · Normal, Student-t and Historical side by side — the same trio as the rest of the projection — with a footnote stating exactly what each model ran with.
Run your own numbers
Load a sample household and open Projection — or start from your own balances. Every figure on this page is reproducible in the app in about a minute.
30 days, every feature, no card and no account — then read-only, never locked.