How the Numbers Work

WorthSync is a manual tracker, so every figure on your screen traces back to balances you typed on dates you chose. This page explains what happens between those entries and the numbers on your dashboard: how periods are normalized, how each planning engine works, and where the outside data comes from.

It’s the long version of the short explanations you’ll find behind the info icons inside the app. Nothing here is a secret formula — if a number ever looks wrong, this page should tell you why it looks that way.

Part 1 — Definitions & normalization

The ledger

There is one source of truth: a snapshot — an account, a date, and a balance. There is no assumed schedule. You can log three balances in March and none in April, and WorthSync will not invent the missing month.

Everything below exists because that freedom has to be reconciled with questions like “how did I do this quarter?”, which assume tidy period boundaries that your ledger doesn’t have.

Closest-to-boundary period math

The period tiles (1W, 1M, QTD, YTD, 1Y) all work the same way:

1. The boundary date is calculated from today — 7 days back, 30 days back, the first day of the current quarter, January 1, or the same date last year.
2. The baseline is your most recent snapshot at or before that boundary.
3. The change is your latest snapshot’s total minus that baseline.

So a “1M” change compares your newest balance to whatever you last recorded on or before 30 days ago — not to a March 31 figure you never entered. Deltas are never computed by bucketing snapshots into months and joining them up.

When a tile shows a dash: if every snapshot you have is newer than the boundary, there is no baseline and WorthSync shows nothing rather than guessing. A brand-new account won’t show a 1Y change until it has a year of history behind it.

When a chart needs one point per month, the same rule applies at month scale: the snapshot closest to the end of each month wins.

The year-over-year pivot

Long histories get unreadable if every month stays on the chart forever, so the trend collapses older data:

One exception: if all your history sits inside a single calendar year, no collapsing happens. Otherwise you’d be left with one lonely point and no trend at all.

This is display normalization only. Your snapshots are never merged, rewritten, or deleted by the pivot.

Archived vs. deleted accounts

These are very different operations, and the difference shows up directly in your net worth.

Because archived accounts are counted by default, the dashboard tells you how much they contribute and offers an Exclude link to take them out of the headline figure temporarily; the reverse link is Include. Excluding is a view preference — it changes nothing in your ledger.

If you closed an account and moved its money somewhere else, Merge is usually the right tool rather than either of these: it transfers the closed account’s history into the destination so your trend line has no artificial cliff. See Accounts & Snapshots.

Carry-forward, estimated, and confirmed

A balance you typed is confirmed. A balance WorthSync produced for you is estimated, and it is always labelled as such — an estimate never impersonates a number you entered.

Estimates come from the Estimated balance source you pick per account:

• Carry forward last balance repeats your last confirmed figure. For a liability with an interest rate and a minimum payment, it applies that interest and payment month by month instead of repeating a flat number. For an asset with a recurring contribution set, it adds those contributions.
• Ticker shares values the account from a symbol and a share count.
• External valuation feed values it from a provider reference.

Three rules bound this:

• An account with no snapshot at all never gets one invented. Carry-forward needs a real starting balance to carry.
• A balance you already logged on or after the target date is never overwritten.
• Confirming the balance yourself clears the estimated mark.

Separately, an account whose last confirmed balance is more than about 90 days old is marked stale. Stale and estimated are different warnings: stale says “this is old”, estimated says “this wasn’t you”.

The dashboard rolls both into a data-health summary. The distinction matters most on the Statement of Net Worth PDF, since whoever reads it is seeing a mix of confirmed and estimated figures — worth clearing the estimated marks before you hand one to a lender.

Household-owned % and the three views

Each account has a Household-owned %, which defaults to 100. Lower it when you only own part of an account — a joint account with someone outside your household, an asset held with a sibling.

That percentage interacts with the net-worth view toggle:

The reasoning: My Net Worth answers “what is on my ledger”, so it shows the whole balance. The household views answer “what does this household hold”, so an account that is only half yours contributes half its balance.

The percentage scales balances everywhere in the household views — totals, top movers, allocation, property equity — not just the headline figure. Accounts at anything other than 100% carry a badge on the account row, and the data-health summary counts them.

Part 2 — Planning engines

Everything in this section lives on the Planning screen, the Debt Center, or the Emergency Fund page, and everything in it is forward-looking — which is to say, an estimate.

Shared assumptions

The planning tools read one shared set of inputs so you enter them once: expected return, inflation, your current age, retirement age, life expectancy, annual spending, and monthly contribution. Change an assumption in one lens and every other lens moves with it.

Two display notes:

• Results default to today’s dollars — future values discounted by your inflation assumption, so “$60,000” means what $60,000 buys now. A toggle switches to raw future dollars.
• Expected return is a pre-inflation assumption about your investments. It is the single input that moves long-horizon results most, and it is a guess, not a forecast.

Monte Carlo projection (P10 / P50 / P90)

A single smooth growth curve is a bad description of markets, so the projection runs your plan through a few hundred randomized return paths instead.

Method. Each month, a return is drawn at random from a normal distribution centred on your expected return, with a fixed volatility assumption (roughly 15% a year while you’re saving, 10% during retirement — you don’t set it). The balance grows by that return, then your monthly contribution is added at month-end. That’s one path; the engine runs several hundred of them.

What the bands are. At each month, WorthSync sorts every run’s balance and plots the 10th, 50th, and 90th percentile:

• P50 — half the simulated futures ended above this, half below.
• P10 — a poor run of markets; 10% of futures did worse.
• P90 — a good run; 10% of futures did better.

An important subtlety: the P90 line is not one lucky future. It’s the 90th-percentile balance at each point in time, so no single simulated path actually traces it. Read the band as a range of outcomes, not three scenarios.

Determinism. The randomness is seeded, so the same inputs always produce the same bands. Reloading the page won’t quietly change your numbers.

What moves the result: starting net worth, monthly contribution, expected return, and horizon. What doesn’t: the volatility assumption and the number of runs are fixed.

FIRE and Coast FIRE

Your FIRE number is annual spending ÷ safe withdrawal rate. At a 4% withdrawal rate that’s 25× your spending; at 3.5% it’s about 29×. There is exactly one FIRE number and it comes from your spending — the real lever is the spending figure, not a multiplier.

Lean / Regular / Fat are shown for context only. They are widely used absolute annual-spend bands — roughly $40k a year or less is Lean, roughly $100k a year or more is Fat, in between is Regular. They are never applied to your own number as a multiplier.

Time to reach it is simulated month by month from your current net worth at your contribution and expected return. If the target isn’t reachable within 100 years the answer is “not at these assumptions” rather than an absurd date.

Your odds are a separate Monte Carlo pass: the share of several hundred simulated paths that finish at or above your FIRE number at your target retirement age. A score in the 80–90% range is generally considered comfortable; below that doesn’t mean failure, it means you’d more likely need to adjust — spend a little less, work a little longer, or claim Social Security later.

Coast FIRE is the balance that would grow into your full FIRE number by your retirement age with no further saving. It’s your FIRE number discounted back by your expected return over the months remaining. Pass it and you still need to cover today’s expenses, but you no longer have to add to investments.

Progress is shown against both your total net worth and your liquid net worth, because a FIRE number funded largely by home equity is not the same plan.

Retirement planner and Social Security claiming

The retirement lens is the most detailed model in WorthSync, and the one that depends most on figures you supply.

Social Security

Benefit amounts are taken straight from your Social Security statement at three claim ages — 62, your full retirement age, and 70. WorthSync does not reconstruct your benefit from an earnings history, and it does not interpolate dollar figures between those three anchors. Claim ages are discrete: 62, full retirement age, or 70.

From those anchors it computes, for each claim age:

• Lifetime total — every payment from the claim age to your life expectancy, each year grown by an assumed cost-of-living increase.
• Present value — those same payments discounted back to today.
• Breakeven age — when the later claim’s running total overtakes the earlier one’s. This is calculated on the same cost-of-living basis as the dollar totals beside it, which puts it later than a simple no-increase crossover.

The recommendation ranks claim ages by present value, i.e. in today’s dollars.

For couples, the model adds two things a single-person calculator misses:

• Spousal top-up — the lower earner can be brought up to half the higher earner’s full-retirement-age benefit. It fills the gap between their own benefit and that level rather than stacking on top of it, it is reduced for claiming before full retirement age, and it is payable only once the higher earner has actually claimed.
• Survivor step-down — when one spouse dies, the survivor keeps the larger of the two checks and the smaller one stops.

The coordinated-claiming view then sweeps every combination of the two claim-age grids and ranks the pairings by household present value.

Income vs. spending drawdown

Year by year, guaranteed income (Social Security plus any pension you entered, each grown by its own cost-of-living increase) is stacked against your spending. The gap is drawn from the portfolio; a surplus is reinvested. The output is whether the money lasts to your life expectancy and, if not, the age it runs out.

A Monte Carlo version replays that same simulation with random annual returns and reports how often the portfolio survives. This is what captures the risk a smooth average can’t: a bad market in your first years of retirement hurts far more than the same drop later, because you’re selling investments to live on while they’re down.

The bridge, sleeves, and required distributions

If you tag retirement accounts by type, the planner splits your portfolio into sleeves — taxable, tax-deferred, Roth, HSA, other — and models retiring before Social Security starts:

• Withdrawals are drawn in order: taxable first, then tax-deferred, then Roth, then HSA, then other.
• From age 73, required minimum distributions are forced out of each owner’s tax-deferred sleeve using the IRS Uniform Lifetime Table. Anything forced out beyond what you needed to spend is moved into the taxable sleeve rather than vanishing.
• Each claim-age combination is scored, and they’re ranked by how often the portfolio survives.

Debt: avalanche, snowball, and 0% promo tranches

Avalanche targets the highest interest rate first; snowball targets the smallest balance first. Both are simulated month by month, because the thing that makes them work can’t be expressed as a formula:

1. Interest accrues on every open debt.
2. Every open debt gets its minimum payment.
3. Everything left — your extra budget plus the freed-up minimums of debts you’ve already cleared — goes to the current target.
4. When the target clears, the whole pool rolls to the next one.

That rollover is why your total monthly outlay stays flat while payoff accelerates. The target order is fixed from your starting balances, which is how both strategies are conventionally run.

The comparison view runs both and reports the interest and months saved. Avalanche normally wins on interest; when the two tie, the shorter payoff wins.

When a plan is flagged as not workable: if a payment never covers the interest accruing on a debt, the balance never falls and no payoff date exists. WorthSync says so rather than printing a date decades out.

Promotional-rate tranches. A tranche is a slice of a debt with its own rate and an optional promo end date. It accrues at the promo rate until that date and at the go-to rate afterwards, and the payoff is simulated month by month across both phases. WorthSync also shows the clear-by payment — the level monthly amount that would clear the balance before the promo expires — which is usually the number that matters.

Payment inference. If you haven’t entered a monthly payment, WorthSync can infer one from your snapshot history by averaging the balance reduction over recent months. Months where the balance rose or stayed flat are ignored, and with too little history it declines to guess rather than inventing a payment.

Emergency-fund runway

Runway is your reserve divided by the monthly essential expenses you set, expressed in months. Two figures are shown because “money you can reach” isn’t one thing:

Retirement accounts, property, and other illiquid assets count toward neither. Liabilities are never part of runway.

Each account’s liquidity label is inferred from its category, asset class, and account type unless you override it, and an account can be split into sleeves so (for example) a brokerage’s cash counts toward runway while its equities don’t. See Accounts & Snapshots.

Without a monthly-essentials figure there is no runway to compute, and the card asks for one rather than showing zero.

Savings rate: contributed vs. market

If you log contributions, WorthSync can separate money you added from money the market gave you:

• Net contribution = deposits − withdrawals over the period.
• Market movement = the change in balance − net contribution.

That’s the whole method, and its honesty depends entirely on your contribution log. If you moved money in and didn’t record it, that transfer will be attributed to the market.

Savings rate is your net contribution as a percentage of an income figure you supply. Without that figure the percentage is hidden rather than shown as 0% — a savings rate divided by an income WorthSync doesn’t know would be meaningless.

Part 3 — Sources & limits

The S&P 500 benchmark

The benchmark answers one counterfactual: if the money you put into your investment accounts had tracked the S&P 500 instead, what would you have today?

Method. Your opening balance is treated as a lump sum bought at the index level on your first snapshot date. Every later contribution buys index “units” at that month’s level, and every withdrawal sells them. The benchmark value on any date is units held × the index level. It uses the exact same contribution stream the savings-rate split uses, priced in index units instead of dollars.

The index series is a monthly total-return S&P 500 series (price plus dividends) reconstructed from Robert Shiller’s public monthly S&P 500 dataset and normalized to 1000.00 at January 2000. It is a bundled monthly table, not a live market feed:

• Values are monthly, derived from Shiller’s monthly averages — not month-end closes. Comparing a single day against it is not meaningful.
• The table is updated periodically. It is not real-time and does not attempt to be.
• Dates before the series starts clamp to the first month; dates after the last clamp to the last.

Which accounts are eligible. Each account has an Investment benchmark setting:

On Auto, WorthSync includes brokerage, retirement, taxable, and HSA-style assets, and excludes cash and savings, liabilities, real estate, vehicles, and vault items. The reasoning is that comparing a checking account or a house against the S&P 500 tells you nothing useful.

Limits worth knowing: the comparison is only as good as your contribution log; it needs at least two snapshot dates; and it is a counterfactual about index exposure, not a measure of your actual rate of return. Benchmarking is part of the Household plan.

Net-worth percentile

The table holds six age cohorts (under 35, 35–44, 45–54, 55–64, 65–74, 75+), each with net-worth breakpoints at the 10th, 25th, 50th, 75th, 90th, and 95th percentiles. The figures are drawn from public 2022 Survey of Consumer Finances-style age-cohort data and are deliberately rounded.

Your position is found by picking your age cohort and interpolating between the two breakpoints your net worth falls between. Above the top breakpoint the estimate extrapolates coarsely and is capped at the 99th percentile; below the bottom one it’s scaled down toward the 1st.

What that means in practice:

• It is directional. The distance between the 90th and 95th percentile is enormous, and rounded breakpoints plus interpolation cannot resolve it. A one-point move is noise.
• It is not adjusted for where you live, your household size, or the year — the vintage is fixed, and it is not inflation-updated.
• Nothing is uploaded, looked up externally, or pooled. Your numbers stay in your account. This is a bundled reference table, nothing more.

It needs your age, which you set in the Planning assumptions, and you can switch it off with Show percentile card if you’d rather not see it.

Valuation feeds are estimates, not appraisals

Ticker-priced and valuation-feed balances are estimates from a third-party source. They are not appraisals, not offers, and not a quote anyone will honour. Feed-valued property in particular can be wide of the mark, and the figures move whenever the provider’s model does.

Any balance that came from a feed is marked estimated for the same reason a carried-forward balance is: you didn’t confirm it. If a number matters — a lender statement, a big decision — confirm it yourself first.

The annual summary is not tax advice

The Annual summary report gathers your year-end figures: net-worth change, contributed vs. market split, debt paid down, and balances grouped by tax treatment (taxable, tax-deferred, tax-free).

What WorthSync doesn’t do

Being explicit about the gaps is part of the method:

• No automatic bank connections. Balances are what you entered or what an estimate source produced. Nothing is reconciled against an institution.
• No transaction-level data. WorthSync tracks balances, not spending, so it cannot categorize purchases or detect an unlogged transfer.
• No cost-basis or rate-of-return tracking. Contributed-vs-market is a balance-difference method, not a time-weighted or money-weighted return.
• No tax modelling anywhere, including in the retirement planner.
• One currency. Balances are not converted or currency-adjusted.
• No inflation adjustment of historical net worth. Past figures are shown as recorded; only forward-looking planning output offers a today’s-dollars view.

Related pages

• Dashboard & Analytics — where these figures appear and how to read the cards.
• Debt & Planning — how to use the planning tools and which plan each needs.
• Accounts & Snapshots — the per-account settings (liquidity, benchmark, estimate source, ownership) that feed all of this.
• Data, Export & Privacy — where your data lives and how to get it out.