The Verification Venue · a head start you cannot buy back
The Ten Years You Only Get Once
Two people save the exact same $300 a month. One starts at 25, the other at 35. Ten years apart, same cheque, and by 65 the early starter retires with more than double. Then the part that sounds impossible: fund only your late twenties, stop cold at 35, and still beat someone who pays in three times as much. It is real, it has a floor, and you can operate both savers until you find it.
The engine is not a promised return. It is a multiplier on time. Every dollar you invest gets multiplied by (1 + r) once a year until you retire, so a dollar put in at 25 is worth (1 + r)10 times a dollar put in at 35 by the time you stop. At a 7% return that multiplier is 1.0710 = 1.97×, and it is true at every positive rate. The head start is real even if markets disappoint. What changes with the rate is how dramatic it looks, not whether it exists.
Below are two savers on one clock. Drag them. Every balance, ratio, and the exact break-even return recompute in front of you from the future value of an ordinary annuity, not from a slogan.
…
Saver A
…
put in … over … · that is … back
Saver B
…
put in … over … · that is … back
…
Minute one. Drag Saver B's start age from 25 up to 35 and watch the gap between the two final balances fan open. Same monthly cheque, ten years of head start, and the balance at 65 goes from a tie to 2.15×.
Minute two. Hit The classic. Now Saver A pays in only from 25 to 35, then never adds another dollar and coasts for thirty years. Saver B keeps paying in the whole time and their money-in counter races past A to three times the dollars. And still B's curve strains and fails to catch A before 65. A put in one third the money and finished ahead.
Minute three, the honest turn. That flip is not arithmetic destiny. Drag the return slider down from 7% and watch the delta banner shrink. At the green break-even tick, near 6.1%, the banner flips colour and B finally wins. You just watched "less in, more out" die with your own hand. It is a fact about high long-run returns, not a law of numbers, and the page will not pretend otherwise.
A convention, not a forecast
The 7% is a convention, not a prediction of your account. US large-cap stocks returned about 10% nominal and 7% after inflation as a geometric average from 1928 to 2024 (Damodaran, NYU Stern). That is one lucky country over one century, with survivorship baked in, and it is the number the detents above are pinned to. Three cautions the fifty compounding blog posts skip:
Order matters: sequence-of-returns risk
Here are five annual returns with an identical average of 5%. Rearranging them changes nothing for a single lump (multiplication does not care about order). But once you are adding money every year, the order changes the ending balance, because a bad decade early lands while you have little invested and a bad decade late lands on your whole nest egg.
…
The pure engine: one dollar, left alone
Strip out the monthly contributions and you can feel raw compounding. This little readout uses annual compounding (a single P × (1 + r)years), separate from the monthly engine above, and clearly labelled so no figure is quoted the wrong way. The default is the famous one: one $1,000 at 10% nominal from 20 to 65.
…
The third caution is quiet but real: fees compound against you exactly as returns compound for you. A 1% annual fee is not 1% of your money, it is a permanent shave off the multiplier, and over forty years it can eat a fifth of the balance. And because 7% here is a real return, it already nets out inflation, so do not "adjust it for inflation" a second time.
The check · every number recomputed in front of you
1. The same-contribution race (both pay $300/mo to 65, 7% real, monthly compounding). Watch the annuity formula FV = M × ((1+i)n − 1) / i, with i = r/12, produce each balance:
2. The stop-early flip. Saver A funds ten years then coasts; Saver B funds thirty. A puts in exactly one third of B's dollars and still ends ahead:
3. The floor. Solved live by bisection for the classic staging: the return where the stop-early saver exactly ties the never-stopping saver at 65. The sign flip straddles the root:
4. The whole comparison, in one table. The durable claims are the ratios (2.15×, 1.15×) and the one-third-in identity: they do not depend on the $300, only on the staging and the rate. Change the contribution above and every dollar figure moves, but these do not.
| frame / saver | starts | pays in until | years in | $ in | ends at 7% | × money in |
|---|
The offline verifier recomputes all of this two independent ways (closed-form annuity and a month-by-month simulation), pins the monthly/monthly convention, proves the ratios are invariant to the contribution, solves the 6.109% floor, and checks the pure-compounding easter egg. Run it: node research/start-investing-at-25-vs-35/verify-start-investing-at-25-vs-35.mjs
Every free choice and every thing this does not promise
One convention, stated and pinned: monthly / monthly / end-of-month. Contributions land monthly, interest compounds monthly, and each contribution is treated as arriving at the end of its month (an ordinary annuity). Switching to an annual convention moves the exact dollar figures by roughly two to nine percent depending on the timing you pick: about 2% for start-of-year deposits, about 6% if you change only the compounding basis and keep monthly deposits, about 9% for a single year-end deposit. Those are real gaps, not rounding, which is exactly why the verifier fails if the two frequencies are ever mixed inside one calculation. The one place this page deliberately uses annual compounding is the "one dollar, left alone" readout, and it says so on its face, because the famous 1,000 × 1.1045 = $72,890 is an annual figure and quoting it as monthly (which would read about $88,000) would be a different number wearing the same words.
The clean "1.97×" is the annual statement of the mechanism. A dollar invested ten years earlier at 7% is worth 1.0710 = 1.97× more under annual compounding, which is the memorable form. Under this page's monthly engine the same ten-year head start on a single dollar is worth (1 + 0.07/12)120 = 2.01×. Both are true; they are one idea at two compounding conventions, and neither is quoted as the other.
The dollar amounts are an illustration; the ratios are the claim. Every balance here is keyed to $300 a month. The whole system is linear in the contribution, so doubling it doubles every balance and leaves every ratio untouched. That is why the headline is "2.15×" and "one third the money in, more out", not "$421,453": the exact dollars depend on a number you chose, the ratios do not.
"Less in, more out" has a floor, and it is not arithmetic. The stop-early flip is contingent on the return being above about 6.109%. Below it, the later saver who keeps paying in wins. At 6% the late saver already edges ahead ($296,093 vs $301,355). The one thing true at every positive rate is the per-dollar head start, (1 + r)years earlier: lead with that mechanism, and scope the staging to the rate.
The late saver "never catches up" is not a general law. At 7% and above they do not catch the stop-early saver by 65, but at lower rates or longer horizons they can, and the instrument shows the crossover age when it exists. Do not sell "never" as arithmetic.
7% is a convention, not a forecast. It is a geometric average over one country's lucky century (US large-cap, 1928-2024, Damodaran), with survivorship and sequence-of-returns risk. Real paths are not a smooth 7%; the order of the good and bad years matters enormously near the end, as the sequence card demonstrates. Treat the dollar figures as an illustration keyed to a named convention, never as a prediction of your account.
What is left out, on purpose. No taxes, no fund fees beyond the note above, no employer match, no changing contributions, no withdrawals. Adding a match or a tax-advantaged account only widens the early saver's lead; fees narrow it. The mechanism (time-in-market is the one lever whose payoff is exponential and cannot be bought back later) survives all of them.