Every plan is most exposed in its first retirement years. A bad sequence early does damage that typical returns later cannot repair, and the same sequence late is survivable. The fan chart shows this only indirectly, by widening its lower band early, so this reads the same simulation from the other end and says it in your plan's own numbers.
What the number counts
For each age in retirement, the app takes the outcomes that reach that age still solvent and holding at least a given balance, and reports the share of those that go on to fund every obligation to your end age.
So "reach 72 with at least the typical balance and 97% of those outcomes fund everything" is a conditional, not a promise. It describes the outcomes that arrive at 72 in that shape, whatever sequence brought them there. That is also what makes it checkable: a year from now you can look at your own statement and see which shape you are in.
When your plan is already above target today, the sentence still says only what was measured, never what will hold. If every later age in the window also clears the target, it says so: "Already above 95% today, and stays above 95% at every age in the window." If the window dips below target at some age even though today reads above, it names that age instead of promising anything about it: "Already above 95% today; the window's lowest reading is 92.0% at 68." Neither form is a persistence claim: that would need next year's re-run, not this one.
The three balances
The threshold is the run's own balance percentile at that age, taken across all outcomes, depleted ones included. So "the typical balance" is the middle of the fan you are looking at, not the middle of the survivors.
- Not in the bottom quarter is the 25th percentile.
- At least the typical balance is the median. This is the line the chart draws and the sentence quotes.
- In the top quarter is the 75th.
Two things it is not
It does not only rise. It usually climbs with age, because a plan that has survived its early years has fewer left to fund. Nothing forces it to: a late cost, such as a survivor's filing status, a Medicare tier or a large one-time expense, can dip it, and the chart draws the dip.
It is not next year's re-run. Re-running your projection from next year's balances is the real update, and the spending check-in reads your live balances. This is the forward statement, made today, of how confidence should evolve if the years go typically. It is the sentence the fan cannot put into words.
Where it stops, and where it cannot be read
The line ends where every outcome has already gone short. There is no pool left to read a share from, and drawing 0% there would report a measurement nobody made.
Under the Historical replay lens the whole run is about seventy windows, so the median threshold reads a share off roughly thirty-five of them, which is coarse to several points. The sentence says so and stands down; the comparison table shows the number with its window count beside it. Switch to a forward lens to read it properly.
Your own place on it
Name a plan as the one you follow, in Projection's Scenario menu, and the app freezes that plan's expected balance track: what a typical outcome held at each age. Your spending check-in then says where you actually stand against it.
It is frozen on purpose. A track redrawn from today says nothing at today's age, because every outcome starts from the balance you have right now. The one your plan drew earlier is the only one that can tell you whether you are still on the course you set. Following the plan again is what redraws it, and the card says which age it was drawn at.
Because it is frozen, its balances are in the dollars of the year it was drawn. The card converts your balance into those dollars before comparing, at the inflation rate your plan assumes, and names both the year and the rate. Without that step a year of inflation alone would move you up a band while your buying power stood still. It is your plan's assumption rather than measured inflation: the app reads no price index, and under Historical replay the run takes inflation from the record instead.
A track drawn before the app recorded that basis cannot be read this way, and nothing can recover it. Your plan may have changed since it was drawn, so today's rate is not its rate. The card asks you to follow the plan again, which redraws the track in dollars it does record.
The card places you and stops there. It quotes no percentage: a confidence worked out under an older plan's assumptions would disagree with the number on Projection, with nothing on the Dashboard to explain why, so the card sends you here for the live one instead. And in the track's first year every outcome still holds what you hold, so there are no bands to be in and the card says so rather than grading you against yourself.
Zero the volatility inputs and every outcome is the same outcome, so the line reads 100% or 0% and nothing in between: the hand-checkable case every statistic here has.