The panel asks one question: each year in retirement, what do I spend? What separates the three answers is what sets the number, and what may change it.
Each choice is a card carrying its own small chart and its own success rate, so you can compare the three before picking one. The card you pick is drawn full size below, with its numbers. The two you have not picked are run at a fraction of the paths, which is enough for the shape and the percentage but not for the last decimal place; the panel's own tiles always describe the choice you picked.
My expenses, as entered. What you typed in Recurring, Housing and Health coverage, every year, whatever markets do. Nothing reacts, so nothing reduces your spending. Pick it when your budget is genuinely not negotiable, or as the baseline the other two are read against.
My expenses, cut back in bad years. The same expenses, with the flexible part trimmed when your withdrawal rate climbs into the danger zone and added back when it falls into the safe zone. How far it can go is your own setting, and the panel states it. Pick it to absorb a bad decade in travel and dining rather than in the balance. The mechanism is the guardrail.
A share of my portfolio. The balance sets the number every year, so spending rises and falls with it. The balance it reads is what you have at the start of the year less any one-time expense your plan commits that year: a house bought for cash is not money to spend a share of. It can move a long way: most this rule can reduce is the whole flexible portion, because a share rule has no floor beyond your essential expenses. Pick it if you want spending to track what you have.
To see the three rules side by side on more than a success rate, run Explore's spending-strategy comparison. Among its columns is confidence by age, which is where a rule that holds up through the early retirement years shows what that is worth.
The two numbers under the picker
Reduction below entered expenses is each outcome's largest one-year reduction below your entered expenses, per month, at the level one outcome in ten reaches or passes. It measures what the rule chose to take off, so a funding shortfall never inflates it, and the choice that adjusts nothing has no tile at all. A $0 reading is a real $0; the line beside it says how often the rule fires. The percentage under it is the same amount as a share of your Recurring expenses, which is the slice both adaptive rules actually move.
Some expenses go unfunded in N% of outcomes is the other side, under all three choices. An outcome counts the first year it cannot pay any one of five obligations: recurring spending, a one-time cost, the ACA premium, the tax bill, or an HSA penalty. That is not the same as running out of money, which is why it carries no dollar figure.
Reading the chart
The bands and the median are planned amounts: they keep following your entered expenses after the portfolio runs short, because a plan that has run out still asks for its full amount.
The separate line marked 1 in 10, as funded is the exception, and the one mark on the chart that can fall. It is what the tenth-percentile outcome actually had to live on: the planned amount less whatever the portfolio could not deliver that year. On a plan that never runs short it sits exactly on the band and you will not see it. Where it drops away is a shortfall, and where it levels off is your guaranteed income, which arrives whatever the portfolio does. It never reads below zero: if that guaranteed income is less than your Medicare premium and the portfolio is gone, an unfunded premium reads as zero spending, not a negative amount, and the shortfall line below still reports that expenses went unfunded. Its end value and the median's are printed at the right edge.
Watch for the two shapes together. A rule whose planned band holds flat while this line collapses is not steadier than one whose band visibly sinks: it is the one that ran out. The success rate beside it is the same fact as a percentage.
Under A share of my portfolio, Spend at most is the ceiling: your entered expenses plus an offset you choose, in percent or dollars a year. What the rule would have allowed above it stays invested rather than being withdrawn, so the band flattens at the cap instead of climbing away from your expenses in a strong market. See how VPW's percentage is set.
For the arithmetic, see how your spending answers to the market.