How improvements are tested

The Ways to improve your chances screen tests common changes on your own plan. It doesn't recommend any of them. It measures them.

Changes tested

Only changes that fit your plan are included:

• Retire later: everyone one year later, and each person separately.
• Save more until retirement, by about a fifth of your flexible spending.
• Spend less: 10% less overall, and 25% less on each of your two largest flexible items.
• Claim Social Security at 70, for each person not already claiming then.
• Roth conversions filling the 12% or 22% bracket, from retirement until the oldest person's required distribution age, kept under the next Medicare surcharge tier. Tested only if your plan has no conversions set.
• Withdrawal order: filling the 12% bracket from pre-tax accounts first, or taxable, then pre-tax, then Roth.
• Spending guardrails at your comfort level, if your plan uses steady spending.
• Moving: to a state with no income tax, or a planned move a year sooner.
• Property: selling a second property, or downsizing your home to about 60% of its value at retirement. Both assume 6% selling costs.
• Fees: lowering fees to 0.10% on accounts that charge more than 0.15%.
• Stock share: 10 percentage points more or less in stocks.

Smallest change to reach a goal

For a target chance of success you pick, the app searches for the smallest:

• delay in retirement, from 1 to 60 months
• increase in savings, from $50 a month up to all your flexible spending
• cut in spending, from 1% to 40%

It tests the largest size first. If that reaches the goal, it narrows in on the smallest size that still does.

Same markets, change by change

Each change runs on the same simulated futures as your plan: the first 300 of them, with the same random seed. The app compares each future with and without the change. This paired comparison removes luck from the difference, so smaller effects can be measured reliably.

Each result shows:

• your chance of success without and with the change
• the difference, with a 95% margin of error
• how many futures the change helped and hurt
• the effect on median ending net worth

Ranking

Results are sorted by how much they raise your chance of success, then by how much they raise median ending net worth. A change counts as clearly helping when it adds at least half a percentage point and the gain is larger than its margin of error. Others are grouped under Little or no clear change.

The headline figure for each change is your plan's full-run chance of success plus the paired difference.

Related tools

What if lets you vary inputs yourself. Its What matters most tab tests each of these up and down: