Top to bottom
Explore the page
- 01
Choose the result to improve
The first full-plan run is the baseline: this plan, with its current income, spending, property, accounts, taxes, funding order, conversions, and estate assumptions unchanged. Every candidate is compared with that same run.
Ranking and guardrails stay in today’s money so changing the display cannot change the recommendation; the Today’s money switch can restate every reported dollar difference in the dollars of the years when it occurs. Grow net worth ranks the final balance sheet with pre-tax accounts counted in full, while also showing and protecting the after-tax reading.
Leave a larger estate ranks what beneficiaries receive after modeled debt, tax, and estate costs. Improve tax efficiency ranks lifetime tax saved, subject to an after-tax-wealth guardrail. Maximize after-tax money to spend adds the spending actually funded across all years to the after-tax wealth left at the end, and rejects a path that raises that total merely by spending less.
- 02
Available actions
The switches under Set the boundaries decide which families may enter the search.
Depending on what this plan actually holds, those families can include contribution and catch-up changes, excess-cash investing and reserve choices, retired-year spending rules, flexible-spending trims, debt-first surplus, withdrawal and forced-sale orders, Social Security claiming ages, investment-property sales, a primary-home sale, retirement timing, relocation, investment fees, Roth conversion schedules and amounts, qualified charitable distributions, gain harvesting, and capital-loss treatment.
The catalog uses discrete, explainable rungs rather than changing every number continuously. Major lifestyle moves begin off; an enabled family with nothing applicable to this plan is named in the coverage notes instead of being presented as tested.
- 03
How an action is tested
Every ordinary catalog single is applied to the same baseline inputs and sent through the funded-plan engine. The run settles annual income, spending, contributions, investment growth, debt, account withdrawals, required distributions, property sales and costs, federal, state, local and payroll tax, Medicare effects, shortfalls, ending balances, and the estate again.
Named and grouped withdrawal orders are also priced as singles; deeper raw permutations are tested only when the explainable orders show that order can materially change this plan. A row appears in the ledger only for a move the search actually priced, so the result is not an estimate pasted onto the baseline.
- 04
Combining actions
The search completes eligible ordinary singles and explores a bounded set of combinations. Safe singles above the materiality floor seed a diverse frontier, with leading moves from different categories and no more than two expanders per family.
Before raw withdrawal permutations use the remaining allowance, a reserved slice tests interactions even when one or both singles look weak: cash reserves with surplus investing, spending with contributions, property sales with debt or tax planning, and other cash and tax interactions. Some of that slice tests pairs outside these explicit rules.
The slice is capped at 24, 72, or 160 runs by depth and at 10% of the remaining allowance. Every complete path must pass the same guardrails. Existing claiming, withdrawal-order, and Roth interactions also remain. Strong safe pairs grow into three-move paths; Thorough reaches four moves and Exhaustive reaches six. Same-family alternatives and standalone sensitivities cannot share a path.
- 05
Choose a search depth
Choose Standard, Thorough, or Exhaustive.
Depth widens both the rungs offered by each lever and the frontier the search can afford to grow. Standard allows paths of up to three moves with a nominal budget of 450 full runs and 15 seconds. Thorough allows four moves, 1,600 runs, and 60 seconds.
Exhaustive allows six moves, 6,000 runs, and 120 seconds, and can walk neighboring withdrawal-order permutations that the shallower modes do not. After ten candidate runs the worker measures this plan’s pace and lowers the remaining run allowance when the time limit would arrive first.
The baseline and the core single-move catalog are completed before guarded combination stages; if a later stage reaches its budget, the result says where the search stopped rather than implying that every possible path was priced.
- 06
How results are ranked
A candidate cannot be recommended if its projection fails to settle, brings the first unfunded year forward, materially increases unmet spending, creates or accelerates an unplanned property sale, or changes which unplanned property is sold unless that is the move being tested. The spending goal also rejects a result bought by lower funded spending.
Tax savings are rejected when ending after-tax wealth falls materially, and a positive gross result for another goal is rejected when it is really deferred tax rather than added after-tax wealth. The materiality floor is the larger of $100 and 0.01% of the baseline value for the selected finish line.
Safe paths above it rank by rounded goal impact, then lower disruption, fewer moves, and a stable id. The winner is checked against its own subpaths; a shorter path that lands within the floor is preferred. Paths producing the same modeled outcome are folded behind one representative rather than shown as different answers.
- 07
Compare results and try a change
The impact strip shows the combined effect of the recommended actions on your selected goal, ending net worth, estate, lifetime tax, and after-tax spending money.
Compare the actions opens the best path, a table of tax actions, and a comparison with less disruptive alternatives and other ranked paths. Additional safe combinations and near-misses appear below.
The ledger shows each tested action on its own: impact, other effects, effort, disruption, first affected year, verdict, and caveats. Exact duplicate runs within a family are grouped together.
Try a whole path from the comparison or one action from the ledger. Both open a What-If preview based on the saved plan. The separate Roth sweep can also pass its schedule into What-If. Choose Keep or Save as new to save, or Revert to restore the baseline.
What changes what
Controls & settings
Check what each control changes and whether it recalculates your forecast.
Chooses the one finish-line measure used to rank every eligible candidate: ending net worth, net estate, lifetime tax saved, or funded spending plus ending after-tax wealth. Changing it marks the visible answer out of date; Update results starts the new search.
Include or exclude whole decision families from the next search. Switching a family on does not guarantee that the plan has an applicable move; the coverage notes explain when it does not. Update results applies the new boundaries.
Sets the floor beneath flexible-spending trims and retired-year rules: 5%, 10%, or 20% below the spending written in the plan. It is available only while at least one spending family is allowed and takes effect with Update results.
Changes the action rungs, withdrawal orders, combination frontier, run budget, and time budget for the next run. Each depth uses its longest supported recommendation: three moves at Standard, four at Thorough, and six at Exhaustive. A deeper run retains the previous winner when the plan and allowed changes still match. The estimate is a budget or the count used by the last search, not a promise that every possible combination will run.
Starts the requested search, including Roth schedule preparation when applicable. Existing results stay visible while it runs. An answer from different inputs is marked out of date and cannot be applied until results are updated.
Stops the active worker and keeps its latest completed checkpoint. Partial results remain readable, and may be tried in What-If only while they match the current inputs and boundaries.
Starts another search allowance from the retained checkpoint when its inputs still match, reusing completed projections instead of pricing them again.
Switches every displayed result between today’s purchasing power and the dollars of the years when the money occurs. It does not rerun, rerank, or change any recommendation or guardrail.
Change which already-priced single moves are visible and how they are arranged. They do not rerun a candidate or alter the ranking.
Applies a complete compared path, one ledger row’s move, or the independent Roth sweep’s exact schedule to a frozen copy of the plan and reruns the normal projection so every chart and plan page can be compared with the baseline. Nothing is saved until the What-If is kept or saved as a new plan.
Reading the result
Read the results
The headline is a difference from the baseline on the finish line selected for that search. Read the companion after-tax, estate, tax, and spending figures before treating a larger headline as an all-around improvement.
Best means best among the bounded candidates that were actually priced and passed the modeled guardrails. It does not mean a global mathematical optimum over every possible amount, date, account order, or life decision.
A recommended path’s combined result can differ from the sum of its moves’ standalone ledger results because taxes, benefit timing, withdrawals, and account balances interact. The path card’s standalone, marginal, and running readings come from separate full runs; the complete path rerun is the answer, not a sum assembled by hand.
Helps, hurts, and immaterial are measured against the search’s materiality floor. Guardrail means the candidate’s chosen metric may look attractive but another modeled outcome made it unsuitable for a recommendation.
Two differently worded actions can produce the same projection when this plan never reaches the account they reorder or when two amount rungs clamp at the same limit. The comparison folds duplicate paths and the ledger folds duplicate singles behind one representative, stating how many equivalent runs it represents.
One plan, several readings
Related pages
Goal Plan →
Choose a specific target and search for a small set of changes that reaches it.
Projection →
Strategy Lab takes its baseline from the same complete funded projection drawn on the Projection page. Trying a ledger move returns there through What-If, with the saved baseline under the edited run for comparison.
Tax Planning →
Roth bracket candidates come from the Tax Planning sweep’s solved schedules, then are repriced against Strategy Lab’s own baseline and allowed to interact with withdrawal order, claiming ages, and the other enabled tax moves.
Plan resilience →
Strategy Lab is deterministic: it ranks paths under the return and inflation assumptions entered in the plan. Plan Resilience is the separate place to run a candidate through many dealt market paths.
Estate →
The estate finish line uses the Estate page’s modeled debt, tax, liquidation, administration, and beneficiary assumptions to calculate what reaches beneficiaries at the end of every candidate run.
Keep following the math
Key concepts
Keep in mind
Model limits
This is a bounded search over the discrete actions and rungs the page exposes. It is not a continuous optimizer and does not prove that no untested amount, date, sequence, or combination could score better.
Every ranking is deterministic under this plan’s returns, inflation, tax snapshot, household dates, and other assumptions. It is neither a probability nor a guarantee that markets, law, income, spending, or behavior will follow the modeled path.
The search never invents a higher investment return. Moving costs, replacement housing after a primary-home sale, transaction execution, tax-lot selection, and rules outside the projection are not filled in on the household’s behalf.
A guardrail is a consistency screen inside this model, not a finding that a move is personally safe, feasible, or suitable. Legal, tax, benefit, and investment consequences still need the relevant records and professional judgment.
Trying a compared path, a ledger move, or the independent Roth sweep is only a What-If. The saved plan remains unchanged until the household explicitly keeps the edit or saves it as another plan.
This guide documents the page as built. The figures on it follow the plan’s own assumptions; none of it is individualized tax, legal, Medicare, or investment advice.