Most plan-quality problems come from a handful of input choices that look small but compound across decades of projection. These are the ones we see trip people up most often. None of them are hard to fix - just easy to miss on the first pass.
1. Using your best-guess age instead of a longer planning horizon
When you guess how long you'll live, you're picking a typical outcome - there's roughly a 50/50 chance you actually live longer. Planning to that typical age means a coin-flip chance the plan ends before you do, which defeats the point of the plan.
The fix is a buffer: set plan-to-age a few years above your honest expectation so the plan covers the harder side of the distribution. A reasonable starting point is 90 to 95 for most singles and 95 for the surviving spouse in a couple plan. If family history points further out - say a parent lived to 89 - push higher.
2. Account balances entered as today's value instead of January 1
The engine works with beginning-of-calendar-year balances. Entering today's mid-year value adds a slice of growth that the projection then compounds on top of, slightly overstating every year of the simulation. Pull the Jan 1 balance from your year-end statement, or use your broker's "balance on a date" tool. The error is small per account but adds up across a portfolio over 30 years.
3. Leaving Taxable account cost basis at $0
The Taxable account row stores your cost basis as a dollar amount, and the field defaults to $0. That tells the engine every dollar you withdraw is capital gain - the worst-case tax treatment. Most real Taxable accounts have substantial basis (the after-tax money you deposited, plus already-taxed reinvested distributions), so taxes in the projection end up materially overstated when the field is left at the default. Pull your basis from your broker's "tax lots" or "unrealized gain/loss" report and enter it on the account row.
4. Forgetting to enter pre-retirement spending
If you aren't retired yet, the planner still needs to model your spending between today and your retirement date. Without it, the projection assumes you spend $0 in those years and your savings grow unrealistically fast, making the rest of the plan look healthier than it really is. Add a recurring expense entry for current monthly spending that ends at your retirement age, then add a separate retirement-spending entry that starts then. The two will rarely be the same number.
5. Entering state or federal income tax as a recurring expense
Federal and state income taxes are calculated by the engine each year from your bracket, filing status, and projected income - you don't enter them. A common precision-tracker mistake is to add a "State tax" or "Income tax" line under Recurring Expenses to match a real budget. That double-counts: the engine taxes you once via its own calculation, and the expense line taxes you again. If a tax row exists under Recurring Expenses, delete it; the year-by-year projection already shows the engine's calculated tax in its own column.
6. Leaving Medicare start age at your retirement age (or 67) instead of 65
Medicare eligibility starts at age 65 for nearly everyone. The Medicare start age field on the Healthcare page defaults sensibly, but it's worth confirming: if it's set later than 65, the engine models a healthcare gap that probably isn't real. Most people enroll at 65 even if they retire later. If you intentionally plan to delay (e.g. you'll stay on employer coverage past 65), that's fine - just confirm it's deliberate.
7. Retiring before 65 without enabling ACA coverage
If you retire before 65 and don't model how you'll pay for health insurance in those bridge years, the planner shows $0 healthcare cost there and the plan looks artificially strong. On the Healthcare page, enable ACA Marketplace coverage and set the period (typically your retirement year through age 64). The engine will then estimate Premium Tax Credit subsidies based on your projected income and apply realistic net premiums. See Retiring before 65 for the full picture.
8. Picking the wrong "today's dollars" vs "future dollars" toggle on one-time expenses and windfalls
The toggle matters a lot. A "$20,000 expense at age 75" entered in today's dollars inflates forward to roughly $45,000 in nominal terms by then. The same entry in future dollars stays at $20,000. Pick today's dollars when you're thinking in current purchasing power ("a new car costs about $20k today"). Pick future dollars when you already know the nominal amount ("my grandfather's trust pays out $20k on my 65th birthday"). See How inflation works in the plan.
9. Using aggressive return assumptions
The default 8% stocks / 4.5% bonds / 2.5% cash assumptions are deliberately set near long-run historical averages. Cranking them up to 10-12% boosts your success rate on screen without changing what's actually likely to happen in real markets - it just hides risk. If you do want to test more or less optimistic scenarios, change the assumptions deliberately and treat the change as a stress test, not a base case. The Monte Carlo simulation already adds volatility around whatever average you set, so the median path is not what you'll experience year to year either way.
10. Not enabling the spouse flag when married filing jointly
If you're married in real life but Profile is set up as a single filer, every tax calculation in the projection runs as Single. That substantially overstates your tax bill across the whole plan, sometimes by tens of thousands of dollars lifetime. On the Profile page, turn on "Include spouse" and fill in the spouse details. Even if your spouse has no income of their own, the joint filing status matters.
The general principle
Plan inputs flow into a stack of compounding calculations across decades. Small errors get amplified. When something looks too good (very high success rate, very high ending balance) or too bad, check these ten first - the answer is usually here.