We ran the same three retirement plans through Boldin, ProjectionLab and our engine

When people ask whether they can trust a retirement planner, the honest answer is: check it against something else. So we did. We built three realistic plans, entered each one by hand into Boldin and ProjectionLab, two of the most complete planning tools available to individuals, and put the results next to ours, year by year.

Where the three tools land together

• The nest egg on the day you retireWithin 1% on two plans; the third differs by a timing choice, explained below
• About ten years into retirementWithin 2% of Boldin on all three plans; within 1.5% of ProjectionLab on two
• Federal and state income tax, year by yearWithin $850 a year of Boldin, every year checked from 2033 to 2059
• Social Security at the claiming ageThe same in all three, to the dollar
• A required 401(k) withdrawal with a spouse 12 years youngerWithin $141 of Boldin
• The order money comes out (cash, brokerage, 401(k), Roth last)The same in all three
• The odds the money lasts, on random returnsWithin 2 points of Boldin on all three plans

The three plans

Every plan used the same stated assumptions in every tool: 6% returns, 4.5% on cash, 2.5% inflation. Balances are savings only, in the future dollars each tool shows.

ProjectionLab can’t move a plan from one state to another, so Morgan and Riley are compared with our engine kept in California. The odds the money lasts: Alex and Jordan 81.7% ours, 80% Boldin; Sam 87.2% and 89%; Morgan and Riley 98.7% and 99%. ProjectionLab’s odds use history rather than random returns, so they aren’t compared.

Where the tools differ, and why

Over 30 or 40 years small choices compound, so the lines drift apart late in each plan. Every gap we traced has a specific cause:

• When in the year money moves. We take each year’s withdrawal at the start of the year; ProjectionLab takes it later, after part of the year’s growth. That alone moves ProjectionLab ahead by about 0.3% a year.
• Saving while one spouse still works. Our engine counts Riley’s future 401(k) saving in the household’s starting figure; Boldin and ProjectionLab add it as it’s made. It’s the same money on a different clock, which is why Morgan and Riley start further apart and finish close.
• Where leftover money goes. When required withdrawals are more than you spend, we and Boldin reinvest the rest; ProjectionLab holds it as cash, whose interest is taxed every year.
• State rules. Georgia excludes up to $65,000 of retirement income a year from 65; Boldin leaves the exclusion out. Boldin grows North Carolina’s standard deduction with inflation; the state fixes it at $25,500.
• Health insurance before 65. Boldin doesn’t estimate the marketplace premium credit, so we entered ours. ProjectionLab estimates its own, somewhat larger than ours for Sam once his pension starts. Most of Sam’s later gap with ProjectionLab comes from this and the timing above.
• The year of a death. ProjectionLab treats the life-expectancy year as the year someone is gone; we and Boldin keep them through that year.

What we refined along the way

Lining three tools up year by year is also the best way to find rough edges. We found a few in our own engine, each affecting a specific situation, and fixed them within a day; the figures on this page include those fixes. The changelog lists each one and who it affects.

Check it yourself

Download the three plans and open one in the app with Inputs → Restore from a file. Restoring replaces the plan saved in your browser, so if you have one, save it first with Back up everything.

Then expand any year in the Projection table and press Check this year yourself to see that year as plain arithmetic: the balance, the growth, the required withdrawal, each Social Security check, and where every dollar went.

Read the full study →Every input, how it was entered in each tool, every year we compared, and every tax figure.

What this doesn’t show

• These are three plans we designed, run once in each tool in October 2026. Tools change; so does ours.
• Year-by-year figures use steady returns, so they compare the math, not the markets.
• The plan files hold October 2026 balances. Opened in a later month, the app counts from that month, so its figures shift a little from these.
• Agreement isn’t proof that any tool is right. It’s evidence that independent tools, built separately, reach the same answers on the same inputs. Where they differ, we’ve said why.
• This isn’t an endorsement or a criticism of anyone. Boldin and ProjectionLab are serious tools; our Boldin and ProjectionLab pages cover how they differ from ours as products.