Already have a plan? Get a second opinion
If you already pay for retirement planning software, you have one answer. What you do not have is a way to tell whether it is right.
Running the same plan through a second engine is the cheapest check available. And when the two disagree, the disagreement is the useful part.
Two planners rarely differ because one is broken
They differ because they assume different things, and most of those assumptions never appear on screen. A ten-point gap in your chance of success usually resolves to a single input that neither tool asked you about directly. Finding which one is worth more than either number by itself.
These are the places we would look first, in the order we would look.
The biggest factors that move results
- 01
The inflation rate
One of the most consequential assumptions in the plan. It compounds against every expense for thirty years, so a difference of a percentage point dwarfs most of what you could argue about.
Ours: We default to 2.54% general and 3.36% for medical costs - CPI-U and its medical-care index over the same 1994 to 2024 window, rather than two numbers chosen independently. Healthcare carries its own rate because it has run persistently hotter than everything else, and retirement spending is unusually healthcare-heavy. Both also line up with the defaults planners commonly publish, so a comparison turns on how the engines behave rather than on where they started.
- 02
What a return rate actually means
The other number that moves everything, and the one most often quoted without saying which number it is. The average of a series of returns and the growth you actually compound are not the same figure, and volatility drives them apart: at a seventeen percent standard deviation the gap is worth about a point and a half a year. Two tools can both say eight percent and be describing materially different futures.
Ours: The defaults are 8% for stocks and 4.5% for bonds, before inflation, and what you enter is the arithmetic average. The simulation sets its drift so simulated returns average to exactly that, which puts the compound result nearer 6.5% for stocks at the default volatility. The standard deviations and the stock-bond correlation are settings you can see and change rather than constants.
- 03
What the simulation actually varies
Some engines randomize only market returns. Others also vary inflation, spending, or lifespan. More moving parts widen the range of outcomes, which can look like caution but is really a different question being answered.
Ours: Returns vary across paths. Cash flows are deterministic, so a difference between two runs is attributable to markets and nothing else.
- 04
How returns are drawn
Drawing from a normal distribution and sampling blocks of real historical sequences produce noticeably different tails from the same average and standard deviation. Downside percentiles are where you will see it first, not the median.
Ours: Both are available, along with a straight-line mode, and the choice is yours rather than buried. Runs use 2,000 iterations.
- 05
Whether a spending rule is switched on
A plan that trims discretionary spending when withdrawals climb will always score better than one that spends straight through a downturn. When this is on by default it is invisible, and it is the largest single source of gaps people cannot explain.
Ours: Adaptive Spending exists and is off until you turn it on, so the baseline number is the unassisted one.
- 06
How your state treats retirement income
States vary enormously in how they tax Social Security, pensions, and IRA withdrawals. A blended or national-average tax model cannot show you that, and the error compounds across the whole projection.
Ours: The full federal and state engine runs on the free tier, including state-specific treatment of retirement income.
- 07
Whether both engines model the same things
The differences above assume the two are doing the same job with different settings. Sometimes they are not doing the same job at all: one may not price Medicare surcharges, or subsidies before 65, or how your state treats a pension. Even success can mean different things, since it counts simulated lifetimes meeting some condition and the condition is not standard. An omission is the hardest thing here to catch, because it never appears as a zero, only as an answer that is quietly cleaner than reality.
Ours: Success means one thing: the money did not run out, with a legacy goal reported beside it and never folded in. The year by year table itemizes what produced each figure, so something we do not model is a column that is not there rather than a number that is slightly low.
Matching the assumptions is the comparison
Two plans built on different assumptions will always produce different numbers, and that tells you nothing. The check only means something once both sides assume the same things, which is why the list above is ordered the way it is: work down it, make the two match, and whatever difference survives is real.
Before any of that you have to rebuild the plan here, which is retyping rather than importing. It goes much faster if you collect the numbers first and enter them second, so we wrote down which ones to gather, in what units, and the handful that are easy to get wrong.
Or print the capture sheet (PDF)
Worth saying plainly: setting your own return and inflation rates is a Plus control, not a free one. The free tier will build the plan, run the simulation, and tax it against your actual state, but it will not let you dial the assumptions to match another tool. That is the one place free does not get you all the way.
Free to start, three months to compare in full
Free, with no sign-up
Enough to find out whether the two engines agree on the core of your plan. Your data stays in your browser unless you choose otherwise.
- The complete plan: accounts, income, Social Security, pensions, expenses, and debts
- Monte Carlo, historical, and linear simulations
- Federal and state taxes against your actual state of residence
- The funded ratio, in today's dollars
Plus, $24* for three months
Long enough to rebuild a plan and compare the parts that matter.
* Summer sale, through August 31. Normally $29.
- Your own return and inflation assumptions, to match the other tool
- What-If Explorer and the Optimizers
- Roth conversions, Medicare and IRMAA, ACA subsidies
- Time Machine, to replay the plan through real market history
A second opinion is only worth as much as the engine behind it
Which is a fair thing to ask of us before you spend an evening on this. So we publish it. Our federal tax calculations are checked scenario by scenario against an independent open-source tax model, and the results are posted with the method and the version.
You do not have to take our arithmetic on faith any more than you should take anyone else's.
Find out where the two disagree
Start with the free tier and no account. If the numbers line up, you have confirmation. If they do not, you have something worth chasing.