The Plan Audit Pack: Every Assumption, On the Record

RetireOdds now lists every assumption your plan actually runs on, tagged as yours, auto-derived, or a default, plus a printable audit report.

I was fairly sure I had entered our expected investment return myself, a specific, considered number Mayur and I had argued about for an evening. Turned out the field had reset at some point, quietly fell back to a model-derived default, and the plan had been running on that instead for weeks. Nothing broke, nothing looked wrong on the dashboard, and that's exactly the problem: a plan can look confident while running on an assumption you never actually chose. You cannot spot that kind of drift by staring at a result. You need the plan to tell you, plainly, what it's actually built on.

What the plan is running on, not what you think you entered

That's the whole idea behind the Plan Audit Pack. It lists every assumption the plan currently runs on — resolved, meaning the actual value in effect right now, not whatever you vaguely remember typing in — and tags each one with where it came from:

• "You set this" — a value you entered directly.
• "Auto" — derived by the model itself, like an expected return calculated from your actual stock-and-bond mix rather than picked out of the air.
• "Default" — a stand-in value in use because nothing more specific has been provided yet.

Seeing that tag next to a number is the difference between a plan you trust and a plan you're guessing about. If "default" shows up next to something load-bearing, that's a five-minute fix. If it shows up unnoticed, it's a plan quietly running on a stranger's assumptions.

The data behind the data

Alongside the assumption list sits a set of data-quality warnings — stale account balances, missing IRA basis, defaults still in use where you meant to enter something specific. Each warning points at the exact field, so fixing it isn't a scavenger hunt through settings you haven't opened in months.

And because the tax rules themselves are inputs too, their provenance gets disclosed the same way: which tax year the constants are pinned to, when they were last verified, and that they trace back to primary sources — the IRS and CMS — rather than a summary of a summary someone else already simplified.

One click, the whole thing on paper

One click produces a printable plan report that pulls everything together: accounts and liabilities, spending, every assumption group with its provenance tags, saved scenarios, and the latest simulation complete with run-record hashes, so a specific result can be traced back to the exact inputs that produced it rather than trusted on faith. It also carries your Roth conversion schedule, tax projection, healthcare projection, key risks, and a known-limitations list — the places the model is upfront about what it doesn't handle.

None of this changes what your plan says. It changes whether you can check it. A projection you can't trace back to its inputs is a projection you're taking someone's word for, and in our case that someone was occasionally a stale default field, not either of us. The report is meant to be handed to anyone who wants to poke holes in the plan, including a spouse who does not trust a dashboard on faith.

Key takeaways

• Every assumption the plan runs on is listed as resolved, tagged "you set this," "auto," or "default."
• Data-quality warnings surface stale balances, missing basis, and defaults still in use before they quietly bias a result.
• Tax constants carry their own provenance: the tax year, verification date, and primary source.
• One click builds a full printable report, including run-record hashes that tie a simulation back to its exact inputs.

Open the Plan Audit Pack and see what your plan is actually running on.