Corrections
I write about numbers that matter to people's money. When I get one wrong, I want it fixed quickly and disclosed honestly — not quietly edited away. Everything I have had to correct is listed below, oldest error and all.
The log
4 corrections to date. Every one is also marked in the post itself — this page is the index, not a substitute for disclosing it where the error was.
- What it said:
- Published with five documents and the headline claim that every crypto whitepaper measured is harder to read than the tax code.
- Corrected to:
- Widening the corpus to thirteen attempts, of which eight resolved, falsified it. Monero's Zero to Monero scores 52.1, above IRS Publication 17's 49.7. The finding is now stated as seven of eight, and the exception is discussed rather than dropped. The original five scores were unchanged — the claim was simply too strong for the sample it rested on.
Self-caught, by widening the sample after publication.
- What it said:
- Gave the projected single-filer standard deduction as $16,600, with ranges for the joint and head-of-household figures.
- Corrected to:
- The rounding was wrong. I had inferred it from the pattern in the published 2026 figures rather than reading 26 U.S.C. §1(f)(7)(A), which rounds the increase to the next lowest $50 — not the adjusted total to the nearest $50. The corrected figure is $16,550, and the ranges were an artefact of the wrong method.
Self-caught, while checking the statute rather than the pattern.
- What it said:
- Gave the single-person subsidy cliff as roughly $62,600.
- Corrected to:
- The correct figure from the 2026 HHS poverty guidelines is $63,840. This one mattered in the damaging direction: it would have told someone earning $63,000 they had lost all subsidy when they had not.
Self-caught, on re-checking against the HHS guidelines.
- What it said:
- Reported a 17.6% seasonal spread in airfares, with November and December both below average.
- Corrected to:
- The analysis averaged the raw index by calendar month without removing the price trend. Airfares rose 25% across the sample, and months missing a 2026 reading were dragged down by its absence. The real spread is about 10 points, and November is slightly above average — the opposite of what the post advised.
Self-caught, while building a separate analysis on the same series.
How I handle an error
If a figure, date, or claim in a published post turns out to be wrong — whether I catch it myself or a reader points it out — I correct it directly in the post as soon as I can verify the right number against its primary source.
What changes when a post is corrected
- Material errors — a wrong figure, a misstated rule, a broken piece of math in a calculator — get fixed in the text, and the post's “Updated” date (shown next to the publish date at the top of the post) moves to the date of the correction.
- Minor fixes — typos, broken links, formatting — are fixed without a separate updated-date bump, since they don't change what the post claims.
- A post is never quietly deleted to make an error disappear. If a post is retired because it's been superseded by a better, more current one, it 301-redirects to the post that replaced it, and that replacement is disclosed in the newer post.
How to report an error
If you spot something wrong — a stale figure, a source that no longer says what I cited it for, a bug in a calculator — contact me with the post URL and what looks off. I read every message and will verify it against the primary source before changing anything.
Why this page exists
Personal finance content carries real consequences for readers. A visible, honest corrections process is part of taking that seriously — see the Editorial Policy for how sourcing and fact-checking work in the first place.