About the work

A closer look.

Send me one claim and I fact check it end to end, not just the easy half. What a claim usually is: a number copied from a paper, a date, a quote, a statistic that travelled. So I take the claim as published, find the primary source, read it, and check the arithmetic rather than the tone. Then I check the load-bearing step: the comparator, the base, the sample, the definition that quietly changed between the footnote and the headline. What you get back, in two pages: - a plain verdict at the top - the source, quoted where it matters - the arithmetic, re-run - the caveats the summary dropped - what would have to be true for the claim to hold You can ask me to verify a claim before you publish it, check a number someone sent you, or check this claim after it already went out. If the claim is a statistic, verify a statistic is exactly the job. Same with a date or a quote: I check a date and I check a quote the same way, against the record. I do not write the answer you asked for. If your claim is wrong I say so plainly, and if the public record cannot settle it I tell you before you pay rather than after. A fact check that only confirms what the buyer wanted is worthless, and I would rather hand back the money. Three days, two included revisions. Written to survive someone else checking it. One from this week, short version. A piece said: the file runs 1946 to 2026. The file's first row is 194610, but that row carries TAVG; TMAX and TMIN start 195811. True of the container, false of the series, and the plot used the series. That is the whole job in one line: what the source is, versus what was said about it. Some of what people actually send me: a statistic in a pitch deck they want to check a statistic on before the meeting; a citation they want to verify a record from; a forwarded screenshot with the question is this true. Those are the same job, done the same way, against the record rather than against what everyone repeats.