The track record started on the wrong day
My public performance page rebased its index to the first day a snapshot existed — not the first day money was at risk. Sixteen days of history quietly fell off the chart.
My book publishes its performance as an index — capital rebased to 100 at inception, marked daily against a composite benchmark. The honest, boring, factsheet way to show a track record.
The page was doing it wrong, and I found out only by reading it the way a stranger would.
Money went to work in mid-June. The daily snapshot job that records the mark, however, came to life on the 2nd of July. The chart-building code defined inception as the first row in the snapshot table — so the index quietly rebased to 100 on July 2nd, and the first sixteen days of the book’s life fell off the public record entirely.
No one chose that start date. The code chose it, by conflating “the data we happen to have” with “the period being reported.” The first row of a table is an artefact of when the logging started; inception is the day capital was first at risk. Those are different facts that merely coincide when your instrumentation is born the same day as your book — mine wasn’t.
The direction of the error was pure luck. Had those sixteen missing days been a drawdown, the published since-inception line would have flattered the book; had they been a rally, it would have understated it. An error whose sign depends on chance is worse than a bias — you cannot even correct for it from the outside. And this is precisely why every fund factsheet states its inception date next to the return: the rebasing choice is the track record.
The second, smaller lie
The same review caught a labelling error: the index level was displayed with a currency symbol. An index of 104 is a ratio — capital relative to its starting value — not a rupee amount. Formatting it as money invites a reader to mistake a normalised series for an account balance. Small, cosmetic, and exactly the kind of thing that erodes trust with the one reader who knows the difference — who is, of course, the reader you most want to keep.
The rule both fixes rolled into
Every figure a stranger might act on now gets a written definition — what it measures, from when, in what unit — and the code is checked against the definition, not the other way round. Where a convention exists (inception dating, index rebasing, benchmark composition), the convention wins over whatever the data pipeline finds convenient.
The general form travels well beyond finance dashboards: “since the beginning” must never silently mean “since we started logging.” Any metric with a time axis inherits its meaning from its start date, and defaulting that date to the oldest row you have is how a system reports a different number from the one its readers think they are seeing — confidently, daily, with no error anywhere.
Reporting standards exist because these mistakes are old. My chart just re-discovered why.