What makes a record previous
A previous record is an entry for a date that has already closed and settled. The archive holds 1,361 of them, covering 1 Jan 2023 to 22 Sep 2026.
The current date is not a previous record until its entry settles at 12:35 IST. Until then the site shows it as awaiting, because a reader who sees a blank cannot tell whether the data is missing or merely early, and those are different facts.
Provisional versus settled
The distinction runs through everything else on this page.
- Provisional. The period is still open. The entry may be revised without ceremony, because nothing has been committed to yet.
- Settled. The period has closed, sources have been reconciled and the entry is published. From this point it changes only through a public, dated correction.
Only settled entries feed the year grids, the month tables and the frequency counts. Mixing provisional values into a statistic means the statistic silently changes underneath anyone who quoted it.
Why archived pages are never edited
When a month closes, its page is frozen at a permanent address. It is not rewritten, not merged into a larger page, and not redirected somewhere more convenient.
The reason is citation. If someone links to a month page and reads a value there, a silent edit leaves them holding a claim that the page no longer makes — and no way to tell. The link still resolves, so nothing looks broken. That failure is worse than a dead link, because a dead link announces itself.
How a correction is published
Corrections happen: sources are re-checked, transcriptions turn out wrong, a reader reports a mismatch. When one is needed, it is published as a dated note on the page it affects, stating what changed, what the value was before, and why.
The correction becomes part of the record rather than replacing it. Anyone who cited the earlier figure can see exactly what moved and decide whether it affects their use. The full process is set out in the editorial policy.
Why past entries say nothing about future ones
This is the most important paragraph on the page. Each entry in this archive is independent of the entries around it. Nothing about the values recorded so far constrains, influences or indicates what any future value will be.
The intuition that says otherwise — that a value absent for a long stretch is somehow
due
, or that a frequently recorded value is running hot
— is the
gambler’s fallacy. It is among the best-documented reasoning errors there is, and it
survives because a long record set will always contain streaks and clusters that look like
structure when you go looking backwards for them.
The frequency tables on this site are published as a description of the archive: how many times each value has been recorded, nothing more. They are not a model, they have no predictive content, and no arrangement of them acquires any. This site publishes no predictions, tips or systems, and treats anyone selling them as selling nothing.
What previous records are genuinely good for
Plenty, as long as the claim stays inside what the data supports:
- Looking a date up — the ordinary case, and the one the structure is built for.
- Citing an entry — a permanent address plus a stated review date.
- Checking a second source — comparing two independent records of the same date.
- Measuring the archive itself — coverage, completeness, and where the gaps are.
Every one of those is a statement about the past. That is the boundary, and this archive stays inside it.
The limit, stated plainly
Historical records describe dates that have already happened. They do not forecast, and no page on this site should be read as though they do.