← home

Methodology

What you read on FloatVerify comes from filed documents and time-stamped market series. This page explains how they are captured, what we correct, and above all what we refuse to smooth over.

How the data arrives

We continuously capture trades and one-minute bars for the tickers that move, and we retroactively rebuild past months. Every evening, documents filed with the SEC enrich those series: country of headquarters, activity code, and shares outstanding filing by filing — not a single value, but their dated sequence.

Four quality guardians

Four checks run continuously. They do not correct prices: they verify that the series are complete and that no session has been mixed with another.

Intraday freshness
Verifies that the series advance while the market is open, and flags a feed that stalls.
After-close catch-up
Fills the gaps left by an interruption, once the session has ended.
Per-symbol index
Maintains the mapping between a symbol and its history, including after a ticker change.
Split exclusion
Discards days where two price scales coexist — they would manufacture gaps that never happened.

The raw-price rule

A price recorded on a given day stays that price forever. We never re-adjust it after a reverse split.

Our archives show what the trader saw on screen that day, not a series re-smoothed after the fact. That is the difference between a history and a reconstruction.

The consequence is deliberate: a reverse split creates a visible jump in the price series. It is not a display glitch, it is the information — and it is precisely why split days are excluded from the statistics rather than quietly corrected.

Why two float measures are not averaged

Shares outstanding filed with the SEC and the estimated tradable float are two different measures, not two versions of the same number. The float is normally lower: their ratio is information in itself. We do not average them, and we do not choose for you.

An alert only lights up on the impossible: a float greater than the declared share capital. It does not mean a number is wrong, but that one source is lagging — most often after an issuance later than the last filing.

Source laggingThe only use of amber on the site: a contradiction between sources.

Issuances and reverse splits

A rise in the share count and a reverse split are not the same information, and we label them separately. An issuance carries a percentage; a reverse split carries a ratio and no percentage — the two sides of a split are not on the same scale, so a variation computed on it would mean nothing.

History depth

We prefer to show the limit rather than hide it. Rebuilding the months before the start of the archives is ongoing; the counters update on their own.

What we will never do

No intrinsic value. No price target. No recommendation. Sourced, dated facts; the decision remains yours.

This is not a v1 limitation, it is the product.A number that answers "what should it be worth" is an opinion on price, and that is the line we do not cross.