← Dilution Check

How this sample was built

The ten companies behind the dilution check are a hand-reviewed validation sample, not an automated scan and not comprehensive coverage of the market. This page states the commitments the sample was built to. The exact accession numbers, timestamps, and excerpts checked for each of the ten companies live in the project's own review ledger, not here — this page is the public method, not the per-ticker paperwork.

Selection contract

  • · US-listed operating companies, each with its own SEC CIK, read directly from their own filings.
  • · At least two facilities that are active or priced, and at least two that are only possible -- authorized but never drawn on.
  • · At least one facility with disclosed remaining capacity, and at least one that is exhausted, terminated, expired, or withdrawn.
  • · A card whose reviewed record does not settle remaining capacity reports it as unknown, with the reason, rather than guessing.
  • · Warrants, convertibles and cash are named as outside this check's coverage rather than left to be read as absent.
  • · At least five of the ten cards anchored to a filing accepted within twelve months of the sample's publication date.

Review procedure

  • · Every figure, date, and quote traces to a specific document fetched from sec.gov during the read: an accession number, the SEC's own acceptance timestamp, a source URL, and a verbatim excerpt of at most 320 characters.
  • · Company press releases, vendor summaries, search results, and issuer investor-relations pages are never cited to support a displayed claim, even when they are easier to read than the filing itself.
  • · Every document read is recorded by (accession, document path, content hash), so a filing that is re-rendered with identical words is the same version and a filing whose words changed is a new one.
  • · What a filing does not establish is recorded as unknown, with the reason, rather than inferred or estimated -- an honest unknown is a first-class answer here, not a gap to paper over.
  • · Each issuer's filing history is swept forward from its last successful read, so a card is never published as current while a later filing that would change it sits unread.

What a card is made of

Every card is read by an automated parser directly from SEC filings. No person reviews one before it is published, and no price data is used anywhere: every fact below comes from a filing.

  • · Shelf registration, its status and expiry -- from S-3, S-3ASR, F-3, F-3ASR filings and the SEC's EFFECT notice, matched to the registration by SEC file number.
  • · An at-the-market or equity-line program, its counterparty and its size -- from the 424B5 prospectus supplement covering it, or the periodic report describing it.
  • · How much of a program is left -- from the issuer's own statement of remaining availability in a 10-Q or 10-K, and only when that statement names the agreement it is about.
  • · Priced takedowns off the shelf -- from a 424B5 with a fixed share count.
  • · Public float -- from the dei:EntityPublicFloat XBRL cover-page fact.
  • · Shares outstanding -- from the dei:EntityCommonStockSharesOutstanding XBRL cover-page fact.

Every figure carries its own status

There is no single word for a company. A card publishes each fact with a status of its own, because a shelf we can cite and a balance we could not attribute are not the same kind of answer and should never be averaged into one.

  • · supportedA filing says so, and the card cites the filing it says so in.
  • · unknownWe read the filings and could not establish it. The card says which reason, in the same words the export and the API use.
  • · conflictingTwo sources disagree, or a calculation would mix incompatible bases -- a quarter's net proceeds against a gross ceiling, say. The card refuses the number rather than picking one.
  • · outside_coverageWe did not look. Warrants, convertibles and cash are outside this check's scope, and the card names them rather than leaving their absence to be read as absence.

When a number is withheld

A total is published only when every part of it is established. If one program's balance is unknown, if a usage report names no agreement we can match it to, if the units differ, or if a document in scope could not be read, the total is withheld — along with every ratio, summary figure and export column computed from it. The card names what was withheld and why.

A withheld total is not a degraded card. Publishing a number computed from a knowingly partial read, and marking the page “partial” beside it, does not retract a figure a reader has already acted on.

“No ATM detected” is a statement about the filings read within a stated window, never a claim that an issuer has no program. The card shows the window it examined, how many documents it discovered, processed and failed to read, and any reference it could not resolve.

What the assessment means

The assessment describes THIS run's processing. It is never a verdict about the issuer.

  • · readyEvery document in scope was read and every published figure is supported.
  • · partialSomething could not be read or resolved. Figures that depend on it are withheld.
  • · unavailableThis run established nothing. A statement about the read, not about the company.

Re-reading

Each issuer is re-read on a schedule, and also whenever a new filing appears, the parser changes, or an instrument's stated expiry passes. A failed attempt is recorded and does NOT count as a check: a card whose last successful read is old says so rather than looking fresh. A correction supersedes the card it replaces and the superseded card is kept, unedited.

What this sample is not

It is not a screen, a ranking, or a universe. It is not a claim about which companies are most likely to issue equity, and an authorized instrument is never evidence that issuance is occurring. It is not comprehensive coverage — a ticker outside the published set is simply not covered, not cleared. Every figure here is research context about what the reviewed filings say, not a signal to act on — see the disclaimer below.

Research context, not investment advice.