Skip to content
Tapeboard. DATA

DATA METHODOLOGY

How Tapeboard sources its market data

Every dated page on Tapeboard is a frozen snapshot of what a named source said at a named time. This page documents the pipeline, the cadence, the retention limits, and the licensing behind those numbers.

The daily data pack

After each US market close, Tapeboard captures the day into a single data pack: top gainers, losers, most-active names, unusual volume, 52-week highs and lows, the day's trading halts, the short-squeeze leaderboard, large-cap market breadth, and news-catalyst headlines for the day's biggest movers. The pack is published as /dated-data/YYYY-MM-DD.json with a markdown render beside it, and the dated pages (/movers/archive/{date}, /halts/{date}) render from that frozen record. Once published, a pack is permanent; the same numbers stay citable at the same URL indefinitely.

Each section of each pack carries its own capture timestamp and status. If a source was unavailable at capture time, the section is recorded as not recorded — the dated page says so plainly. Gaps are never filled with estimates or invented values.

What each section is, and where it comes from

WhatSourceCadenceLicense / basis
Movers (gainers, losers, most-active, unusual volume, 52-week lists) Tapeboard scanner snapshots, frozen per market day by the movers-archive job Per market day, ~4:30 PM ET (after the US close) Schwab delayed quotes; delayed-redistributable
Trading halts NASDAQ Trader tradehalts feed, parsed by Tapeboard's halt pipeline Live feed; frozen into the pack on capture Public exchange feed
Short-squeeze leaderboard FINRA short interest (bi-monthly prints) + IBKR stock-loan availability + market data Recomputed daily; the pack freezes the day's board Licensed market data; scores are Tapeboard-computed
Market heat (large-cap breadth) Tapeboard's tracked S&P 500 large-cap set, regular-session close Per market day Schwab/Yahoo delayed quotes
News catalysts Yahoo-sourced stock-news dataset via Tapeboard's bridge, with publisher headlines Dataset refreshes daily; the pack freezes headlines on capture Publisher headlines, linked to source

Capture cadence and the halt-feed retention limit

Mover snapshots are frozen by the archive job shortly after the US close (~4:30 PM ET). The daily pack is then assembled the same evening, so the halt feed still holds the day's halts: the upstream NASDAQ Trader feed retains currently-halted and recent entries, and a same-evening capture is the complete record for the date. Since 2026-09-26 the halt pipeline also writes a per-day archive snapshot of every halt it observed (updated as resume times post), so a pack rebuilt later for a missed evening reads that archive instead of relying on feed retention. Days with no snapshot at all mark halts, the squeeze board, market heat, and catalysts as not recorded rather than guessing.

Licensing and disclaimers

Mover, heat, and squeeze rows are built from delayed market data (at least 15 minutes) under a delayed-redistributable license. Market data is for informational purposes only and is not investment advice. Short-interest figures are bi-monthly FINRA prints; each squeeze row is stamped with the as-of date it was computed from, and borrow-fee data reflects stock-loan availability as reported. News-catalyst headlines are the publishers' own words and link back to the source.

Frequently asked questions

What is a dated data pack?
One JSON file per US market day (tapeboard.com/dated-data/YYYY-MM-DD.json) holding that day's frozen snapshot: top gainers, losers, most-active names, unusual volume, 52-week lists, trading halts, the short-squeeze leaderboard, large-cap breadth, and news-catalyst headlines for the day's top movers. A markdown render of the same pack sits next to it. Packs are generated after the market close and are permanent once published.
Why do some sections say "not recorded"?
Tapeboard only publishes numbers its own pipelines actually captured. If a source was unavailable when a pack was generated, the pack records that section as not recorded and the dated page says so — it never fills gaps with estimates, averages, or invented values. Halts have a known retention limit: the upstream feed keeps currently-halted and recent entries, so a same-evening capture is the complete record for the date. Since 2026-09-26 the halt pipeline also archives every halt it observes during the day into a per-day snapshot, so a missed evening is recoverable from that archive rather than lost.
Are the quotes real-time?
No. The mover snapshots are built from delayed quotes (at least 15 minutes) under a delayed-redistributable license, which is what allows Tapeboard to publish the rows on public pages. Live, real-time streaming data is a product feature inside the Tapeboard scanner, not part of the dated public records.
Can I cite or reuse these numbers?
Yes, with attribution and a link. Every dated page and pack carries its own capture timestamps and source stamps, so a citation can point at the exact frozen record (the page URL or the JSON pack). The short-interest inputs are bi-monthly FINRA prints; each squeeze row is stamped with the short-interest as-of date it was computed from.
What happens when data is wrong?
A pack is a frozen snapshot of what the source said at capture time; it is not retroactively edited. Material corrections are handled on the corrections page and, where a correction affects product data rather than the frozen record, fixed in the live pipeline so the next capture is right.

Where to check next

Data sources and disclosures — the per-vendor table for every data type on Tapeboard · Movers archive — browse dated records · Corrections — our correction policy.