Verify the record
A stock screener whose record you can download and check yourself.
Most screeners hand you a number and ask you to trust it. You cannot see the method, and you cannot see how the output has actually done. Tapeline publishes both: the six factors behind every score, and a record of each recorded daily top-10 pick, back-checked against SPY the next session — including the days it went nowhere. Entries are not re-ranked or deleted, and every correction to a recorded value is dated.
- The published record — free to read, no account
- 30-day Premium trial — $0 today, cancel in one click
- Pro from $8.25/mo · $99.00/yr
- 30-day money-back guarantee
Three rules that make the record checkable, not decorative
A track record is only worth something if the person publishing it can’t quietly change it after the fact. These three rules are what let you treat the archive as evidence instead of marketing.
Not re-ranked or deleted, and every correction dated
Entries are not re-ranked or deleted. We have corrected recorded values twice, and said so: prices on 25 August 2026, and scores from 18 May to 12 June capped on 15 June 2026. Both corrections are listed on the scorecard, in the changelog and inside the download itself, and trading days with no list are named the same way. If the method changes, the change is dated in the changelog and earlier lists are not re-ranked.
Losing days kept
The sessions where the top-10 went nowhere sit on the page in the same styling as the sessions that went well — same weight, same size, no trophy on the good ones and no burial of the bad ones. Removing them would defeat the point of publishing at all.
Checked the next session
Each pick is scored at one session's close and evaluated at the next session's close, and its move is set beside SPY's move over the same window. The comparison is the same arithmetic for every row, so you can reproduce it from the raw file.
Verify it yourself, in three steps
- 1Download the raw record. Take the archive as a spreadsheet or as JSON — no login, no card. It runs to sessions 7 days old, the same delay as the public page.
- 2Re-run the arithmetic. Every row carries the pick’s move and SPY’s move over the same next session. The comparison is the identical calculation on each row, so you can recompute it — and count the losing days for yourself — straight from the file. The scorecard page says how many of its back-checked entries are newer than the download.
- 3Spot-check a single ticker. Open the recorded entries for any symbol — for example /scorecard/AAPL — and confirm the row on the page matches the row in your download. Cross-reference the current factor readings on its ticker page and the weighting on how it works.
What’s in the download
Each row is one recorded pick. These are the fields you can pull and re-check — descriptive readings, not a performance summary.
| Field | What it is |
|---|---|
| Date | The session the pick was recorded. |
| Symbol | The ticker, as it was ranked that day. |
| Rank | Position 1-10 in that session's list. |
| Tapeline Score | The 0-100 composite recorded at the close. For 18 May to 12 June 2026 every score reads 100: on 15 June 2026 every score above 100 was set to 100, the originals were not kept, and which entries that changed is not known. |
| Prices | The official close on the session and on the next session (restated on 25 August 2026 for entries up to 21 August). Four older entries the data vendor can no longer price, and the flag prices on the 24 August 2026 list, are still the last trade including after-hours trading. |
| Next-session move | How the ticker moved the following session, beside SPY's move over the same window. |
The record is deliberately raw rows and plain counts — nothing annualised, compounded or turned into a hypothetical account balance. The exact column definitions live alongside the export in the data sources page.
What the record currently shows
The sample is small and the result so far is close to a coin flip against SPY — it sat below an even split for weeks. That figure stays off this page and out of every headline on the site by design, so the presentation reads the same whether the number is flattering or not. The numbers, with the sample size next to them, are on the scorecard.
Why a checkable record is the whole point
If nobody outside the company can see the method and nobody can see the record, the score is unfalsifiable — and an unfalsifiable number is a marketing asset, not an analytical one. Publishing the archive, keeping the losing days, and handing you the raw file is what turns “trust us” into “check us”. That is the entire difference Tapeline is built around.
The longer version — the five questions worth asking of any scanner before you pay, and how most of them quietly fail the record test — is in how to evaluate a stock scanner you can actually trust. The reasoning behind publishing losing days at all is on why, the named factors are on how it works, and the honest list of what this product is bad at is on limitations. Two companion explainers go wider than Tapeline: whether stock screeners actually work and how to judge one, and what a real stock screener track record has to contain.
Verifying the record — questions
Can I verify Tapeline's record myself?
+
Yes. The scorecard is downloadable as CSV or JSON with no login and no card; entries appear 7 days after the session. Each row is one top-10 daily pick with its rank, its recorded Tapeline Score, its prices, and how the ticker moved the next session beside SPY's move. The file also lists every correction and every trading day with no list. You can re-run the comparison yourself from the file.
Has the record ever been changed?
+
Entries are not re-ranked or deleted. We have corrected recorded values twice, and said so: prices on 25 August 2026, and scores from 18 May to 12 June capped on 15 June 2026. The score cap was not disclosed until 14 September 2026. Both corrections are dated on the scorecard, in the changelog and in the download.
Does the scorecard keep the days the picks lost?
+
Yes. Losing sessions are published in the same styling as winning ones, with the sample size shown next to the summary. Nothing is deleted after the fact. Publishing only the good days would make the record a marketing asset instead of a checkable one.
How is each pick checked against SPY?
+
A pick is scored at session N's close and evaluated at session N+1's close. Its move over that window is placed beside SPY's move over the same window. It is the identical calculation on every row, which is what makes the archive reproducible from the raw download.
Is the data free to download?
+
Yes — CSV and JSON, no account required. The endpoints are on Tapeline's own origin (/api/scorecard.csv and /api/scorecard.json) so you can script against them directly.
What can't the record tell me?
+
It is a small, growing sample, it is descriptive rather than predictive, and it is not advice. A high score is not a recommendation, and the honest list of what a six-factor screen is bad at lives on the limitations page. Read that before you pay for anything.
See the record before you sign up.
The scorecard is public and the raw data is a click away — no account for either. The scanner takes an email and a password on the free plan; the 30-day Premium trial is the part that takes a card.
More transparency artefacts
Tapeline's moat is everything you can audit before you sign up. These pages all live publicly — each is independent evidence.