Methodology
Six factors. Named. Public record.
Every score is a weighted blend of six named factors — no black box, no mystery AI, no chat assistant required. You see each factor's contribution on every ticker.
The six factors
Composite = weighted blend of 6 sub-scores
Trend
How far price has moved over a multi-month window, and where it sits inside its own 52-week range.
Relative strength
The stock's price change minus the broad-market benchmark's, over three horizons. Not sector-adjusted.
Fundamentals
Reported margin, return on equity, EPS and revenue growth, and an earnings multiple.
Smart money
Net direction of insider transactions disclosed to the SEC on Form 4.
Macro
The prevailing market regime, applied identically to every ticker on a given tick.
Momentum
Short-horizon rate of change: a momentum-quality reading blended with a short-horizon return.
Each factor has its own page: what it measures, how the reading is derived, which data feeds it, and where it is weak. The honest limits are stated on each factor’s own page rather than collected out of sight.
How the blend is weighted
The composite leans most heavily on Trend and Relative Strength, then Fundamentals, Smart Money, and Macro, and least on Momentum— short-term momentum on its own tends to mean-revert, so it's the smallest input. Each sub-score is normalised to 0–100 and the same weighting is applied to every ticker.
The factor set is fixed and visible, and each factor's contribution is shown on every ticker. Any change to the factors ships through the public scorecard and changelog.
Per-ticker confidence
Not every signal has the same evidence behind it.
We surface a confidence percentage on every row. It varies based on which underlying data feeds returned data for that ticker — not every ETF has a P/E, not every stock has recent insider filings, not every name has analyst coverage.
Signal labels
Descriptive, not prescriptive.
Each score maps to a label that describes the state of factor data. We never tell you what to do with it.
Common questions
Methodology FAQ
How is the Tapeline Score calculated?
+
Each ticker gets six sub-scores (Trend, Relative Strength, Fundamentals, Smart Money, Macro, Momentum), each normalised to 0-100. The composite is a weighted blend of the six — weighted most heavily toward Trend and Relative Strength, and least toward Momentum. The factor set is fixed and public; any change ships through the public changelog.
What do the signal labels mean?
+
Six descriptive tiers map to score ranges: HIGH CONVICTION (85-100), STRONG SETUP (70-84), CONSTRUCTIVE (55-69), NEUTRAL (40-54), CAUTION (25-39), WEAK (0-24). Labels are descriptive, not prescriptive — Tapeline is not a registered investment adviser and does not issue buy/sell calls.
How often does the score update?
+
Scores are recalculated on each worker pass, about every 60 seconds during US market hours, and persist between sessions. Prices are delayed about 15 minutes. Most score inputs are daily readings (daily price bars for trend, relative strength and momentum, and the macro regime), so a score usually changes about once a day; fundamentals refresh on company filing cadence. Insider Form 4 filings are read from SEC EDGAR and re-checked about every two days per stock (about monthly for ETFs).
What is the per-ticker confidence percentage?
+
Confidence reflects how many of the underlying data feeds returned data for a given ticker — not every name has fundamentals coverage, recent insider filings, or analyst coverage. 95%+ means full data on every signal feature; under 40% means sparse data and the score should be deprioritised.
Is the methodology really public?
+
Yes. We name all six factors, show each factor's contribution on every ticker, and log each day's top 10 to a public scorecard that's back-checked against SPY the next session. Since the record began on 11 May 2026, four trading days have no list: 31 August, 2 September, 4 September and 9 September 2026. Any change to the factor set ships through the public changelog. The moat is the data spine plus that public track record — not a secret list of factors.
Before you trust any of this
Read the parts that argue against us.
Limitations
What Tapeline is not good at: a small public sample, a blunt six-factor screen, and data that can be late or wrong.
Public scorecard
Each day’s top 10, back-checked against the next session, with the sample size shown and losing days left in. Four trading days since the record began have no list.
Why it works this way
A note from the founder on publishing the method and the record, including the days the picks went nowhere.
Methodology changes, data errors and corrections are dated in the changelog.
See the scores in the scanner.
An email and a password puts you on the free plan, where every scan returns its top ten scored rows. A card starts the 30-day Premium trial — every matching row, $0 charged that day, first charge on day 30, one click to cancel.
Create your account →More transparency artefacts
Tapeline's moat is everything you can audit before you sign up. These pages all live publicly — each is independent evidence.