Independent Cowboys media analysis

Evidence
before echo.

Cowboys Review traces reports to their source, separates news from speculation, and shows exactly why a claim earns its score.

All ratings are provisional while the 2022–2026 pilot is underway.

Scoring model Auditable inputs

Illustrative claim

“Dallas is expected to complete a player trade this week.”

72/100
Credible report

Supportive evidence, meaningful uncertainty.

Evidence quality30%
Reporter history25%
Independent roots20%
Source directness10%
Specificity10%
Freshness5%

Example only · no real story is being rated here

01 / The model

One rumor.
Three different questions.

01

Story credibility

How strongly does the evidence support this exact claim, right now?

0–100
02

Reporter reliability

How have qualifying original reports resolved in this specific category?

By topic
03

Outlet integrity

Does the publication attribute, link, label, headline and correct responsibly?

6 signals

02 / Source chains

Four posts can still be
one source.

Repetition is not corroboration. We map every citation, embed and rewrite back to its earliest reporting root—and count only truly independent confirmation.

1
Original reporting rootReporter A
Independent
2

Outlet articleCites Reporter A

3

AggregatorCites the outlet

4

Video reactionCites aggregator

1independent
reporting root
4total
occurrences
3maximum
chain depth

03 / Backtesting

Built to survive
messy stories.

We are testing easy confirmations and hard edge cases before any reporter leaderboard goes public.

2022Injury timeline

Dak Prescott’s shifting return window

Can a careful model separate an initial medical estimate from later, better information?

Why it’s in the pilot
2024Contradictory reporting

The Mike Zimmer–Rex Ryan sequence

Expected hire, public denial, renewed contact, then an agreement: every verb matters.

Why it’s in the pilot
2025Source-chain test

Micah Parsons traded to Green Bay

Joint reports, precise compensation and a same-day resolution test speed without rewarding echo.

Why it’s in the pilot

12 stress-test clusters · 2022–2026 · development, validation and holdout sets

See the complete pilot

04 / Non-negotiables

Specific claims.
Visible evidence.
Public corrections.

An unresolved private conversation is not a miss. An aggregator does not inherit a scoop. A correction never vanishes. And a 5-for-5 sample does not outrank a deep record without adjustment.

See every rule and formula