49ers Truthiness

How we rate the outlets and creators we aggregate. Published methodology, version 1.0, October 2026.

What it is

We rate the work, not the person.

A disclosed, data-driven rating of 49ers coverage, in the spirit of Ground News but built for sports media. Readers see at a glance how much of a source's output is sourced reporting versus speculation, how often its verifiable claims prove correct, and whether it corrects itself when wrong. These scores track published work: what was claimed, what checked out, what got corrected. They say nothing about anyone's character or motives. Think batting average, not report card.

The three dimensions

Each scored 0-100 from the aggregation log. All three use only observable evidence, never vibes.

1. Sourced

Share of aggregated items that qualify as sourced reporting: named sources, anonymous sources with a stated basis, documents, or direct observation like practice reports and press conferences. Opinion and analysis pieces are excluded from this dimension entirely. Opinion is not a sin; undisclosed speculation is.

2. Track Record

Of the source's falsifiable claims (trades, signings, injuries, roster decisions), the share proven correct. Each claim is logged when we aggregate it and marked confirmed, refuted, or left pending. Ambiguous outcomes stay pending. Pending claims older than 90 days are dropped, never counted as wrong.

3. Accountability

Whether the source issues corrections or updates when it gets something wrong. Every source starts at a neutral 70. Prompt, clearly labeled corrections move it up; refuted claims left standing move it down. Scored from observable corrections only. We never guess at intent.

The badge

Truthiness is the average of the three dimensions. 80-100 earns a gold High badge, 60-79 a Solid badge, below 60 a Mixed badge. The badge always links back to this page. Scores recompute from the log on every run and are never hand-edited.

When ratings appear

No source is scored until it has 30 days of aggregation history and at least 20 stories in our log. Before that, its card reads "Collecting data." A rating without a sample is just an opinion with a chart.

Safeguards

We rate the work, not the person. Paid outlets are never penalized for the paywall. A source can dispute a score by pointing to a correction or resolved claim we missed; disputes are resolved within 7 days. If our log has a gap, affected scores show an asterisk until the sample recovers.