.png)
Methodology
How the LogicLens score works.
The current score is a neutrality-oriented reading signal. It estimates how much of eligible content appears affected by detected reasoning and framing issues, then gives more weight to issues that appear more severe. This system can change as LogicLens improves.
What the score is
LogicLens currently uses a 0 to 100 score for supported article and blog-style content. A higher score generally means detected issues affect less of the text, appear less densely, or are lower severity. A lower score means the content may deserve closer review.
The score is not a truth rating. It does not mean an article is correct or incorrect. It is a reading signal about neutrality, framing, and persuasive pressure. The individual insights are usually more precise than the single number.
When scoring applies
The score is currently shown only when enough readable text is available and the content type supports scoring. For unsupported content, or where there is not enough scoreable material, LogicLens may show N/A instead of forcing a misleading number.
What counts against the score
The current scoring system focuses on scoreable issues such as loaded language, weak reasoning, misleading framing, missing context, and other persuasive tactics. Neutral or positive observations are not treated the same way as issues that suggest rhetorical pressure.
The two penalties
Coverage penalty
LogicLens estimates how much of the text is affected by scoreable insights. A larger affected portion can lower the score more, especially when the issues appear more severe.
Density penalty
LogicLens also looks at how frequently scoreable issues appear relative to the length of the content. Repeated issues can matter more than a single isolated example.
In plain English: a single small issue should not sink a long article, but repeated or high-severity problems should matter. The exact scoring approach may evolve over time.
Score bands
Author, sources, and unattributed text
When the text allows it, LogicLens can group insights by author narration, quoted sources, and unattributed material. The combined score summarizes the whole piece, while the breakdown can help show where the rhetorical pressure appears to come from.
Important caveat
More emotional topics can produce lower scores even when a writer is trying to be fair. The score can be wrong or overly simple, so it is best treated as a starting point for reading the insights, not as an absolute moral grade.
Current system
This page describes the current LogicLens scoring system. It may change as the product improves, as categories are refined, and as we learn where the score helps readers versus where individual insights are more useful.
Read the FAQ