How Scoring Works

Every sermon gets a single PSR score from 0–100 — a weighted average of eight categories. Biblical Accuracy is denomination-neutral: it measures whether the text is handled faithfully, not whether the theology is “right.”

Category	Weight
Biblical Accuracy	25%
Time in the Word	20%
Passage Focus	10%
Clarity	10%
Engagement	10%
Application	10%
Delivery	10%
Emotional Range	5%

The Three Colors

74+ (green) 60–73 (yellow) below 60 (red)

Grounded in the text

Before scoring, PSR builds an independent study of the sermon's primary passage — consulting the original languages and trusted commentaries — so the score reflects how faithfully the sermon handles what the passage actually says, not just how it sounds.

Admin & technical reference (bonus points, detection patterns)

Hebrew / Greek / Aramaic detection

Scans the transcript for references to the original biblical languages. Matches within 20 characters are de-duplicated.

Language + noun — "the Greek word agape"
In the language — "in the Hebrew, this means…"
Language introduction — "Greek:" / "Hebrew,"
Translation reference — "translated from the original"
Original text — "the original Greek says…"

Church history detection

Scans for ~50 historical figures (Augustine, Aquinas, Luther, Calvin, Wesley, Spurgeon, Edwards, Bonhoeffer, …) plus event/term patterns (councils, creeds, reformation, revival, church eras, century references). De-duplicated within 30 characters.

Bonus point formula

Admins can nudge the score with bonus points. Each row totals independently:

Row Total = min(Max, |Count × Bonus|) × sign(Count × Bonus)

Total Score = PSR + Total Bonus (clamped 0–100)

Biblical Accuracy & Time in the Word rows

These pull the sermon's AI category scores (read-only) and let an admin apply a manual adjustment when the AI over- or under-scored a category.