Monthly Roundup: January 2026 — Avatar's Spectacle Paradox and Mercy's Courtroom Collapse

Welcome to the inaugural Hollywood Metrics Monthly Roundup — a data-driven autopsy of the month's biggest winner and most notable underperformer, examined through the lens of our 20 quantitative screenplay features and their correlations with box office and critical outcomes.

Best Performer: Avatar: Fire and Ash

68% RT | 91% Audience Score | $1.475B Worldwide | $400M Budget

James Cameron's fourth Avatar installment is a fascinating case study in what our correlation data can and cannot capture. The film earned $1.475 billion worldwide on a $400M budget — a 3.7x ROI that puts it firmly in profitable territory, if not the runaway success of its predecessors.

The screenplay metrics tell a story of spectacle-first construction. Avatar films are built on extreme CAPS density — all those "MASSIVE EXPLOSION" and "VAST OCEAN VISTA" descriptors that signal visual scope. In our dataset, CAPS density carries a correlation of r = 0.160 with worldwide gross, one of the strongest commercial predictors we measure. Fire and Ash leans into this aggressively.

Similarly, the film's action ratio (the proportion of action description vs. dialogue) correlates positively with box office at r = 0.061. But here is where the paradox emerges: action ratio correlates negatively with critical score (r = −0.025) and audience score (r = −0.062). The very structural choices that drive commercial performance are the ones critics penalize.

This explains the 68% RT score — the lowest of any Avatar film. Our model would predict exactly this divergence: high CAPS density and action ratio pushing gross upward while dragging critical metrics down.

The 91% audience score, however, is the outlier. We attribute this partly to exclamation density (r = 0.069 with audience score) — Cameron's scripts are littered with exclamatory dialogue that audiences read as emotional intensity. And partly to the "Cameron Exception" — visual innovation and immersive world-building that exist beyond what text-level features can measure. Some filmmakers simply operate in a dimension our parser cannot reach.

• Key Features: High CAPS density, high action ratio, elevated exclamation density
• Predicted Outcome: Strong commercial, weak critical — confirmed
• The Exception: Audience score outperforms feature predictions (Cameron effect)

Worst Performer: Mercy

25% RT | 6.1 IMDb | $53.7M Worldwide | $60M Budget

Chris Evans and Leighton Meester star in this AI courtroom thriller that promised a contained, real-time trial structure — 90 minutes of trial roughly mapping to 90 minutes of runtime. On paper, the concept is elegant. In practice, the screenplay metrics reveal a structure at war with itself.

Mercy's premise — a defense attorney arguing against an AI prosecution system — suggests a dialogue-heavy courtroom drama. Dialogue ratio correlates positively with both IMDb (r = 0.042) and audience score (r = 0.060). A strong courtroom script should lean into this aggressively. But the film also attempts thriller-adjacent pacing, pulling the action ratio higher than a pure courtroom drama warrants and diluting the very feature that should be its strength.

More telling is question density — the frequency of interrogative sentences in dialogue. For a courtroom drama, this should be among the highest in its class. Question density carries strong positive correlations with RT score (r = 0.122) and critical score (r = 0.089). The structural opportunity was enormous: a trial script is essentially a machine for generating questions. But reviews consistently cite the film's dialogue as declarative and on-the-nose rather than probing, suggesting the screenplay failed to exploit the format's natural questioning rhythm.

The contained structure also means a likely low scene count and low total pages. While fewer pages can help ROI (less production overhead), total pages correlates positively with worldwide gross at r = 0.142. A thin screenplay generating only $53.7M on a $60M budget means negative ROI — the worst outcome for a contained thriller, a genre that typically thrives on efficiency.

• Key Features: Misaligned dialogue/action ratio, underutilized question density, thin page count
• Predicted Outcome: Courtroom premise should drive critical metrics upward — execution failed the structure
• The Lesson: High-concept formats amplify both good and bad screenwriting; the format alone cannot save weak dialogue

See you next month for the February 2026 roundup. Explore these correlations yourself in the Hollywood Metrics Correlations lens.