Zach Cregger spent $38 million dollars. He got back $270 million. That is a 611% ROI โ the kind of return that makes studio executives reconsider every $200M tentpole on their slate.
Weapons is the follow-up to Barbarian (2022), and structurally, it doubles down on everything that made Barbarian work. Our feature analysis reveals a screenplay optimized โ whether consciously or not โ for the metrics that matter most.
The ROI Features
Our correlation data identifies three features with the strongest positive relationship to ROI:
Character Intro Rate (r = 0.068 with ROI). This measures how frequently new characters are introduced per page. Weapons uses a chapter-based, multi-perspective structure โ six interwoven character arcs (Josh Brolin, Julia Garner, Alden Ehrenreich, Benedict Wong, among others). New characters appear throughout the script, each with their own entry point into the mystery. High character intro rate signals narrative density: the audience is constantly processing new information, new allegiances, new threats.
Sentiment Variance (r = 0.032 with ROI). Weapons oscillates between quiet domestic tension, creeping supernatural dread, sudden violence, and dark humor. The emotional range is extreme. Our data shows this variance drives both commercial engagement (audiences talk about emotional roller coasters) and repeat viewing (each watch reveals different emotional textures).
Question Density (r = 0.033 with ROI). What happened to the 17 children? Why 2:17 AM? Why were their arms outstretched? Cregger's script is built on unanswered questions โ a puzzle-box mystery that drives audience engagement through ambiguity. Our data consistently shows question density as one of the strongest predictors of critical reception (r = 0.122 with RT score) and it contributes to ROI through word-of-mouth: people who leave the theater with questions talk about the film.
The Budget Efficiency Signal
Total pages correlates negatively with ROI (r = โ0.063). Weapons, at an estimated 110-115 pages, is lean. Cregger does not waste a page. Every scene either introduces a character, deepens the mystery, or delivers a scare. This page-level efficiency is the structural signature of high-ROI films: say more with less.
Compare this to Mission: Impossible's 150+ page, $400M approach. Both made money. Weapons made profit.
The Puzzle-Box Architecture
The multi-perspective, chapter-based structure of Weapons is structurally unusual โ and our engine would flag it through several features. Scene length variance would be elevated (each chapter has its own rhythm). Character intro rate would be high (new protagonist per chapter). Top 3 character dominance would be suppressed (no single character dominates).
This combination creates a fascinating prediction profile. Our model would likely place it in B-Tier commercially (the structure is too unconventional for massive mainstream appeal) but A-Tier for ROI (lean production, high engagement). The actual result โ $270M on $38M, 94% RT โ validates the ROI prediction while exceeding the commercial forecast.
Predicted tier: A-Tier (ROI), B-Tier (gross), A-Tier (critical). Actual: 611% ROI, $270M, 94% RT. The efficient script wins.