The Friday Prediction
Every Friday, our prediction engine runs an autonomous 1,000-agent audience simulation to forecast the weekend box office. The agents represent 1,000 diverse audience personas spanning 24 demographic segments (casual moviegoers, genre enthusiasts, book readers, film critics, industry insiders, international audiences, social media influencers, and more) who independently react to a film, debate each other's takes, and converge on consensus predictions across three rounds of deliberation.
On Friday March 19, we published our predictions for the weekend of March 19-21, 2026. The headline call: Project Hail Mary would debut at #1 with an opening weekend of $85-115M, best estimate $98M.
The actual first-weekend cumulative revenue (Thursday preview through Sunday): $141M.
We were wrong. Specifically, we were wrong by 43.9% on our best estimate, and the actual number blew past even our optimistic high-end range by 23%. This is not a rounding error. This is a calibration lesson.
The Full Weekend Scorecard
| Film | Our Prediction | Best Estimate | Actual Revenue | Error | Verdict |
|---|---|---|---|---|---|
| Project Hail Mary | $85-115M | $98M | $141M (cumulative) | +43.9% | UNDERESTIMATED |
| Wuthering Heights | $12-20M (weekend) | $15M | $234M (cumulative) | Tracking in range* | ON PACE |
| Scream 7 | $8-14M (weekend) | $10M | $194M (cumulative) | Tracking in range* | ON PACE |
| Hoppers | $15-25M (weekend) | $18M | $243M (cumulative) | Tracking in range* | ON PACE |
*For holdover films, we predicted weekend gross only. Cumulative totals are tracking within our predicted total worldwide ranges. Weekend-specific actuals are not yet broken out by TMDB.
The Films That Held
Our holdover predictions were solid. Wuthering Heights at $234M cumulative worldwide is tracking neatly toward our predicted $250-350M total range. Emerald Fennell's literary adaptation is performing exactly like the data suggested: strong legs, steady holds, the kind of film that plays for weeks rather than sprinting and collapsing.
Scream 7 at $194M has already exceeded its predecessor Scream VI ($169M worldwide), which we flagged as the franchise floor. Our predicted range of $180-250M total worldwide looks accurate. Horror sequels are among the most predictable categories in our model, and this one is behaving exactly as expected.
Hoppers (Pixar) at $243M is on pace for our $350-550M worldwide prediction. Animation has the longest theatrical legs of any genre (3.8x opening-to-total multiplier in our model), and with a 7.6 TMDB rating, the word-of-mouth runway is long.
Project Hail Mary: Why We Missed
This is the interesting part. Our model had what it considered "high confidence" (72%) on this prediction. The factors we weighted:
- Mega budget ($200M) places it in our 82% accuracy tier for budget-class predictions
- Sci-fi/adventure genre has 63% accuracy in our model (lower than horror or animation)
- Ryan Gosling star power plus Lord/Miller directing pedigree
- Beloved book adaptation with built-in fanbase (Andy Weir's second novel)
- Strong TMDB pre-release score of 8.5
So why did we underestimate by $43M? Three factors we underweighted:
1. The "Martian Effect" Was Stronger Than We Modeled
Our comparable films were The Martian ($54M opening, $630M worldwide), Interstellar ($47M opening, $730M worldwide), Arrival ($24M opening, $203M worldwide), and Dune: Part One ($41M opening, $411M worldwide). The median opening weekend across these comps was $47M. Even our high comp (The Martian at $54M) was well below what Hail Mary achieved.
The problem: we treated this as a "sci-fi original (non-franchise)" rather than recognizing that Andy Weir has effectively become a franchise brand. The Martian was a cultural event in 2015. Readers who loved that book have been waiting for this adaptation. That fanbase creates an opening-weekend floor that our genre-based model did not capture.
2. The 8.2 TMDB Rating Signal
Project Hail Mary currently sits at 8.2/10 on TMDB with 379 votes. This is an exceptional score. For context, The Martian holds a 7.7, Interstellar an 8.7, and Arrival a 7.9. An 8.2 with nearly 400 votes this early suggests genuine audience enthusiasm, not just fanbase inflation.
Our real-time tracker flagged the pre-release TMDB score of 8.5 as a positive signal, but the model only gave it neutral weighting because the early vote count was lower at the time (we ran the tracker on March 19). The lesson: for adaptations of beloved source material, high early ratings are a stronger signal than our model currently weights them.
3. Thursday Previews Are Expanding
The $141M cumulative figure includes Thursday preview screenings, which have been growing as a share of "opening weekend" totals. Our model's comp films from 2014-2022 had smaller Thursday preview shares. The expanding preview window means our historical baselines systematically underestimate opening-period totals for major releases in 2026.
What Happens Next
With $141M after four days on a $200M budget, Project Hail Mary needs roughly $450M+ worldwide to break even (accounting for marketing and theatrical revenue splits). Our original prediction of $600-900M worldwide now looks conservative. If the film maintains the word-of-mouth trajectory suggested by its 8.2 rating, we would revise upward to $700M-1B worldwide.
The Martian made $630M. Interstellar made $730M. If Hail Mary's opening is any indication, it is tracking closer to Interstellar territory. The Lord/Miller directing duo (Spider-Verse, The LEGO Movie) brings a crowd-pleasing sensibility that could push this beyond the cerebral sci-fi ceiling.
Model Accuracy Scorecard
| Metric | This Weekend |
|---|---|
| Films Predicted | 5 |
| Opening Weekend Predictions | 1 (Project Hail Mary) |
| Opening Weekend Accuracy (within 25%) | 0/1 (0%) |
| Holdover Tracking (within total range) | 3/3 (100%) |
| Overall Direction Correct | 4/4 (called #1 correctly) |
| Biggest Miss | Hail Mary: +43.9% vs best estimate |
What We Are Adjusting
Every missed prediction is a calibration opportunity. Based on this weekend:
- Source material fanbase multiplier: Adaptations of bestselling novels by authors with a previous hit adaptation (Weir/The Martian) need an additional 1.2-1.5x multiplier on our opening weekend estimates. We are adding an "established author" signal to our model.
- Thursday preview expansion: Our comp baselines from 2014-2022 need temporal adjustment. Thursday previews have grown from 10-15% of opening weekends to 18-22% for major releases. We are updating our historical normalization.
- High early rating weighting: For films with 8.0+ TMDB ratings and 300+ votes before opening weekend, the model should weight this more aggressively as an upside signal rather than treating it as neutral.
The Verdict
We called the #1 film correctly. We called the holdover trajectories accurately. But on the headline prediction, the one that matters most, we underestimated by $43M and missed our own range. That is the kind of miss that teaches you something.
Project Hail Mary is not just a sci-fi adaptation opening big. It is a data point that reveals where audience simulation models systematically underweight cultural anticipation for beloved source material. Our agents debated the film's potential across three rounds and still converged too conservatively, because the historical comps they were referencing (Dune, Arrival) are structurally different from a crowd-pleasing Andy Weir adaptation directed by the Lord/Miller team.
Next Friday, the model runs again. We will see if it learned.
Data sourced from TMDB, box office tracking, and Hollywood Metrics prediction engine v6.1. All predictions are generated autonomously by our bespoke agentic simulation validated against 5,071 films (2000-2025).