Quantitative model measuring casting collision probability across A24, HBO, and independent feature projects.
Quantitative model measuring casting collision probability across A24, HBO, and independent feature projects.
This explorer is a source authored scenario, not a measured probability that two actors compete for a role. Its three archetype sliders begin at ninety two, eighty seven and eighty six. Their mean is two hundred sixty five divided by three: eighty eight point three percent when rounded. The code clamps that mean between forty five and ninety nine point four. The initial HTML says eighty eight point five, but startup replaces it with the computed eighty eight point three. Move the everyman slider to its native minimum of twenty. With the other two unchanged, the sum becomes one hundred ninety three and the displayed mean becomes sixty four point three. The anxious best friend now has the largest weight, so it becomes the primary archetype. The score drops below seventy and the label becomes Divergent Roles. Those labels are thresholds in the code, not independently observed casting outcomes. Putting all three sliders at twenty would average twenty, but the floor displays forty five. At one hundred each, the ceiling displays ninety nine point four. Export JSON to inspect the actual inputs, result and selected dialogue entry. Offline, the required D three library is missing: the numeric sliders and export can work, but the network graph cannot initialize. No missing library was fetched and no source was repaired. The actors, quotations and narrative links are saved page content rather than facts verified by this explainer.