The festival list that kept forgetting quiet films
In the dim festival office, a programmer hit play and a sorter spat out a tidy ranked list of films. A volunteer frowned, last year a quiet short from a small town stole the show, now those voices barely appeared. It felt like hiring, films are applicants, the list is the first screen.
They traced a film’s path. First, who even hears the call. Then quick checks and auto flags that decide what gets watched. Then a jury picks the final line-up. Then reviews and reactions shape next year’s sorter. If you only check the final invites, you miss the earlier gates.
The coordinator started a blunt list. Prestige and mates’ networks can tilt things. People drop out if travel is hard, care duties pile up, or a space feels unsafe. The tools can miss too, ads shown unevenly, awkward forms, voice and face checks that fit some people better. Small nudges stack up.
They realised they’d been asking different fairness questions. Is the line-up balanced, or are the scores equally right for everyone. Are some films pushed so low they’re never seen. Is the process respectful. A simple pass rate can look fine while the top of the ranking stays narrow.
Fixes came in three places. Before sorting, they could change the call and the examples used to tune the sorter. During sorting, they could stop it leaning on easy hints like names, faces, or accents. After sorting, they could reshuffle the first pages, but that can mean handling sensitive details at decision time.
Then a worry. Most neat records sit at the early steps, clicks, forms, scores. They knew far less about what happened later, whether people were treated well, earned enough, or kept going. Their info also leaned on a few languages and places, and it often missed disability or forced gender into a tight box. A sudden shock can make last season’s checks stale.
On the final night, the coordinator stopped staring at the sorter like it was the whole story. They watched the whole pipeline instead, who hears, who gets filtered early, who reaches the top of the list, and what feedback writes next year. Tools that judge people from video and faces felt like the riskiest shortcut, so they chose extra restraint.