judging-criteria-scoringlisted

# Scoring against the real rubric ## Stage One is a gate, not a criterion Most hackathons screen on pass/fail before anything is scored: does the project fit the theme, and does it genuinely apply the required technology? This is where submissions die cheapest — a missing demo video or an unused required API is a disqualification, not a low score. Run the gate first, and **do not score what fails it**. A number on a non-compliant idea makes it harder to abandon, which is the opposite of what the gate is for. ## "Equally weighted" does not mean equally valuable When the rules say ties break by "the first applicable criterion listed above, then the next," the ordering carries weight the percentages do not show. The top-ranked criterion decides every close call. Record the rank, and when it changes a ranking, say so. ## The score ceiling is often not 5 Bonus contributions are usually worth a fixed increment per published piece up to a cap — 0.2 each to a maximum of 0.6 is common, which makes the real ceiling 5.6, not 5. That is roughly eleven percent of the maximum available for a few hours of writing, and most entrants skip it entirely. Treat it as score, not as marketing. The mechanics matter: the piece must be genuinely public (not unlisted), must carry the host's required disclosure language verbatim, and usually must use a specific hashtag. Missing any of those forfeits the points. ## Choosing a track is an expected-value problem Each track is a separate competiti