One question per round: which side will the AI judge vote for? Read the matchup, take a position, and the ballot settles it when the round ends. This page is a design preview, so nothing here takes a call yet.
Preview only · Nothing is playable · No credits move · No cash value at any pointEverything below is the real interface running on seeded data, kept here so the shape of the product is visible. The matchups are AI debater personas, not live rounds between real debaters, and the outcomes come from a rating model rather than a judge reading an actual ballot. Calls are locked off until that changes.
Every call you have made, and how you are scoring as a predictor. Conviction rewards calling it early: a Window 1 read (motion hidden) is worth more than a live read.
Called rounds: the AI judge's winner, the crowd's favorite, and whether they agreed. When they split, that gap is the show. Each card carries the reason for decision.
Two boards: the sharpest predictors of the judge, and the debaters the room shows up for.
Not just who wins most. A blended score: how often you call winners, whether your picks beat the crowd, how early you commit, and how hot you are right now. New accounts are shrunk toward the mean, so a 3-0 streak does not outrank a proven read.
Who the room shows up for. Draw Score blends unique backers, watch time, repeat spectators, and how often a debater splits the judge and the crowd. Controversy counts: a debater who keeps winning the flow but losing the room is a walking comment section.
Watch, call it, and the round settles. The whole loop in three steps, then the lens it reads through.
Read the matchup and the round context. Every round shows its format and judge lens before you pick.
Use play credits to back the debater you think takes the round, before the window locks.
The exhibition round resolves from the matchup model, the crowd read appears beside it, and the round is called.
You are predicting the AI judge's winner, shown by the lens on each round. The crowd vote is shown for comparison. Same motion, different lens, different verdict.
The lens on a round is its default. Open a format for what the judge prioritizes and the questions that decide the round.
What the judge prioritizes: clear claims, evidence quality, clash, weighing, and final-focus consistency.
Common deciding questions: Who won the clearest offense? Did either side weigh impacts? Were the key responses extended through summary?
What the judge prioritizes: harms, inherency, solvency and topicality, plus disads, counterplans and dropped arguments depending on the round's lens.
Common deciding questions: Did the aff prove solvency? Did the neg win a disad that outweighs the case? What got conceded on the flow?
What the judge prioritizes: comparative contribution across four teams; extensions and the whip writing the ballot by issue.
Common deciding questions: Whose extension was most necessary? Who won the framing clash? Did the closing add new terrain or just recap?
What the judge prioritizes: whether the motion holds across the likely cases; general-knowledge mechanisms over tagged cards, no fabricated citations.
Common deciding questions: Did Gov defend a fair, central case? Was the burden met across the round? Who won the weighing in the rebuttals?
What the judge prioritizes: the value and criterion resolve before any contention; offense flows under the winning framework.
Common deciding questions: Whose framework controls the ballot? What offense links cleanly to it? What got dropped and conceded?
What the judge prioritizes: reach, magnitude and probability across the principled and practical layer; matter, manner and method scored.
Common deciding questions: Who won the principled clash? Who grounded the impacts? Did the reply crystallize the right two issues?
Most paradigms sit between the flow decides (a dropped argument is conceded) and the better idea decides (real-world plausibility wins). The AI judge names where it landed, so the verdict is never a black box.
In a spar or live round, both debaters can agree the paradigm before the first speech and the ballot holds to it.
The call: who won and the margin, so a near-split and a blowout do not read the same.
The hinge: the one exchange the round turned on, named.
The why: what was extended, dropped or outweighed, judged in the format's paradigm.
The read: per-speaker notes in the register the format scores.
Every event runs two markets: the result market settles the round, and the crowd votes who won. When they split, that disagreement is the whole show. Nothing on The Floor pays cash or prizes.