Set your odds before seeing what the market believes. Each question is a real, CFTC-regulated event contract trading on Kalshi right now — with real money behind the price. When each one settles, reality grades you both.
Copying the market scores zero by definition. You only earn edge where you disagree — and turn out right.
The market’s current price is hidden until you commit.
Each settlement updates your calibration curve. Come back tomorrow morning — the verdicts arrive while you sleep.
Once your forecasts settle, this is drawn from your own record: when you say 80%, how often does it happen? Perfect calibration hugs the diagonal.
Roughly a fifth of the world’s oil moves through that strait. The reopening reports leaked midday Wednesday; official confirmation and the fee backlash followed Thursday and Friday.
KXWTIMAX-26AUG31 · $95 strike · here is the offer price through Wednesday’s repricing:
Anchor — Wednesday midnight ET: ≈11¢ (mid of 7 bid / 15 ask)
Days of official confirmation and fee outrage later — where does it trade now?
Every question here is a real event with a real, verifiable outcome — a Fed decision, an inflation print, tomorrow’s temperature. Each one trades as a yes/no contract on a regulated exchange (Kalshi), where the price is the crowd’s probability, backed by real money: a contract at 36¢ means the market collectively believes 36%.
You are not asked to beat a quizmaster. You are asked to out-judge that crowd — and reality itself grades the paper. A crossword’s answer exists when you solve it. Predicta’s answer key doesn’t exist yet.
Play: five questions each morning. For each, drag the slider to your probability — not yes or no, but how confident — and lock it. Only then does the market’s price appear.
The gap is the game. Lock 30% on a hurricane question the market prices at 22¢ and you’ve staked an 8-point disagreement on the YES side. If the storm comes, that gap banks in your favor at settlement; if not, it counts against you. Match the market and nothing is staked — which is why looking up the price first is pointless, not forbidden: copying the crowd scores zero by definition.
Scoring, plainly: when a question settles, both you and the market are graded with a proper score — your probability versus what actually happened. Your running result is always shown relative to the market’s. Beating a real-money crowd, even slightly, over dozens of settlements — that’s the badge.
The calibration curve is the deeper mirror: across all your settled forecasts, when you said 70%, did it happen 70% of the time? Most newcomers discover a dip in their 70s — confident claims that land barely better than coin flips. Watching that dip flatten over weeks is the real progress bar.
Play: a real headline, the related market, and its price from before the news. You dial in where you think the price sits now, then see the actual tape. Ninety seconds, scored on direction and magnitude.
What it trains: was that “news” actually news? Markets reprice on leaks and surprises, then shrug at official confirmations. When a dramatic announcement moves the price a penny, the information arrived earlier than the headline. Announcements aren’t information. Surprises are.
Questions settle on their own clocks — tonight’s weather by morning, Tuesday’s CPI at 8:31, the Fed in September, the long game in 2028. Each edition seeds your week with verdicts-in-waiting; the morning visit opens with whatever reality decided overnight. Showing up to be graded is the habit that improves you.
Occasionally, after you lock a big disagreement with the market, a gated panel appears. It opens only for players with a proven calibration record in that domain — at least 20 settled forecasts there, with a positive edge over the market across your last 20. Evidence, not enthusiasm. No record, no nudge, ever: Predicta measures judgment first and refuses to point unproven forecasters at a trading floor.
Predicta is stake-free: no deposits, no prizes for predictions, nothing to buy. Prices come from Kalshi’s public market data via our own service, shown as the midpoint of the best bid and ask. Your forecasts are stored under a random anonymous ID created by your browser — no account, no name, no email. We collect anonymous gameplay analytics (which screens are used, when forecasts lock) with no cookies, no IP addresses kept, and nothing sold to anyone. None of this is investment advice; any trading anyone chooses to do happens on a regulated exchange, as an adult, at their own risk.
Your forecaster ID in this browser is —. The ID — and the record attached to it — lives only where it was created: playing from another phone or browser starts a separate record under a different ID. Keep playing where you started if you want one unbroken track record; a way to claim and merge records under one name is planned alongside the leaderboards.
Questions, bugs, ideas: support@predicta.game