Predict closer. Win bigger.
Accuracy-focused price forecasting on Solana: how you win, how payouts work, and why closeness beats direction alone.
Trepa is a forecasting contest on Solana, backed by Colosseum and Balaji. You predict asset prices at short intervals, starting with Bitcoin. You pay a fixed entry fee per round, the same for everyone.
Traditional prediction setups are often binary: yes/no, up/down, or right/wrong. On Trepa, two players can both miss the exact price, but the one who was closer wins when errors are ranked. Winning means beating the median error in that round. Because the cutoff is the median, about half of the field wins each time in a typical pool, so your baseline expected win rate per round is roughly 50% before skill. The exact rule and edge cases are on Winning and losing.
If you win, you get your entry fee back in full, plus a share of the prize pool built from losers’ fees. Among winners, the closer you were to the outcome, the larger that share. If you lose, you forfeit your entry fee to fund that pool and fees. How money moves in aggregate is summarized on Payout overview.
Every round you also get a Precision Score (0–1000) for leaderboards and streaks. That score is separate from the win/loss cutoff but still rewards precision.
