Welcome to Trepa, a precision forecasting platform where you predict real-world outcomes and get rewarded based on how accurate you are. This page explains how prediction pools work, how results are determined, and how payouts are calculated during the beta phase on Solana devnet.
Trepa is not a typical binary prediction market. Instead of betting “Yes” or “No,” you forecast where a real-world metric (like CPI, stock price, or ETH gas fees) will land — and your reward scales with your accuracy, not just correctness.
Forecast Range: Each pool asks you to submit a numerical forecast for a specific real-world metric (e.g., “What will the NASDAQ-100 close at today?”).
Stake: You back your forecast with a USD-denominated stake. You can start with as little as $0.10.
Time Cutoff: Pools lock prediction submissions 1 to 24 hours before the final value is revealed or announced by an official source. Cutoff times vary by event and are clearly shown in the UI.
Edits Allowed: You may update your forecast until the cutoff, but doing so may reduce your time bonus (early forecasts get rewarded more).
Data Source: Each pool specifies an authoritative or official reporting entity (e.g., Nasdaq, BLS, BEA) as the source of truth for the final result.
Final Outcome ("Expiration Value"): This refers to the verified value published by the data source at the designated reporting time. If you'd prefer a different term, let us know, for now, we’ll use “Expiration Value.”
Settlement Timing: Pools are typically settled within 24 hours of the data becoming available.
Trepa uses a parimutuel-style payout structure: losers fund the winnings for the most accurate predictors.
Your payout is based on:
Prediction Accuracy: Defined by your normalized error: {Normalized Error = |Your Forecast – Expiration Value| / Benchmark Error}. Values are scaled between 0 and 1, where lower is better.
Stake Size: The more you stake, the more you stand to earn.
Time Bonus: Submitting predictions early increases your reward weight. Late edits reduce this bonus.
Return Cap: Maximum ROI is capped at 100x your stake.
Prediction Window: The period during which forecasts can be submitted and edited.
Cutoff Time: The moment beyond which no more edits or new predictions are accepted.
Expiration Value: The final, verified outcome reported by the authoritative source.
Normalized Error: Your forecast error scaled between 0 and 1 for fair comparison across pools.
Pool: "What will the NASDAQ-100 close at today?"
Your forecast: 17,401.3
Official value (Nasdaq): 17,398.7
Your absolute error: 2.6
Your normalized error: e.g., 0.18 (depending on the benchmark)
If your prediction ranks among the most precise, you’ll win a portion of the reward pool, potentially up to 100x your stake.
Trepa is in beta. If something looks off or you want to dispute a pool outcome, join our Discord and contact the team within 24 hours of resolution:
