Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.trepa.io/llms.txt

Use this file to discover all available pages before exploring further.

The is a 100–1000 rating of how good your prediction was on its own, without comparing you to other players. It uses log return error so that being 10% high or 10% low counts the same, and scores stay comparable across BTC price levels. Precision Score is not used for payouts; it drives streaks and leaderboards. The score is calibrated to recent market volatility so a typical 1-minute move is well defined. A score of 500 means your error was about one typical move; above 500 you did better, below 500 the market’s random move was more informative than your guess.
How accurate you werePrecision Score
Perfect prediction1000
About half a typical move off~700
One typical move off500
Two typical moves off250
Three typical moves off~125
Each extra unit of typical error roughly halves your score until the 100-point floor.

​
Formulas (for verification)

  • Log-return error: εi=∣ln⁡(xi)−ln⁡(y)∣\varepsilon_i = \bigl| \ln(x_i) - \ln(y) \bigr|, where xix_i is your estimate and yy is the outcome.
  • Precision Score: PSi=max⁡ ⁣(100,  1000×exp⁡ ⁣(−λ×εi))\mathrm{PS}_i = \max\!\left(100,\; 1000 \times \exp\!\left(-\lambda \times \varepsilon_i\right)\right) with λ=ln⁡(2)/σ\lambda = \ln(2)/\sigma and σ\sigma the recent 1-minute log-return volatility (see Volatility calibration).
Last modified on April 28, 2026