ConvexPi
← Competitions

The Open Leaderboard

active

The always-open competition. Submit any strategy and see how it ranks out of sample against reference baselines and everyone else. No deadline.

Data & scoring: Graded on a hidden synthetic market: a simulated cross-section of stocks with planted alphas to discover, split into in-sample and out-of-sample. Your score is the out-of-sample Sharpe — the part that can’t be overfit.

Ready to compete?

Sign up for free, run Mission 1 in Colab, and submit your strategy in under 30 minutes.

How it worksSign up free

How scoring works

OOS Sharpe

Measured on a hidden holdout period. The only trusted score.

Overfitting ratio

OOS ÷ IS Sharpe. Target ≥ 70%. Penalizes in-sample curve fitting.

Forward Sharpe

Updated nightly on fresh synthetic data. Tracks real generalization.

How to participate

  1. Follow the getting-started guide — takes about 30 minutes.
  2. Run Mission 1 in Colab. No local Python install needed.
  3. Create an account and submit your strategy from the editor.
  4. Your OOS Sharpe appears on this leaderboard after grading (under 5 minutes).