Back to Community

Ensemble methods for combining multiple alpha signals

COCovarianceKid
Jan 29, 2026
1,004 views
15 posts
strategy
machine-learning

I've been experimenting with different ensemble techniques to combine signals from my individual models:

  • Simple averaging -- Equal weight across all signals
  • Inverse-variance weighting -- Weight by 1/variance of each signal
  • Stacking -- Use a meta-learner (Ridge regression) on out-of-sample predictions
  • Dynamic blending -- Adjust weights based on recent performance
  • Stacking has given me the best results so far. Anyone else exploring ensemble approaches?

    14 Replies

    19
    SHSharpeShooterJan 30, 2026

    Great analysis! I've been using a similar approach with rolling z-scores and it's been working well for mean reversion signals.

    20
    GRGradientHunter3d ago

    Be careful with the Kelly criterion for position sizing. Full Kelly is way too aggressive. I use quarter-Kelly in practice.

    9
    RERegimeDetectorFeb 22, 2026

    The competition scoring docs could definitely be clearer. I spent 2 hours debugging what turned out to be a normalization issue.

    Post a Reply