How We Invest
Lone Bear Capital manages a systematic statistical arbitrage strategy designed to generate consistent, market-neutral returns across diverse regimes. The strategy is fully quantitative: every decision, from signal construction to portfolio weighting, is governed by rules developed through rigorous research and proven through extensive years in the field.
Approach
We identify and exploit short-term statistical mispricing across a broad universe of instruments. Signals are constructed from statistical, macro, fundamental and alternative data, organised into thematic families, and subjected to a five-gate selection process before any capital is allocated. Portfolio construction incorporates regime-aware weighting, with a Hidden Markov Model layer used to modulate exposures dynamically.
We draw on tropical geometry, non-additive measure theory and dynamical systems —tools less common in quantitative finance than the standard factor toolkit. We try to fish in a different pond. Whether we succeed is a question markets answer daily, not us.
Universe
US, European , Asian and Latin American equities. ETFs, 23 FX USD pairs, crypto perpetual futures, and futures. Approximately 3,400 instruments monitored daily.
Signal Framework
Signals pass a five-gate selection process: in-sample Sharpe thresholds, Probabilistic Sharpe Ratio test, Deflated Sharpe Ratio for multiple-testing correction, out-of-sample validation, and portfolio-level contribution analysis. Over 200 active signal families across six research categories.
Risk Management
Controls operate at signal level (position limits, drawdown caps), portfolio level (daily loss limits, GMV controls), and regime level (HMM-triggered exposure scaling). High-watermark tracking and systematic drawdown-based de-risking throughout.