Market-Neutral Portfolio of Trading Strategies as Universal Indicator of Market Micro-Regimes: From Rare-Event Forecasting to Single-Example Learning of Emerging Patterns

摘要

Discovery of stable statistical arbitrage opportunities becomes more challenging due to increasing number of intelligent market participants and market-related technological advances. Existence of practical econometric-type forecasting models leading to profitable trading strategies is questionable. Technical trading strategies, directly optimized to achieve desirable return/risk objectives, have more practical value but still cannot warranty stability across different market regimes or timely regime switching. Recently proposed boosting-based optimization framework can discover portfolios of multi-scale trading strategies for a particular instrument(s) with stable (non-resonant) performance over wide range of market regimes. Here it is argued that such market- neutral portfolios can have much wider and more generic scope of applications when used as universal indicators of market micro-regimes. Among the most interesting applications of these "strategy-based" indicators could be such challenging problems as rare-event forecasting and single-example learning of emerging patterns leading to trading strategies exploiting these difficult-to-model regimes.