Generalized statistical arbitrage concepts and related gain strategies

The notion of statistical arbitrage introduced in Bondarenko (2003) is generalized to -arbitrage corresponding trading strategies which yield positive gains on average a class scenarios described by -algebra . This contains classical as special case. Admitting general static payoffs strategies, done Kassberger and Liebmann (2017) the case one pricing measure, leads -arbitrage. We show that even under standard no-arbitrage there may exist gain yielding specified scenarios. In first part paper we prove characterization (2003), no being equivalent existence an local martingale measure with path-independent density, not correct general. establish this equivalence holds true complete markets derive sufficient condition for -arbitrages. As main result second construct several classes profitable respect various choices particular, consider forms embedded binomial follow-the-trend well partition-type strategies. study compare their behavior simulated data also evaluate performance market data.