This book investigates online portfolio selection (OLPS), which aims to efficiently allocate capital among a set of assets. The principles of OLPS mainly include follow the winner, follow the loser, and pattern matching based algorithms. Based on various existing algorithms, the authors designed four algorithms to advance the state of the art, which led to some very powerful tools for portfolio selection. Simulations using real market data are used and show that these strategies are effective and provide strong confidence for potential deployment.