Sequential Asset Ranking within Nonstationary Time Series Gabriel Borrageiro*, Nick Firoozye and Paolo Barucca Department of Computer Science, University College London, 66-72 Gower Street, London, WC1E 6BT, London, United Kingdom. *Corresponding author. E-mail: gabriel.borrageiro.20@ucl.ac.uk; Contributing authors: n.firoozye@ucl.ac.uk; p.barucca@ucl.ac.uk; Abstract Financial time series are both autocorrelated and nonstationary, pre- senting modelling challenges that violate the independent and iden- tically distributed random variables assumption of most regression and classification models. The prediction with expert advice frame- work makes no assumptions on the data-generating mechanism yet generates predictions that work well for all sequences, with perfor- mance nearly as good as the best expert with hindsight. We conduct research using S&P 250 daily sampled data, extending the academic research into cross-sectional momentum trading strategies. We intro- duce a novel ranking algorithm from the prediction with expert advice framework, the naive Bayes asset ranker, to select subsets of assets to hold in either long-only or long/short portfolios. Our algorithm generates the best total returns and risk-adjusted returns, net of transaction costs, outperforming the long-only holding of the S&P 250 with hindsight. Furthermore, our ranking algorithm outperforms a proxy for the regress-then-rank cross-sectional momentum trader, a sequentially fitted curds and whey multivariate regression procedure. Keywords: financial time series, online learning, prediction with expert advice, learning to rank, cross-sectional systematic strategies 1 arXiv:2202.12186v1 [cs.CE] 24 Feb 2022