This paper provides discussion on factor-based modelling for hedge fund returns, and demonstrates replication via both rolling windows and Kalman filters. In particular, we focus on estimating time-varying hedge fund returns exposure through various asset-based style (ABS) factors. It is shown that certain hedge fund strategies are more susceptible to cloning, suggesting a higher likelihood of creating transparent liquid replication products as either an alternative investment vehicle or as a benchmarking / style analysis tool for institutional investors