I design (and trade) intraday systematic strategies (for automated trading of US stocks). Each time I start work designing something new, I begin under a sort of guiding assumption that long set-ups may be treated simply as mirror images of short set-ups. Therefore, I design my strategy to handle both. And when I analyze the finished product, I consider the various metrics (e.g. # trades, % Winners, % Average Trade, Sharpe Ratio, Cumulative Return, etc) mainly on an agnostic blended (i.e. longs/shorts together) basis. Obviously, there are times when running a âblendedâ (i.e. both long and short) strategy when say intraday long trades do better than short trades (perhaps, for example, in an overall rising intraday market). Easily enough, of course, a signal might be designed for inclusion in the strategy to say switch off the short trades in these times. But there are other ways in which long does seem to differ from short, regardless of overall market direction; for example, when a market falls, it will often seem to fall faster than it climbs, fear and greed making their marks differently on longs and shorts. So, I am wondering whether I would be better off always âsplittingâ a strategy into long and short from the outset, i.e. always analyzing, optimizing, and trading, and altogether treating separately, the long trades from the short, essentially ignoring the strategy's "blended" characteristics? Wasted effort? Fundamentally flawed?