I’m getting confusing results from some tests while working on a long/short strategy.
I combined a long strategy with an identical short strategy (albeit with margin) in a Simulated Book, which should yield a 0% annualized return. Using no leverage and allocating 50% of the capital in the Book Simulation to the long strategy and 50% to the short strategy, the Book Simulation had a drawdown of 53% (Book Simulation) with daily rebalancing. The distortion is more pronounced during periods of high volatility, but it is present during periods of low volatility as well. As it stands, the discrepancy is significant enough that I’m not certain which reflects the model’s true performance.
Using random tickers, I also tested a few workarounds discussed on the forum for replicating long-short strategies in Book Simulations. The first (mentioned by Hemmerling here) involves leveraging up the long and the short Simulated Strategies to 2X and combining them in a Simulated Book at 50% weights. The second (mentioned here) calls for assigning a negative weight to a placeholder portfolio in the Simulated Book to combine the long and the short Simulated Strategies at 100% weights. I tested these workarounds on a two-ticker long-short strategy with differing results (First Workaround, Second Workaround). Hemmerling’s method makes sense to me, but most short strategies throw an error at 2X leverage without a long strategy to hedge. I get the feeling that there is something amiss with the second method – has anyone successfully used the second workaround, or another workaround, to replicate a long/short strategy in a Simulated Book?
Any insights into either of these issues would be appreciated.
Will