Hi, I want to do a backtest and compare it against buy and hold. My question is what should I use for my order sizing function if I want the strategies to be comparable? Should it be: 1) An absolute share amount 2) A percent of equity 3) A fixed dollar amount For example here is a simple strategy on the UVXY. I would like the buy and hold strategy to be - Sell UVXY using 100% of net liq at the beginning of the backtest. This equates to shorting 5 shares. In this scenario, UVXY prices are adjusted for splits. Therefore we will realize all our PnL volatility at the beginning of the backtest due to the large price decline. This first picture shows our strategy (black- invest a fixed dollar on every signal) vs the buy and hold (red-short 5 shares at beginning of backtest ). As you can see these are not very comparable. On the fixed share basis we get comparable outcomes, but it is not realistic. In the beginning, we are trading $1million on every signal and at the end $100. The Percent of equity looked very similar to the fixed dollar. So far it seems, the only way to compare these strategies is through a table (compare trading statistics) rather than actually plot them. Any help is much appreciated. Thank you. Ps. Would a better buy and hold strategy be - rebalance every month to keep the position at our original net liq? PPs. I can add source code if anyone is interrested
Unless there is some sort of portfolio-formation impact from having a fixed dollar amount, why change from that popular standard?
How about randomizing the timing of your trades? For example, suppose you have three trades lasting 2 days, 7 days and 5 days over a one month period. Keep those durations, but randomize the entry times (and keep the dollar amount invested fixed rather than the number of shares).