@Epi, In your backtest, did the 1-, 3-, 6-, and 9-month momentum yield better results than the 1-, 3-, 6-, and 12-month momentum, which is more commonly used in the literature?
@Spouh In my backtests using MTD, 3, and 6, the strategy reacts more quickly to trends than Faber's 1/3/6/12 variant. And both performance and maximum drawdown benefit as a result.
@DEW_7240 Yes, the results were clear and better across various models—higher CAGR, higher Sharpe ratio. The reason? The momentum effect likely diminishes after 9 months. After 12 months, it’s barely detectable anyway. Furthermore, 9M roughly corresponds to the SMA150 used.
@Spouh I'm currently testing a weighting of 50% MTD / 25% 3M / 25% 6M for performance calculations. In backtests so far, this has yielded better results than an equally weighted calculation.
However, this does not apply to leveraged products! In those cases, the monthly swings are too extreme and create volatility.
@Epi Performance, CAGR, Sharpe ratio. But it also became clear that the selection of ETFs and their volatility, in turn, influence the results. Therefore, these factors must always be considered together. The mDD was actually influenced more by the selection of the long list than by the performance calculation.
