I don't understand what you're saying.Marseille07 wrote: ↑Wed Dec 01, 2021 9:47 pmBecause it's not "better" in a meaningful way.
If you roll a die 1000 times, some faces turn up more than others. Do you start arguing that "5" backtested better? Of course not.
This is why understanding why something might look better is important.
They're not rolling a die. The backtest isn't some random probability generator.
It's applying the model to the data. EoM looks better because it has been better over that time period.
Quite frankly, the backtest shows me that using the 200-day MA isn't "better" in a meaningful way.
anoop - If you want a smaller max drawdown and lower overall return outside of a convergence point, just have a more conservative AA and be done with it. If 25% is too much for you, go to 20% or 15%. Whatever. Pick your lowest non-0 comfort point (where you know you can just not touch it at all, no matter what happens) and go with that. But let it ride. Don't rebalance. Not in the next crash, and not in the next bull. I'm guessing you might need the next bull. Better let it happen.