First time I'm reading one of the performance review articles and again, I am inspired. I'll definitely be trying to replicate this portfolio tracking on my own over the next couple of months! Also, I'm definitely looking forward to the TSX and Long Volatility Mini-Portfolio of systems articles when they come out as I need a lot of diversification!
Interesting approach, using a rolling Sharpe Z-score as an allocation guardrail is intuitive. My only caution is that a negative Z-score identifies below-normal recent risk-adjusted performance, but does not by itself imply higher expected returns. Have you tested subsequent returns conditional on the Z-score in a strict walk-forward/out-of-sample framework and across different lookback windows? If the mean-reversion relationship remains stable, this could be a compelling allocation overlay. BTW congrats on your returns!!!
Yes it’s just one metric to use in context of everything else. Tho it does look to be a decent predictor based on live results. I have not tested historical tho I should.
First time I'm reading one of the performance review articles and again, I am inspired. I'll definitely be trying to replicate this portfolio tracking on my own over the next couple of months! Also, I'm definitely looking forward to the TSX and Long Volatility Mini-Portfolio of systems articles when they come out as I need a lot of diversification!
Amazing man let me know how it goes!
Interesting approach, using a rolling Sharpe Z-score as an allocation guardrail is intuitive. My only caution is that a negative Z-score identifies below-normal recent risk-adjusted performance, but does not by itself imply higher expected returns. Have you tested subsequent returns conditional on the Z-score in a strict walk-forward/out-of-sample framework and across different lookback windows? If the mean-reversion relationship remains stable, this could be a compelling allocation overlay. BTW congrats on your returns!!!
Yes it’s just one metric to use in context of everything else. Tho it does look to be a decent predictor based on live results. I have not tested historical tho I should.
Maybe it could be an idea for your next post!! Thanks for sharing your thought process and progress, very inspiring, as always!