Galileo Galilei was deemed heretical but was later vindicated when the world caught up to his brilliance. This post might make me a quant-heretic without the brilliance part.
you can think of a tilt with an example, albeit abit forced. suppose you know the weekend theta decay is not priced by options, but trading this inefficiency is not compensated by costs enough. But you already have an open position, which you intend to close on Friday. Instead, you close on Monday, harvesting the premia and not incurring further costs.
trend on alpha signals are not always recommended, particularly if the signal itself is mean-reverting. You can learn about it in literature for trend on trend (e.g. Alex Greyserman managed futures). This is intuitive. Suppose trend following is compensation for taking on tail risk. By trend-on-trending, you take on less tail risk > your compensation is lower. The same effect is observed when you shorten the trend following period, since this lowers your tail risk even further.
Just an idea on your Strategy Filtering. I use CSI data with over 40 years of continous futures series and I havent found a single significant filter for any of my strategies. What I use for the "regime" differences is just having constant $ risk sizing based on ATR, that means I risk the same amount of money but based on the current volatility of the futures, stock or what ever.
Thanks for the comment. I already have vol scaling as part of portfolio construction flow, I'm just testing out filters to see if they are merits of including them. Appreciate your input :)
great job :)
you can think of a tilt with an example, albeit abit forced. suppose you know the weekend theta decay is not priced by options, but trading this inefficiency is not compensated by costs enough. But you already have an open position, which you intend to close on Friday. Instead, you close on Monday, harvesting the premia and not incurring further costs.
trend on alpha signals are not always recommended, particularly if the signal itself is mean-reverting. You can learn about it in literature for trend on trend (e.g. Alex Greyserman managed futures). This is intuitive. Suppose trend following is compensation for taking on tail risk. By trend-on-trending, you take on less tail risk > your compensation is lower. The same effect is observed when you shorten the trend following period, since this lowers your tail risk even further.
Just an idea on your Strategy Filtering. I use CSI data with over 40 years of continous futures series and I havent found a single significant filter for any of my strategies. What I use for the "regime" differences is just having constant $ risk sizing based on ATR, that means I risk the same amount of money but based on the current volatility of the futures, stock or what ever.
Thanks for the comment. I already have vol scaling as part of portfolio construction flow, I'm just testing out filters to see if they are merits of including them. Appreciate your input :)