I have just read this article in today's FT and wonder how does this approach differ than simply going long the VIX? https://www.ft.com/content/2e28fb42-de8a-40d7-8716-88cb5febe006 Is all this computational power and army of PhDs necessary really?
Blimey. Whether you need all that power or an army of phds is a moot question, but there is an awful lot more to this than just being 'long the VIX' Rather than trading one future, you're probably trading over a hundred. Also you're going long and short. And if anything you spend more time being short the VIX specifically than long it. Here's a graph of being long VIX Here's my own trading strategy live returns (it's basically a CTA style but with zero phds) Here's the HFRI index referred to in the article Spot the difference? The correlation between the VIX and the other two graphs is close to zero. GAT
YTD, they are all up. But what does that prove? The correlation of daily returns for the YTD period (Long VIX, my CTA) is 0.07. Basically noise. Correlation of monthly returns is -0.29 FWIW (only 5 observations) GAT