First, I have to point out that this post references one of my all-time heroes (you can find more information on the Investor Superhero page under Menu), Cliff Asness, whom I have an unhealthy propensity to scour the web for any tidbit of info he posts/shares.
The article does a wonderful job at showcasing statistically insignificant short time intervals in proper context vis-à-vis a larger sample. A few thoughts:
- If risk and return are in fact related then seeking out recent back tested strategies that have faired extremely well in one particular market could very well play out to produce a lower expected return in the future (Meaning you get out of sample returns with lower Sharpe ratios than the benchmark used for comparison). Not a bad idea to test things, but simply looking at history and not applying an economic or behavioral rational to the outcome (Combo of the two likely better) has a high probability of being a fools errand.
- If you stop and think about it, at a very basic level this makes since…Risk and Reward must be interrelated at the large sample size level, and yet doesn’t appear as such when viewed from a narrow lens. Think about all the silly things you did once, twice or possible three times as a youth, and yet it didn’t materially harm you. Take that same event and imagine doing it 1000 or 10000 times…Probably pretty hard now to imagine that you didn’t experience at least a few really bad outcomes, right? You might be able to catch a falling knife once without a scratch, but chances are you’ll need a box of Band Aids if you try it 100 times.