One of the major themes of War and Peace will resonate with all practitioners of stochastic finance.
Essentially, Napoleon's nemesis, the Russian general Kutuzov, keeps dropping back before the invading French until at last he spots a weakness in the French army and pounces.
Tolstoy tells us that indecision and chaos is the natural state of things, and the logical conclusion is that we should delay decisions until the time is right.
Positive autocorrelation implies momentum, so we predict the sign of the next day's return to be the same as the previous, and vice versa for mean reversion.
AutoCorrelation(500 Daily Returns) * abs(Previous Day's Return) / Previous Day's Return
'Drop the Mic' doesn't trade for decades, you can see this by the broadly straight downward sloping line of the Sharpe Trajectory (as it is a demeaned time series) from 1982 or 1988 until 2008 when it makes a killing benefitting from super negative autocorrelation and mean reversion.
Then, for all intents and purposes, closing up shop again in late 2009.
Other than that it makes all its money in the first half of the 70's.
6 odd years of profits and half a century of sitting on your hands not risking much ends up with a Sharpe of 1.6.
Like a barbell-style strategy.
The prescient questions are,
- How long until the strategy jumps back into action and takes advantage of market situations?
- Does this work across enough asset classes, geographies or sectors to keep our AUM working?
- Are these events idiosyncratic? I.e. does the mana from heaven rain or pour down at the same time?
- Can we build a rotation style strategy?
- Lastly is positive auto correlation truly a thing of the past? Are we stuck in a neutral ('momersion') or mean reversion world?
Perhaps one person's dead air is another's nugget of a rotation strategy.
In a characteristic change of heart - yesterday day I knocked strategies which don't trade and inflate statistics; now I am praising them as being the guts of profitable rotational strategies.