The Unlikely is Not Impossible

And sometimes it's even likely. This is a paradox I often find myself wrestling with as a trader, particularly since I like to think I approach the forex market with a statistical mindset. Why do you think I go on about Bollinger Bands so much? They're one of the more statistical indicators out there, based on a measurement of 2 standard deviations from a specified moving average. But if you trade long enough, you discover that your statistical ideas about market behavior often run headlong into the painful realities of how the market actually behaves.

What set me off on this particular tangent was a recent article in The Economist. (Seems I reference them in every other post. Such is intellectual laziness.) It was commenting on the recent volatility in world financial markets, seen in the meltdown of the Chinese stock market, the serious decline in US markets, and similar turbulence in all the other markets I don't really follow. What really got my attention were these remarks about the statistical likeliness (or rather, unlikeliness) of all this marketplace drama:

"According to Goldman Sachs, the latest jump in the Vix (a measure of stockmarket volatility) took it eight standard deviations from its average. If conventional models are correct, such an event should not have happened in the history of the known universe. Then again, the move in energy prices that caused the collapse last year of Amaranth, the hedge fund, was a nine standard-deviation event. [I wrote about the Amaranth collapse a while back, follow that link to learn more about how they screwed up.] Perhaps modellers do not know the universe as well as they would like to think."

Eight standard deviations. Nine standard deviations. These make the 2 deviations of my Bollinger Bands look painfully inadequate to accurately gauge possible market fluctuations. The key of course is that these types of events are extremely unlikely. Just like the global financial implosion triggered by Russia's loan defaults and the devaluation of the Thai baht back in 1997. Extremely unlikely events, and yet they brought Long-Term Capital Management crashing down. [Wrote about them too. Follow the link for more hedge fund meltdown fun. So, got any assets in hedge funds?]

Extremely unlikely, and yet they happened. In fact, if you look at historiy it seems inevitable that extremely unlikely events like these will happen again. So what does that make them - likely? Depending on your timeframe, yes and no (century: likely, month: unlikely). And they're definitely severe enough to include in whatever rosy model you have for trading and investment success. If you live in an earthquake zone (which I do) you don't expect one every day. But you do prepare yourself for the day when the floor starts moving. So how strong is your trading floor?

Related article:
The Economist: Grey Tuesday - An overdue sell-off flusters exchanges and sobers investors