Newsletter: Choosing Binary Classification Factors

I thought that I'd do something technical but different in this month's newsletter.

As I mentioned in the hypertext book, the factors in all the market climatology models are binary classification factors. That is, they're either on or off. Equity risk, for instance, uses three implied volatility factors which signal high or low volatility. But how do I choose these factors? It goes back to two things:

• What am I trying to measure, and
• How do I know this factor does that?

So in this newsletter I'm going to describe my approach. It's actually quite simple, which is what makes it interesting.