Watch out for sector rotation and selection bias

Strategies might be attractive on the surface but fail moving forward. Why? Here we discuss the qualities of effective trading strategies, and the many traps one should avoid.

Watch out for sector rotation and selection bias

Postby Overload » Wed Feb 22, 2006 2:59 pm

Here’s an issue about which I don’t have a clear answer.

Let’s say your trading system is based on the Nasdaq 100 Index stocks. And let’s say that the Nasdaq 100 rotates its stocks as much as 10% per year. Since the Nasdaq 100 is made up of the 100 largest non-financial stocks in the Nasdaq, the types of companies included can change significantly over the years. For example, the Nasdaq 100 might be heavily influenced by Internet stocks one year, but then be taken over by Biotech companies a couple years later.

The big unanswered question is whether you should continually update your Nasdaq 100 sector list as you move forward, or whether your Nasdaq 100 component list should remain exactly as you back tested it. There are pros and cons either way.

If you continually update your sector listing with the new rotation, you must be careful that your trading system doesn’t become obsolete in the process. Since a trading system is often built according to the unique characteristics of the sector, a rotation from Internet Stocks to Biotech Stocks may create uncharacteristic signals and therefore throw off your performance moving forward.

On the other hand, if you don’t update your Nasdaq 100 sector list, you’ll be introducing a new variable into your trading system: Stocks that “were” a part of the Nasdaq 100 that might no longer be. I won’t even guess what the impact of such a variable might have on a trading system.

In addition to the above, there’s also what might be called “selection bias” within the back test of a sector like the Nasdaq 100. When you run your back test against the Nasdaq 100, you’re running against stocks that are members of the Nasdaq 100 as of today. But they likely weren’t at the beginning of your Evaluation Period. Thus, there’s a “selection bias” for these stocks, in that they’re companies that were destined to become one of the 100 largest non-financial stocks in the Nasdaq 100. Such foresight likely isn’t possible moving forward.

It’s a personal decision whether or not to use sectors that have these issues. But either way, it’s helpful understanding how these issues can affect your trading systems moving forward.

Posts: 2248
Joined: Wed Nov 30, 2005 12:14 pm

Return to Curve-Fitting and Other Pitfalls