Static models are used frequently in medicine. For example, the methods that apply to diagnosing the flu do not change from week to week or month to month.

In cardiology, radiology, and psychology, static models are used to make diagnoses and recommend treatments for patients with all types of symptoms. In The Naked Portfolio Manager, I gave examples in each of these disciplines where statistical prediction models proved more reliable than the judgment-based methods of clinicians in the field. For example, the Goldman Chest Pain Decision Aide model proved more reliable at making better triage decisions than the emergency room doctors and cardiologists at Cooke County Hospital in Chicago.

Although some traditional portfolio managers argue that decision-making, when it comes to the stock market, is different. People with chest pains do not change their symptoms based on what other people with chest pains are doing to gain an advantage. But investors, at least to some degree, change their behavior based on what other investors are doing. For this reason, many traditional judgment-based portfolio managers reject "naked strategies" because they say a rule-based method is inflexible and thus cannot change based on market conditions. This is why the best portfolio managers that rely on judgment, they argue, are able to outperform naked managers.

Phooey!

Naked strategies don't have to be static. They can be flexible and adjust to market conditions. The rules for adjusting to market conditions need to be created when the strategy is designed. The Cripps Model, discussed in The Naked Portfolio Manager, is a perfect example of this. This model measures how investors react to news on a stock and then applies those measurements to calculate a fair price. Stocks are purchased when the market price is substantially below this calculated fair price.

Note here that the Cripps Model calculates the value of a stock based on the collective judgments of all market participants and then gives a buy signal when the stock price deviates substantially from that price. This is much different than an analyst forming his own opinion about what a stock should be worth.

Labels: Cripps Model, Goldman chest pain decision aide, judgement