Many market analysts, Mutual Funds, and Hedge Funds have one "key model" that they get all their timing signals from. I spoke to a Hedge Fund Manager the other day, who said that their primary model worked in bull markets, bear markets, and trading ranges.
Does it work "just as good" in all these markets? I asked. "No, the model works better in some markets, and not as good in others", was the answer.
I did not ask how bad "not so good" was. Instead, I asked, "Forget bull and bear markets for a minute, how does it work on large cap stocks versus the NASDAQ 100, and mid cap stocks?" "I don't know, why are you bothering to ask that?", he replied. The reason why, was because different models produce different results for stock groups that are different in terms of their avarage capitilazion size.
Why was "my question" extremely important to You as an Investor?
I recently did a very important correlation study. I took 1 key variable in one of our Market Models and back tested the profit and lost results from January 3rd. 2006 to August 22nd., 2007. Specifically, I tested the key variable against the "Top 75 core holdings" that Institutions owned, the NASDAQ 100, and the S&P 600 Mid Caps.
Thus, the study covered large Caps and S&P Mid Caps. The NASDAQ 100 sits in between the two groups. The three charts below show the performance differences one model can have over differences in capitalization.
Let's look at each previously mentioned indexes to prove the point:
Chart 1: Our Mountain Model against the Top 75 Stocks owned by Institutions. (Remember that we only used one key element in out Big Cap Model. Pay particular attention to the % of stocks that made a profit and the % of stocks that had a loss. ... here are the results:
The results were not bad. 84% made a profit and the worse loss was 20% of the capital invested in that particular stock. (I assumed $100,000 was invested in each stock for the test.) Now, you might be saying that this was because they were the Top 75 Institutional stocks ... but that is not the case, because we ran the same test on the DJI 30 stocks and the results were within 1% of this test.
____________________________________________________________________________________
Chart 2: Our Mountain Model against the NASDAQ 100. Again, pay particular attention to the % of stocks that made a profit and the % of stocks that had a loss. ... here are the results:
The results were not as good. Only 67.3% made a profit and the worse loss was close to double the loss on the large caps .... coming in at 40.75% of the capital invested for that particular stock.
____________________________________________________________________________________
Chart 3: Our Mountain Model against the S&P 600. Again, pay particular attention to the % of stocks that made a profit and the % of stocks that had a loss. ... here are the results:
The results were worse again. Only 58.4% made a profit and the worse loss was more than double the loss on the large caps .... coming in at 49.18% of the capital invested for that particular stock.
____________________________________________________________________________________
Conclusion: Investing models have huge differences in Results when they are applied to bull markets, bear markets, trading ranges AND the capitalization size of a stock. Not only did the % of wins drop from 84% to 58.4% on the same model between large caps and mid caps, but just as important, the largest loss on a single stock on the mid caps was 238% larger than the worse loss on the Institutional large caps.
The point is, there is not "one" magical formula that works across all conditions, and there is a critical need for analysts to run their models against multiple conditions and criteria.
