The gain is astonishing, +123% over six months for such a mixed value-growth group of stocks as the Classic Portfolio.
Let us examine what happened. The 40-weeks results which appear in the legend show the gains or losses over the full period for each tactical method in the legend box compared with the SP500.
The chart depicts the outcome of a very simple system.
1- When the trend is up, buy the top-ranked stocks. 2- When the trend is down, sell short the bottom-ranked stocks.
(The trend indicator appears every week in each of the 100 Stocks--3 Portfolios' tables. See the top row of the Classic Portfolio table here).
All four curves start at 100 on September 5, 2008. Twenty-six weeks later, the buy & sell short results peak at 223, more than double the invested funds in six months. Impressive. But would anybody have gotten them, including me?--I doubt it--because the system is asymmetrical. It uses two top-ranked stocks when the trend is up and seven bottom-ranked stocks when the trend is down. Who would have guessed? Why those two numbers?
Therein lies the beauty and danger of data mining. It gives you the absolute best results for a set of data. It also can mislead you into the impression, if not the belief, to expect that the next set will be just as good or better. The next set will not. (The remedy is to run the test with 400 or 500 weeks instead of 40. That would improve reliability and credibility.)
Why 40 weeks in this test? My tools that I used for this analysis are limited to that capacity for the time being.
The 223% peak value was an unexpected byproduct of the test. It is attributable solely to the discovered combination of using top 2 and bottom 7 ranks for buying long and selling short. I will show other (symmetrical) combinations in a minute.
The big value rise (no leverage used) seems to me unusual for the time frame--499% per year annualized over six months, September 5, 2008 to February 27, 2009. These are 'Classic', 'blue chip' type stocks. The value spike is the inverse counterpart to the SP500 value which collapsed -41% during the same interval.
Ten weeks into the trades, by November 14, 2008, profits were above 60%. Would it not be outstandingly prudent to opt out of the program for a while--at least until the next durable buy signal? My advice to myself is, You bet it would! By this simple voluntary pre-exit against the 'rules', my total buy & sell short results clock out at 85% instead of the 30%* reported in the legend box above. Bye bye, Greed. Hello Happiness, the reward of flexibility.
Maybe a good sub-rule for any system on this website, or anywhere else, might be: whenever any cumulative portfolio profit exceeds 60%, take it at once and wait for the next buy signal.
*40.3% per year annualized
Next page: Reliability and Gains--here
|
Back
| Subscribe
|
