Interesting Stock Analysis Results

It’s ironic… I just declared my wish to focus on Internet Marketing, yet today I whipped out my excel program which analyzes historical U.S stock market prices. Anyway, this week is for me to slack and I’m kind of interested in the actual results that certain indicators will produce in the U.S stock market.

Just a bit of background info: I studied Financial Management in University and my textbook states that Efficient Market Hypothesis is well demonstrated in its weakest form, which means that historical prices should not affect how the stock market moves. However, I have never believed in my textbook, not even for a second, because there are tons of technical analysts making money. What I believe is that it’s extremely difficult to identify the current trends, because trends will tend to change based on many different factors. A chart pattern might cause massive selling ten years ago, but the same pattern might not cause any effect in the stock market today. Therefore, the tough role of a technical analyst is to constantly look out for profitable chart patterns and make the decision to ditch patterns when it’s no longer viable.

Anyway, let’s look at the results of my macro that ran certain indicators on the U.S stock market from January 2003 to July 2006. This is only the first time I’m running this test, which means that it can be much further improved, but the figures are somewhat promising for a new bird. My indicators picked up around 600 buying signals and assuming that you invest a fifth of your portfolio to buy any stock that is signalled, your portfolio would have grown three times. The test is based on selling on the 7th day closing price and assumes a bid-ask spread of 20 cents, which is pretty conservative, since I’m only running on stocks which have plenty of volume.

Note however that it’s pretty obvious that we won’t have so much cash to act on all 600 signals, so we’ll assume that we can act on 100 signals per year, which is pretty conservative, since we only invest a fifth of portfolio per signal. Therefore, it’s pretty conservative to conclude that based on these signals, your portfolio should grow by 3 times in six years, which sums up to 20% annualized returns.

There are quite a lot of assumptions in this test, but nevertheless, it still provided me with a meaningful afternoon. It’s fun to look at the figures and see my own programming skills being put to use in a different way. My afternoon is now over and it’ll probably be quite some time before I look back at Stock Analysis. My focus for this week is to slack and play games, thereafter my focus will be Internet Marketing. In the unlikely event that I get another boring afternoon like today, I might run another test, but that’s it for now.