Online Stock Trading

This is a continuation of this article. In that article, which was simply a beginning examination of what stock price, exclusively, can seem to indicate about price performance, we determined that stock price alone could seem to be used (for the time frame I selected to examine) to achieve a higher yield; higher when compared to the market overall, but only neck and neck with the performance of the major indices. As promised at the time, I’m now coming back to this to assess some of the same things, intra-index. In other words, let’s narrow the scope. Also, a reader, bzak, in response to the last article, suggested some ways we could test some inferences that we need to make before we can use results like these. His implied question was, “What’s the cause of the effect we see?” In other words, if higher priced stocks perform better, what causes that? He offered some excellent suggestions. In a following article, I will begin discussing these too, and will look at ways we can identify and capture those suggestions.

So let’s get started

In the previous article, we discovered that stocks over about $25 on average performed better than all stocks, but only mixed when compared to the major indices. This time, we’ll take a similar look at the same measurement within those indices to see whether even more can be said.

We’ll start by looking at the Dow Jones Industrial Average. This index and a graph are never a good marriage. Too few stocks. In this case, trying to assess any correlation between price and yield is difficult because they all have generally high stock prices. The graph below shows the results using a 5-stock, backward oriented smooth. After an initial spike (caused by phenomenal gains from General Motors and Alcoa) the stocks performed within a tight range between about 6 and 8%. Not much to ascertain. You could have picked any of them, only 3 were flat (Intel, Home Depot and IBM), and none fell in value.

The NASDAQ is difficult in a graph because it was highly volatile at almost all price ranges. Ignoring the initial poorest performance of the lowest priced stocks, all price ranges ranged between about 4 and 8%. This graph uses a 200-stock, backward oriented smooth. One interesting thing to note is this: the very cheapest stocks (on the extreme left limit) performed clearly worse than the whole, and the most expensive stocks (on the extreme right limit) performed clearly better than the whole. In fact, once you pass about $35, following from which the previous 200 stocks yielded on average over 8%, the absolute worst you could do on average was over 7% and the best over 9%. This would seem to indicate that indeed, of the stocks over $25 that we looked at from the first article, those in the NASDAQ that were even more expensive than the $25, did best of all. In other words, initially, we could say that stocks over $25 did best, but not better than the indices, but now we can say of those over $25 in total, those that were over $35 and in the NASDAQ did better than the stock market as a whole and better than any index. Now we have a useful subset. In fact, the average yield of the this subset was 8.35%. And I don’t even know the companies’ names.

Now let’s look at the S & P 500. She was the most beautiful beast ever beheld by gridlines on this day. This graph is a 100-stock, backward oriented smooth. Something more clearly interesting happens with her. While the stocks on aggregate, do trend up overall, there is a clear indication that the very cheapest stocks did significantly worse than the most expensive. In fact, once you’re over $58, there’s no holds barred. In fact, just between $58 and $63, there’s a full 1.5% hike, followed by more moderate gains all the way to a spike at the end (smell that? I think Google might be around). After that $63 threshold, you would have, as with the NASDAQ’s sweet spot, both soundly beaten the market as a whole and the major indices.

Here are some next things I’m going to consider. Because we all know that none of us could easily buy “this 200 stock section” or that one of an index, and thus any smaller subset would make literal my repetition of the phrase “on average” (i.e., a gamble), I’m going to begin some density functions (or probability distributions) using some of these subsets. For investors like us who don’t like to gamble, this should certainly be some fun gaming.