The fundamental distinction in SPC is between random variation and special-cause variation. Random variation is
usually defined as variation between the control limits. Special-cause variation usually falls outside these limits. However, there are other considerations. Runs above and below the mean can
indicate unusual conditions. Several points in a given zone of the control chart can also suggest something is wrong. After you've looked at control charts for years, you get an instinctive feel
for unusual conditions. This intuition is even stronger if you have really studied and understand the process. Some years ago, I left a poorly managed company and began working
for a well-managed company. Among the details I had to deal with as a result of the change was my Individual Retirement Account (IRA). As any of you who read this column know, I tend to be a
little simple-minded about things. As such, I decided to sell all my stock in the poorly managed company to buy stock in the well-managed company. This decision brought howls of protest from my
financial advisor. She wanted me to diversify and ride the wave of the soaring Dow. She wanted me to grow up. Well, I didn't want to grow up. If I had, I never would have
become an engineer. I certainly didn't want to become a stockbroker. In my mind, stocks were something you bought as part of the company retirement plan and worried about in a few decades.
Besides, they didn't seem like real money, the kind you put in a Coke machine. I had also read somewhere that the best strategy for stocks was to buy and hold. This appealed to my simple-minded
approach, so I put all of my money in the stock of this well-managed company. I found it hard not to watch the stock price. Even though I was just supposed to buy and hold, I
still checked it every few days. For a while it went up. Then it went down. The longer I stared at the graph of the stock price, the more it reminded me of a control chart. I had always believed
that control charts were of no use in the stock market. A friend of mine once plotted General Motors' stock on a control chart and found that it was operating in a state of statistical control
(i.e., it wasn't changing in any statistically significant way). We both believed that there were many small influences on the price of the General Motors' stock and that it was showing only
random variation. Each day, as I listened to the news, I felt a superior contempt for the news people who not only reported small rises and falls in the market but also gave what I believed were
superstitious reasons for these fluctuations. I continued to track my stock for a year or so; then I thought I recognized a pattern in its price fluctuation: It had a tendency
to overcorrect. When it went up, it seemed to go up too high and then fall back. When it went down, it tended to go down too far and then come back up. The line graph of the stock price didn't
form a line with random variation in it; it formed a series of peaks and troughs that reflected the overcorrections. This all remained a vague intellectual exercise. After
all, I was the guy who was going to buy and hold my stock, right? But then a strange thing happened. The price of the stock shot up dramatically; it increased about 50 percent in a few days.
Every day I told myself I didn't know much about the stock market and should leave it alone. But a little voice in the back of my head kept saying, "This is statistics."
Because I was beginning to understand the underlying process, I was able to move beyond statistics. With the very strong feeling that I still didn't really understand the stock market, I
sold my stock. I was rewarded in a few days when the price dropped back close to where it had been. Since then I have bought and sold the stock a few times, when the price seemed very unusual to
me. As a result, the 1,000 shares I started with have now grown to 1,900. Of course the value of the stock fluctuates, so I haven't made a 90-percent profit. But if I had done nothing, I would
only have 1,000 shares instead of 1,900. Because I'm planning to keep this stock for about 20 more years, the daily fluctuations in price are of little concern to me. And because this stock is in
an IRA, I don't have to pay capital gains for a long time. It would be nice if I could say that I plotted the price on a control chart and it told me what to do, but the
process was more subtle than that. After many years of working in process control, I have come to believe that simply studying the process is far more important than any of the statistics that
might get used. I call this "becoming one with the process." I may not be "one with the stock market," but I can still recognize unusual events and react to them. This may all
look like child's play to a stock market analyst; I wouldn't know. And before all is said and done, I may lose my shirt. But somehow I doubt it. A good understanding of variation might just be
worth its weight in gold. About the author Gregory P. Ferguson is quality manager
of Parker Hannifin's Tucson, Arizona, facility. He has published technical articles and assisted in the publication of two books. Comments can be e-mailed to him at gferguson@qualitydigest.com . |
