by Mark Harris
The stocks charted in Figure 1 look for all the world like typical charts you might find in a chartbook. I
have added a few uptrend and downtrend lines and shown how 52 (in the upper chart) seems to provide
long-term resistance, then support, and in the latter stages resistance again. I can't tell you what stocks
these are, because they aren't charts of real stocks: the charts were generated by my computer along with
thousands of others. They are shown as a family here because they all possess common statistical
Some years ago I became interested in modeling stock charts using stochastic methodsĄ that is, using
random numbers to drive the simulation.
I started off looking at random walks by letting p be today's closing price and getting tomorrow's price
from p + a x R, where R is a random number between -0.5 and +0.5 and a is a scaling factor. This
produces squiggly lines, but compared with a real plot of closing prices, the charts are neither
aesthetically pleasing nor do they produce the urge to dive in and apply technical analysis. what, in other
words, should the squiggles look like to approximate a real stock chart?
Here, naturally, we are far from scientific bedrock and getting into the realm of art. If a successful
simulation produces a chart that looks to me as if it came out of Daily Charts, it's not far from my
proclaiming, "I don't know much about art, but I know what kind of pictures I like to look at."