two signals went short only one day apart from each other. In late October and
early November, two signals two days in a row went long. In the case of the short
signals, clearly the second signal was better than the first signal. In the case of the
long signals, the first signal was better. However, the point is that in either case,
the second signal most likely would have gone unnoticed and unanalyzed, because
you already would have been in a trade generated by the first signal.
Now, pretend that a few of the signals that came first in a row of two or more
signals really didn’t happen, or that you for one reason or another failed to act on
them. Then you would have had to rely on the second signal. Transfer this reason-
ing to a real-life trading situation. How many days have you had when you saw a
nice setup forming and readied yourself to place the trade if the stock in question
only reached a certain level? Have they all resulted in a trade? Probably not. The
next day, the stock takes off in the anticipated direction, leaving you behind until
late in the day, when you actually get the signal you were looking for yesterday.
With the market already having made a good move in the anticipated direction, do44 PART 1 How to Evaluate a System
FIGURE 3.7
Microsoft performance signals—Fall 2001.