Now, pretend the columns in Figures 2.1 to 2.3 represent percentage
moves (never mind the scale on the y-axis) on a one-share (one-contract) basis.
Which system version would you go with, assuming you know nothing else
about the systems? Again, your answer should have been System 3 (Figure
2.3). Again, it’s because it’s the only system that works equally as well all the
time, no matter the trend or the quality of the different market conditions.
Because the outcomes from all the trades are very similar, it’s the only system
that gives all the trades the same weighting in deciding the parameter settings
of the system.
The other systems will give more weight to the earliest trades (Figure 2.1) or
the latest trades (Figure 2.2). This is especially dangerous in Figure 2.2, because if
market conditions change back to what they were at the beginning of the test peri-
od, the system will start to produce significantly less profitable trades, and you
won’t know why until it’s too late. A system like that represented by Figure 2.1
might, on the other hand, surprise you positively by starting to produce trades sig-
nificantly more profitable than what you expected. But who in his right mind
wants to trade a lousy system only on the assumption that it is more likely to sur-
prise you in a positive way than a negative way? What if it doesn’t surprise you at
all and just continues to be plain lousy?18 PART 1 How to Evaluate a System
FIGURE 2.2
System 2 distribution of trades and profits.