Way of the turtle

54
•
Way of the Turtle

---

But why make it so complicated to forecast probability? One
could ignore the math and formulas and still construct a graph like
the one shown in Figure 4-1 by using this simple method: First, go
to a place where you will find a lot of women, such as a college
campus. Next, find 100 women at random and measure their
height. Finally, divide those heights into 1-inch intervals and count
the number of women in each interval. You are fairly likely to get
around 16 women at 64 inches, about 15 at 63 and 65 inches,
about 12 at 62 and 66 inches, 8 at 61 and 67 inches, 4 at 60 and
68 inches, 2 at 59 and 69 inches, and one each at 58 and 70 inches.
If you created a bar chart showing the number of women at each
particular height, it would look like the chart shown in Figure 4-2.
The type of graph shown in Figure 4-2 is called a histogram. It
graphically shows the frequency of a particular measure com-
pared with other nearby measures (in this case the measure of a
woman’s height). The graph in Figure 4-2 has the same shape as
Think Like a Turtle 
•
55
Height Histogram
10%
20%
40%
60%
80%
100%
54
53
55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74
0
2
4
6
8
10
12
14
16
18
Number of W
omen
Cumulative Percentage
Figure 4-2
A Histogram of Women’s Heights
Copyright 2006 Trading Blox, LLC. All rights reserved worldwide.

---

the normal distribution graph in Figure 4-1, but it also has the
advantage of being something you can construct without using
complex mathematical formulas. You only need to be able to
count and categorize.
A histogram like this can be constructed from your trading sys-
tems to give you an idea of how the future might turn out; it pro-
vides you with a way to think in terms of probabilities rather than
prediction. Figure 4-3 is a histogram of monthly returns from a 20-
year test of a simplified version of the Turtle system, the Donchian
Trend system. In addition to being simpler, it has a better per-
formance record than the Turtle system.
The histogram sections in Figure 4-3 are divided into 2 percent
segments. One bar lists the number of months with between 0 per-
cent and 2 percent positive returns, the next bar lists the number
between 2 percent and 4 percent, and so on. Note how the shape
of the histogram resembles the normal distribution of heights
described above. The notable difference is that the shape is elon-
gated toward the right. This elongation represents the good months
and sometimes is referred to as skew and fat tails.
The histogram shown in Figure 4-4 represents the distribution
of the trades themselves. Figure 4-4 shows how individual trades
are distributed. The section on the left is for losing trades, and the
section on the right is for winning trades. Note that the scales for
each section include both a number scale on the outside left and
right and a percentage scale in the middle from 0 percent to 100
percent. The cumulative lines move from 0 percent to 100 percent
from the center of the graph outward. 
The numeric legends on the left and right indicate the number
of trades represented by each 20 percent section of the graph. For