vided you’ve done your homework correctly. Unfortunately, however, very few
systems designers (professionals and amateurs alike) know how to evaluate a sys-
tem’s drawdown, as the following little story will show you:
As an editor first for Futuresmagazine and later for Active Tradermagazine,
one of my duties was to visit many of the various trading seminars and confer-
ences, held annually around the country. During one of these conferences, I attend-
ed a seminar where a very popular and famous system designer talked about
system design and evaluation, and especially about how to limit a system’s maxi-
mum drawdown.
One of the most important aspects of system design and evaluation, he
explained, was to keep the drawdown to a minimum. To illustrate his point for the
large and sacredly quiet audience, he first showed a performance summary for a
system with a certain maximum drawdown amount, measured in dollars. Then he
went on explaining how he modified the system to lower this amount, which he
then showcased with yet another performance summary.
Now, this would have been all fine and well had he moved on to talk about
the many aspects of exactly why the second, lower drawdown was better than the
first one. But this never happened. Instead, he went on with example after exam-
ple for about two hours straight, with most systems altered with a trend filter of
some sort or a fixed dollar–based stop loss. Needless to say, at the end of his sem-
inar, the audience was awestruck by his system-design capabilities and analytical
mind. To me, it was painfully clear that not only the audience didn’t have a clue,
but worse still—neither did the speaker!Drawdown Compared to Equity
To compare two drawdown figures to each other, you simply cannot only look at
these two numbers and say that the lower one is better than the higher one—not
even for two similar systems, traded on the same market, over the same period.
Instead, you must put the numbers in relation to where and when the drawdowns
took place, in regards to both your latest equity peak and the level at which the
market traded.
For example, by altering a system ever so slightly by adding a trend filter, a
drawdown that was caused by a series of long trades when the market was in a
downtrend can be washed away completely. This means that the second, improved
drawdown must have taken place at a completely different time and place, in
regard to both the account equity going into the drawdown and the level at which
the market traded.
Say, for example, that the drawdown for the first, unimproved system was for
$60,000, at the end of a couple of years of good trading that had taken the equity
from an initial $100,000 to $300,000. In that case, the drawdown depleted the trad-
ing capital by 20 percent (60,000 / 300,000), leaving you with $240,000. For the58 PART 1 How to Evaluate a System