provided you go long or short the same number of shares throughout both trading
sequences, a high drawdown still can be better than a low drawdown, if the mar-
ket also is sufficiently higher in the case of the high drawdown compared to where
it is during the low drawdown.
This is easier to understand if you isolate your drawdown to one single trade.
For example, if a stock is priced at $60, and you buy 1,000 shares and lose $5 per
share, you have lost a total of $5,000 (1,000 * 5), or 8.3 percent (60 * 1,000 /
5,000) of your investment. If, on the other hand, the stock is priced at 30, and you
buy 1,000 shares and lose $3 per share, you have lost a total of $3,000 (1,000 * 3),
or 10 percent (30 * 1,000 / 3,000) of your investment. Thus, even though the high-
er priced market (and its inherent larger price swings) forced you to invest and lose
more dollars, the fewer dollars lost in the lower-priced market resulted in a rela-
tively larger loss of equity.
Thus, if you apply a trading system that buys or sells a constant amount of
shares on a trending market, the value of the losing (and winning) trades should
be sufficiently lower in a low-priced market, compared to a high-priced market.
That is, the wider the price swings, measured in dollars, the larger the drawdown,
also measured in dollars, should be. If that is not the case, there is something
wrong with the logic behind the system.
One such logical blunder people make all the time is to apply a fixed
dollar–based stop that will have you stopped out after a trade has moved against
you a specified amount. However, the effect of such a stop is that it will have you
stopped out repeatedly on a high-priced market, but hardly ever on a low-priced
market: It will be too tight for the high-priced market, but too wide for the low-
priced market, in relation to the respective markets’ normal price swings.
Furthermore, this system-design blunder can be very hard to detect via the
drawdown because, in the case of the low-priced market, a larger than necessary
drawdown can be a result of a few but larger than necessary losers, while in the
case of the high-priced market, a larger than necessary drawdown can be a result
of a large amount of unnecessarily small losers.Mean–Median Comparisons
Another oft-forgotten attribute of the maximum drawdown is how it relates to the
average drawdown. Ask yourself the following questions, “Would I rather experi-
ence plenty of tiny drawdowns than a few really large ones?” I think most of us
agree that we prefer even an infinite amount of smaller drawdowns to one large one.
However, when designing and examining a trading system, one large draw-
down doesn’t necessarily have to be that bad, as long as you understand what
might have caused it and trust all the other aspects of the system. The trick is to
get a feel for what created this drawdown and how it relates to all the other draw-
downs created by the system.60 PART 1 How to Evaluate a System