Thus, since it is the speed that is behind the final result, and not the other way
around, why not aim for the source directly? Obviously, plenty of other factors
influence which system to trade, but we will get to those in a little while. For one
thing, the average profit per trade, looked at all by itself, says nothing about how
reliable the system is and how likely it is that it will continue to perform well in
the future.
Even so, a couple of other reasons exist for why the net profit should be
avoided when evaluating a trading system, no matter how rigorous the testing has
been and how robust the system seems to be. For example, the total net profit tells
you nothing about when your profits occurred and how large they were in relation
to each other. This is especially important if the markets you’re comparing are
prone to trending.
If you’ve tested the system with a fixed amount of shares for each trade, it is
likely that the dollar value of each trade has increased with the increasing dollar value
of the market. This, in turn, means that the profits are unevenly distributed through
time, and the net profit is mostly influenced by the very latest market action. In a
downtrend, the opposite holds true. Notice, however, that the trend of the market says
nothing about whether the system has become more robust or not. In a market with
several distinctive up and down trends, this matter becomes even more complex.
In a portfolio of markets, the total net profit tells you nothing about how well
diversified your portfolio is. This is especially true if you stick to trading a fixed
amount of shares for all stocks, because what is considered a huge dollar move in
some stocks or markets is only considered a ripple on the surface in others.AVERAGE PROFIT PER TRADE
Knowing nothing else about the different characteristics of three different systems,
which one would you rather trade? One that shows an average profit per trade of
$290 over 101 trades; one that shows an average profit of $276; or one with an
average profit of $11. Assume all systems have been tested on the very same mar-
ket, over the same time period, ending today, and with a fixed dollar amount
invested in each trade. All profits are measured in dollars.
I take it most of you answered the first system, with an average profit of $290.
But what if I told you that the first system also produced a distribution of trades like
in Figure 2.1? Would this system still be your first choice after you’ve compared
Figures 2.1 to 2.3? If not, which one would you have picked this time? Why?
If you said that System 1 (Figure 2.1) is not a good alternative because it’s
obvious that the trades these days are way below the value for the average trade
over the entire period, you’re on the right track—sort of. But if you also said that
System 2 (Figure 2.2) is the best alternative because it’s equally as obvious that the
average trade for that system is way below what can be expected, as judged from
the last few trades, you’re still not right enough to justify real-life trading.16 PART 1 How to Evaluate a System