shares bought and the dollar amount invested with the changes in the equity. Only
then will your profits per trade grow, as shown in Figure 2.2, no matter the current
trend or market conditions. The discussion for why that is will have to wait until
we get to Part 4. To get to this stage, you first need to test your system so that it
trades equally as well, on average, at all times on a one-share basis, with the prof-
its and losses measured in percentage terms, which is what we’re discussing now.
We want to measure the system’s performance in percentages, because as the
dollar value of the market changes, so will the dollar values of its moves. Thus, the
move in relation to the current market value will stay approximately the same. This
means that a low-priced market is more likely to produce small dollar moves than
a high-priced market, and vice versa. If you test a system on a low-priced market
for an average profit of $1 per share, produced by an average move of 2 percent,
the average profit in a higher-priced market will not be $1 per share, but 0.02 times
whatever the market value of that market (everything else held equal and provid-
ed the system is considered robust and stable in the first place).
Had you not known that the average profit per share traded was higher than
what your testing showed you, you might have discarded the system as not good
enough. Worse yet, reverse the logic and you might have traded a system that you
thought had a high average profit per trade when in fact it had not. Thus, the sin-
gle most important thing to think about when building and examining a trading
system is to do it in such a way that you can estimate what the profits and losses
are likely to be at today’s market value (or at any other point in the future), not
what you could have made in the past. Only then can you make your results for-
ward looking and, for example, calculate how many trades it should take you to
make a certain amount of money or trade your way out of a drawdown, as this
example will show you:
If you know that your average profit per trade, given the current market values
of the markets you’re trading and the number of shares you trade in each market,
equals $500, and you currently are in a $7,200 drawdown, the estimated number of
trades to get you out of the drawdown will be 15 [Integer(7,200 / 500) 1].AVERAGE WINNERS AND LOSERS
Note that a 2-percent move doesn’t seem to be that much ($2 for a $100 stock and
$1 for a $50 stock), but for now, we’re talking on average over all trades, counting
both winners and losers. The average winner might be larger, say 4 percent, but
since we’re bound to a have a few losers as well, the average profit for all trades
might not be any larger than this.
Say you have a system with a stop loss of 1 percent and a profit target of
4 percent. This system is going into each trade with an estimated risk–reward
relationship of 4:1. However, because not all trades will be winners and fewer still
will be stopped out with a maximum profit of 4 percent, the actual risk–reward20 PART 1 How to Evaluate a System