acteristics of the system, so that the very best values don’t fall on the very edge of
the chart. If they do, there is no way of knowing how the system will perform if
the conditions happen to change in the wrong direction and cause the best variable
combinations to fall outside of the chart.
The larger square in Figure 19.1 encapsulates an area that is relatively large
and flat, which suggests that all trade-length/stop-loss combinations surrounding
the four-day/1-percent combination in the little square will produce similar results.
Therefore, the best alternative in this chart is the four-day/1-percent combination,
and if it just so happens that adjacent combination would have worked better in
real-life trading, at least we’ve gotten close enough to that combination to make a
profit, or at least avoid a catastrophe.
However, just looking at one chart like this isn’t enough; and that is where
Figure 19.2 comes into play. Figure 19.2 shows the risk–reward ratio as the aver-
age profit per trade, divided by the standard deviation of all trades. The greater the214 PART 3 Stops, Filters, and Exits
FIGURE 19.2
Surface chart with risk–reward ratio as output variable.