Source: Microsoft Excel.

Figure 17.8 shows the Buy (above the x-axis) and Sell (below the x-axis) indicators for AT&T over the 10-year period as defined by the preceding trading strategy. Figure 17.9 shows the closing price of AT&T during this period, with the Buy and Sell indicators plotted. Buy indicators appear as diamonds above the price curve; Sell indicators appear as squares below the price curve.

Table 17.7 shows the results of the individual trades resulting from this approach. During the 10-year period of consideration, 16 trades were made for an average gain of 21.2 percent per trade. The average annual gain of each trade was 33.6 percent. With the exception of trade number one, all trades produce a positive return. For short positions, we simply calculate the return as the percentage gain between the Buy and Sell price.

Figure 17.10 shows similar chart analysis for IBM. Table 17.8 shows the performance results of the individual trades. During the 10-year period of consideration, 20 trades were made for an average gain of 34.3 percent per trade. The average annual gain of each trade was 33.5 percent.

Figure 17.9 price data and corresponding buy/Sell indicators for AT&T.

FIGURE 17.10 PRICE DATA AND CORRESPONDING BUY/SELL INDICATORS FOR IBM.

170 +♦,,

a 130

i1 1104 «

£ I <J

90 +

70 i

50 i

4v >•£, f

> it i

30 -»

-ft1 U"

N N

CM CM

Date

CM CM CO

0)0)0>0)0)0)0)0)0>

Conclusion

Statistical Network data mining provides a highly automated ontogenic approach to interpreting technical indicators for daily price data. It allows creativity in defining a robust set of input indicators, and flexibility in defining different trading strategies. The results shown here demonstrate the validity of this approach.

Endnotes

1. Takens, E, "Detecting Strange Attractor in Turbulence," Lecture Notes in Mathematics, D. Rand, L.Young (Ed.), Berlin: Springer, 1981.

2. Minsky, M., and Papert, S., Perceptrons, Cambridge, MA: MIT Press, 1969.

Table 17.8 Trading results for IBM