back start next

[start] [1] [2] [3] [4] [5] [6] [7] [ 8 ] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54]

8

1.5 Performance

This section establishes some guidelines on evaluating trading sjstem performance using the TradeStation Performance Report. For a thorough evaluation of a trading sjstem, refer to Stridsmans book Trading Systems That Worh [30]. As the trader wiU discover, the kej to any trading sjstem is to analyze rts drawdown in terms oflosing streaks and the size ofthe losing trade. Based on these data, we can calculate the appropriate amount ofcapital to risk per trade.

Table 1.8 shows a sample performance report. The Total Net Profit and Percent profitable mmibers are alluring, but the imponant number is the Profit Factor: the mmiber of dollars gained for each one lost. In this exanqde, dividing the Gross Profit of \$447,001.50 by the Gross Loss of \$174,787.00 yields a profit factor of2.56.

Reviewing some other ratios, the ratio ofthe average win to the average loss is \$4,217.00 divided by \$2,361.99 equals 1.79. The holding period ratio is the average mmiber of bars in the winners (30) divided by the a\erage mmiber of bars in the losers (16), approximately 1.88.

Table IS. TradeStation Strategy Performance Report - A Sjstem QQQ-lOmin

Total Net Prof it \$272,214.50 OpenpositronPA. \$0.00

Gross Profit \$447,001.50 Gross Loss (\$174,787.00)

Total # oftrades 180 Percent profitsble 58.89%

Ratio avgwiii/avg loss 1.19 Avg trade (win & loss) \$1.512.30

Max consec. Winners 7 Max consec. losers 5

Avg#barsinwinners 30 Avg #bars in losers 16

ProfitFactor Z56 Max # contracts held 9,500

To assess the impact of drawdown, multiply the largest losing trade (\$5,280.00j by the maximum consecutive losers (5) to get \$26,400.00. The actual maximum drawdown (not shown in the table) was \$19,720.00. The maximum intraday drawdown of \$21,993.00 occurred when the system was short before a surprise interestratecul,sowearefortunatetohavethispriceshockintheresuhs.

We look for month-to-month consistency with any trading sjstem, as shown in Table 1.9. Day traders should expea consistent weekly profitability. A swing trader should expect occasional losing weeks because the combination of time and losing streak makes rt almost impossible to avoid a losing week. For example, ifthe trader makes five trades a week and the maximimi consecutive losers is four, then the odds of a losing week are highly probable. Compare the actual monthly performance with the expected monthly income in Table 1.3 to set reasonable profit goals

Table 1.9. Montlil>Analyas

 Period NelPrl ProfitFactor # Trades % «14,s3z00 9.04% 2.21 13.50 48.15% february «8,994.25 4.00% 1.79 12.50 56.00% march »18,932.7s 9.21% 7.10 8.00 75.00% apru «19,467.50 10.16% 2.21 19.00 57.89% «24,333.00 11.53% 3.04 17.00 64. % june «18,020.50 7.6m 3.30 13.00 76.92% july «22,304.50 8.80% 6.79 16.00 68.75% august «27,699.00 10.05% 3.88 16.00 68.75% september «5,841.00 1.93% 1.59 10.00 50.00% october «10.592.00 3.43% 1.57 13.00 46.15% november «9,011.00 7.65% 1.64 11.50 60.87% december «20,508.50 12.79% 3.27 15.50 64.52%

As displed in Figure 1.14, the Eqidty Ciir\-e (EC) is a graph ofthe cumulative profit of a set oftrading sjstems. The vertical distance between each point on the chart represents the profit or loss ofan individual trade. The EC isjust hke a price chart-it has tiendandit has pullbacks (the distance from peak to tioughis the drawdown). Technical indicators such as the moving average and ADX can be calculated forthe curve to assess the strength of atrading sjstem.

Analyze the Equity Curve from a three-month perspective because a trader should expect flat periods lasting up to thirty or sixty dsys for a system. The EC in Fire 1.14 has rouly the same net profit for each three-month period. Plot the EC every month to determine whether or not the system performance is deteriorating,e.g.„ it advances half as much over consecutive periods.

Fure 1.14. Equity Curve

Finally, measure the distance (in terms of number oftrades) between successive equity peals and troughs to approximate the cjcle ofthe sjstem. This cj-cle is a function ofthe number ofwinning and losing streaks made by the sjstem.

1.5.1 A Tale ofTwo Stocks Ciena and Cigna

We finish the introduction with the tale of two stocks: Cigna and Ciena. Just one letter apart, the two stocks couM not have been more opposite in personality, one a staid insurance company and the other a volatile optical stock. Since Cigna was listed on the New York Stock Exchange and Ciena on the Nasdaq, a bitter livahj developed, so the two stocks requested a performance review from the Acme trading sjstems.

The performance reports in Tables 1.10 and 1.11 evaluate the unfiltered performance ofthe trading sjstems for both ofthe stocks. Clearly, any sjstem that generates a profit factor of 0.65 is useless for trading but instructive. For Cigna, the sjstems fared poorly, with only five out of eighteen winning trades. In contrast, Ciena performance results were just the opposite-a profit factor of 4.49 with only three losing trades.

The whole point ofthis exercise is to demonstrate that a sjstem or sjstems cannot be blindly apphed to a universe of stocks. First, we need to identify the characteristics that differentiate these stocks through a learning process known

:e Report - Acme All Strategies CI-Daily

First, we mine the trading filters to extract the characteristics that separate the trading stocks from the non-trading stocks. Then, we iterate through the characteristics to explain the difference in performance between two stocks. Clearly, we need to re-apply the trade filters ev-erjnight to create anew stock imiv-erse: a trading stock can reven to a non-trading stock and vice versa.

The first distinguishing characteristic is volatilit}. As shown in Figures 1.15 and 1.16, the HVjc for Cigna is 0.36 and the HVjc for Ciena is 1.09. Cignas HV does not meet the minirmmi thresliold ofO.5 set by the indicator, altlioueli its HV is turning up, and it may soon become a trading candidate. The choice of a minimimi threshold is a balance between discretion and automation, i.e., the higher the value, the fewer the mmiber of charts to review; a lower value means the trader exercises more judgment during the stock selection process.

The trader should test each Acme sjstem by stock sector. For example, the Acme N sjstem is a momentimi sjstem that performs wall on technolcgy stocks but fares poorly on cjchcal stocks. By testing each sjstem per sector over distinct time fi nes, the trader wiU develop an appreciation for the cychcal sjmbiosis between sector volatility and the Acme sjstems, just another way to obtain an edge (see Chapter 8). The stock selection process is methodical; aH siocks are funnelled through market and sector filters to obtain the best trading candidates. Through experience and experimentation, the trader wiU leam how and when to apply the sjstems.

CIEN LAST-Daily

ciLAST-Daily

Acme ¹(30,0.5) 1.14 0.50

ie.ooo ie.ooo

12.000 10.000

D.SO «40

Figure 1.16. High Volatility: Ciena

LIS. liw V.>:,t,liiv fif.iw

446499999999999

[start] [1] [2] [3] [4] [5] [6] [7] [ 8 ] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54]