Books
in black and white
Main menu
Home About us Share a book
Books
Biology Business Chemistry Computers Culture Economics Fiction Games Guide History Management Mathematical Medicine Mental Fitnes Physics Psychology Scince Sport Technics
Ads

Mechanical trading systems - Weissman R.L.

Weissman R.L. Mechanical trading systems - Wiley publishing , 2005 . - 240 p.
ISBN 0-471-65435-3
Download (direct link): mechanicaltradingsystems2005.pdf
Previous << 1 .. 27 28 29 30 31 32 < 33 > 34 35 36 37 38 39 .. 82 >> Next

Mean Reversion Systems 79
TABLE 4.4 RSI extremes with 200-day moving average filter and 2.5% stop.
Asset Profit # Trades # Days Max Draw MDD MCL P:MD P:L Ratio %W Time %
ES 8711 50 16 -14562 742 5 0.60 1.24 48.00 28.94
IEURUSD 4764 39 20 -13922 1764 6 0.34 1.10 51.28 28.99
IEURJPY 30742 46 16 -15793 557 4 1.95 1.50 58.70 27.24
IEURCHF 14902 37 28 -8343 890 2 1.79 1.65 72.97 38.63
IEURCAD 1 7073 42 19 -30085 1385 6 0.57 1.19 47.62 29.13
IAUDCAD 17364 33 21 -10294 937 3 1.69 1.63 66.67 25.97
ICADJPY 5681 46 20 -20425 1017 8 0.28 1.11 50.00 32.78
IGBPAUD -9101 46 20 -24959 2541 5 -0.36 0.90 43.48 34.23
IGBPCHF 10052 33 28 -20123 791 5 0.50 1.23 60.61 34.23
Totals 100188 372 20.4 -44202 801 11 2.27 1.27 54.57 31.06
Note: All trade summaries include $100 round-turn trade deductions for slippage and commissions. Data source: CQG, Inc.
value of superior results in columns such as percentage of winning trades and maximum consecutive losses. (Note: Although this particular system experienced a large string of consecutive losses, as will be shown in Tables 4.5 and 4.6, this is usually not the case.) Experiencing more winning trades than losing ones or enduring a smaller string of consecutive losses can make the difference between sticking with a particular trading system long enough to reap its rewards and abandoning it.
Bollinger Bands with 200-day Moving Average Filter
One of the pitfalls of RSI extremes was that the average duration of a trade was the same as that of our intermediate-term trend-following systems. This problem can be eliminated by changing its exit condition #1 criteria, but to showcase another trend-following mean reversion system, I have chosen instead to introduce the fading of our trend-following Bollinger band system with the addition of a 200-day moving average criteria to filter out countertrend trades.
Using CQG, the programming code for this trend-following mean reversion system is written in this way:
Long Entry:
Close(@)[-1] XBELOW BLO(@,Sim,20,2)[-1]AND Close(@)[-1] > MA(@,Sim,200)[-1]
Long Exit—Condition #1 set “Price” field to:
BMA(@,Sim,20)[-1]
80
MECHANICAL TRADING SYSTEMS
Long Exit—Condition #2 set “Price” field to:
EntryPrice(@,0,All,ThisTradeOnly)-(.025*
EntryPrice(@,0,All,ThisTradeOnly))
Short Entry:
Close(@)[-1] XABOVE BHH@,Sim,20,2)[-r AND Close(@)[-1] < MA(@,Sim,75)[-1]
Short Exit—Condition #1 set “Price” field to:
BMA(@,Sim,20)[-1]
Short Exit—Condition #2 set “Price” field to:
EntryPrice(@,0,All,ThisTradeOnly)+(.025* EntryPrice(@,0,All, ThisTradeOnly))
Table 4.5 presents the backtested results from December 31, 1992, to December 31, 2002, for this system on our mean reversion portfolio.
As expected, the average duration of trades, time percentage in the market, and P:MD were all reduced considerably when compared with RSI extremes. This is because our mean reversion Bollinger band system exited at the 20-day moving average instead of holding trades longer to profit from the anticipated resumption of the longer-term trend. This same factor was also instrumental in the improvement of this system’s winning trade percentage.
TABLE 4.5 Bollinger bands with 200-day moving average filter and 2.5% stop.
# Asset Profit Trades # Days Max Draw MDD MCL P:MD P:L Ratio %W Time %
ES 6192 36 5 -8239 680 3 0.75 1.39 66.67 6.43
IEURUSD 530 33 9 -8801 1468 2 0.06 1.02 66.67 10.27
IEURJPY -9691 35 8 -20200 2426 5 -0.48 0.75 57.14 10.08
IEURCHF 14787 44 9 -4068 396 2 3.63 2.41 81.82 13.77
IEURCAD 7099 33 7 -14885 1037 3 0.48 1.17 69.70 7.23
IAUDCAD 17033 33 9 -3813 703 1 4.47 2.76 78.79 9.54
ICADJPY 1496 46 8 -10464 1467 3 0.14 1.05 65.22 12.31
IGBPAUD 20102 40 8 -10433 549 3 1.93 1.44 6.50 10.08
IGBPCHF -6442 34 9 -22436 2391 3 -0.29 0.83 61.76 10.46
Totals 51106 334 8 -26170 1780 4 1.95 1.44 68.26 10.20
Note: All trade summaries include $100 round-turn trade deductions for slippage
and commissions. Data source: CQG, Inc.
Mean Reversion Systems
81
NONDIRECTIONALLY BIASED MEAN REVERSION SYSTEMS

Thus far we have examined only trend-following mean reversion systems. In reality, most mean reversion traders have no bias against countertrend trading. Thus, now I offer some nondirectionally biased mean reversion trading systems.
Bollinger Bands with ADX Filter
Although I am certain that there are instances in which ADX improves the performance of trend-following systems, in general I have found greater success with this indicator as a filter for mean reversion systems. Here we take the mean reversion Bollinger band system previously used and replace the 200-day moving average filter with ADX. This removes the directional or trend-following bias and replaces it with a filter that is intended to ensure a nontrending market condition.
In addition, as opposed to exiting with profits at the 20-day moving average, this system will exit based on a percentage of the asset’s value at the time of trade initiation. Because certain assets are more volatile than others, we will set both the stop loss and profit exits at 2.5 percent of entry level for the E-mini S&P 500 and Japanese yen crosses. All other instruments will use a 1.25 percent move as the exit criteria.
Previous << 1 .. 27 28 29 30 31 32 < 33 > 34 35 36 37 38 39 .. 82 >> Next