A strategy for trading 45-minute bars of the EUR/USD will be used to illustrate the application of MFE and MAE analysis. The strategy tracks the market trend, and if the trend is up, it buys after a pullback has occurred. If the trend is down, the strategy goes short after a rally.
The trend is determined using a concept devised from Gann swing trading: If the market makes three consecutive higher highs, the “swing” is up until the market makes three consecutive lower lows, at which point the swing is down. The trend turns up when price breaks above (by a certain number of ticks) a previous resistance point established by a swing peak. The trend turns down when a previous support level formed by a swing bottom is broken by a certain number of ticks.
This strategy uses a two-bar swing rule. It is an example of a buy setup. If the trend is up, go long when:
1. The market pulls back by forming a pivot low and then trading five pips above the high of the low bar, which is the setup bar. (The setup bar cannot be the high bar of the recent uptrend and entry cannot occur at the trend’s previous high or new high.)
2. If not filled in the next bar when the bar is an inside bar, the buy signal moves to five pips above the high of the inside bar, which becomes the new setup bar.
3. If filled, place an initial stop five pips below the setup bar’s low.
4. If long, once a bar makes a low above the setup bar’s high, thes top moves up to five pips below the low of the most recent bar or, if a pivot-low forms, five pips below the pivot low.
It shows a sell setup which reverses the rules of a long trade.
It contains the results of testing this strategy over two weeks of 45-minute EUR/USD bars. The system assumes the moment a bar closes, the position can be entered at the same price on the next bar.
There were only eight winning trades (16 losers, 1 breakeven), for a total loss of 107 pips. The average winner was 15 pips and the average loser was 14 pips. However, the average MFE was 20 pips and the average MAE was 14 pips. This implies some trades had open profits that vanished by the time they were exited.
It shows the profit/loss (P/L), MFE, and MAE for all the trades sorted by MAE. The majority of the large losing trades had very little, if any, positive MFE and most of these trades had MAEs larger than -20 pips. This suggests setting the maximum stop-loss point at 20 pips could reduce the size of the worst losses without giving up profits.
The average MFE was 20 pips, and trades reached an open profit of 10 pips or more 76 percent of the time. Taking this information into account and setting a 10-pip target and a 20- pip stop for each trade, a second back-test resulted in the equity line. Using the MFE-MAE analysis to set trade management parameters transformed this losing system to a profitable one on a hypothetical back-test.
This is only the first step — the research and development of the procedures. The next step would be to review more data (from a different period), as these parameters and procedures may be working only because of a volatility level unique to the particular time period that was tested. (This article has shown “insample” development of trading parameters and procedures. Applying these to other price data is called “out of-sample” testing.)
Also, keep in mind that volatility does trend. Therefore, procedures that worked or did not work two years ago will not necessarily perform similarly today. The lesson is that once you have settled on a trading approach based on both in-sample and out-of-sample testing, this MFE and MAE analysis should be performed on real-time trades on a regular basis. Then you can adjust your procedures to exploit the current market conditions. When you are trading in real-time you can compare how well you are executing your strategy under the pressure of real-time conditions vs. what the procedures call for you to be doing. This ongoing training process helps raise your trading skills to a level above trading on hunches.
Subscribe to:
Post Comments (Atom)
Post a Comment