The Forex MegaDroid Backtesting Blues – Track 2

2

After a lot of experimentation we finally managed to work out how duplicate the first trade in 2009 shown on the Forex MegaDroid website. Not only that, but using a demo account from FXDD that supplies historical data with only 4 decimal places, and working out exactly the right values for the Risk Level and GMT Offset input settings allowed us to duplicate the first 75 trades, up to and including the loss of $4672.00 on March 18th. After that though, the differences between our backtest results and those shown by the Forex MegaDroid team become more and more obvious.

For trade 76 on March 20th the entry and exit take place at the same time and price, but there is a small difference in size.  This could be explained by the two backtests not using exactly the same value for the Risk Level setting.  The trouble is Albert and John are obviously not using the 0.3 setting they talk about in their manual, so how can we be sure?

By trade 94 on April 24th the differences become more pronounced.  The entry is the same, long at a price of 1.3134 at 00:59 FXDD time, a time when traders in Europe and America are not sat at their desks, their eyes glued to their big screens. A bit further into the small hours of the morning the exits are different though. 1.3135 at 4:13 for us, 1.3144 at 1:18 for them. 1 pip for us, 10 pips for them. By the time you and I woke up on the morning of April 24th the bottom lines of the two simulated Droids were very different. $166.50 profit for the Trading Gurus. $1636.00 profit for the Forex MegaDroid team. How to explain that huge difference?  Only the Droid knows, but the Droid does not speak.

By May 22nd our version of the robot really started having problems.  On the EUR/USD chart it's little face was still smiling happily, and its comments told us it was version 1.11.  Here's what it did to our demo account:

# Time Type Order Size Price S / L T / P Profit Balance
213 2009.05.22 00:16 sell 107 20.01 1.3906 1.3938 1.3856
214 2009.05.22 00:25 close 107 20.01 1.3896 1.3938 1.3856 2001.00 83741.04
215 2009.05.22 00:58 sell 108 20.50 1.3905 1.3937 1.3855
216 2009.05.22 01:28 sell 109 16.76 1.3908 1.3969 1.3848
217 2009.05.22 02:58 s/l 108 20.50 1.3937 1.3937 1.3855 -6560.00 77181.04
218 2009.05.22 09:38 s/l 109 16.76 1.3969 1.3969 1.3848 -10223.60 66957.44

By way of contrast here is what the Forex MegaDroid team achieved in their simulated sleep:

# Time Type Order Size Price S / L T / P Profit Balance
209 2009.05.22 01:28 sell 105 19.60 1.3909 1.3973 1.3849
210 2009.05.22 01:46 sell 106 16.02 1.3912 1.3944 1.3862
211 2009.05.22 01:54 close 106 16.02 1.3902 1.3944 1.3862 1602.00 81699.70
212 2009.05.22 09:00 close 105 19.60 1.3908 1.3973 1.3849 196.00 81895.70

Do you see any similarity?  I sure don't. Once again random fluctuations in the fabric of simulated space-time in MetaTrader land adversely affected the Trading Gurus but not the Forex MegaDroid team.  Our bottom line a loss of $14,782.60.  Their bottom line a profit of $1798.00.  Everyone was asleep at the time, except the two robots of course.  They don't need any sleep.  Maybe ours needs a bit of maintenance now and again though?

Filed under Trading Systems by  #

Comments on The Forex MegaDroid Backtesting Blues – Track 2 Leave a Comment

March 4, 2010

jose @ 11:57 am #

what happen here? I just buy md, and I'm co
never see someyhing like this, who can
explain this? md soupport? nice work guys.

Leave a Comment

Fields marked by an asterisk (*) are required.

Subscribe without commenting