The first thing I looked for and confirmed is that 22-23 GMT retrospective tests are really producing results much better than their counterparts 21-22 GMT. This time even the GBPCHF is best. In the light of these tests, I think 22-23 GMT is easily the best time interval to run.
Important Edition 10.05.2011: according to users (see comments below) and the author, in the light of recent changes (there were several new versions since I wrote the article) the interval 21-22 GMT is now better. I’ve changed the interval of one hour of my test forward and I will leave that trade with its range of default, for the moment, at least until you get the opportunity to do some retrospective tests of my own with the latest version.
Let’s take a look at the differences between the retrospective tests simulated ECN (lower spread with 0.8 pips Commission) and retrospective tests (STP further spread, no commissions). In many cases, STP retrospective tests came out better. Why? Because couples with low spread, such as EURUSD, the 0.8 pips Commission brings the cost per operation above costs by operation for a corridor of STP. One thing to keep in mind when making this comparison is that the differentials that I used for ECN agent are maximum values normalized, while for the DDN corridor / STP were normal. We are comparing the worst cases to the ECN STP case medium, so more or less peanuts and apples, but in the end, if we were doing the same procedure that on average in comparison with the average I think STP would not far behind. The question that naturally follows is: see these differences, is that really worth running it in an ECN broker? And my answer would be: Yes, provided you have a balance sufficiently high to allow the use of the money management with an increment of 0.1 field size.
Changing the subject, let’s compare the retrospective tests true spread against retrospective testing fixed-spread. In all cases, things were much worse and I am asking myself: why? As I mentioned in the review EURClimber, historical data of Dukascopy spreads are wider than current spreads on ECN brokers, Dukascopy data is quite old and spreads again in 2007 were not what they are today. However, I wanted to see what is the differential medium that is causing this to happen, so I ended up writing a small for this. When backtested, this calculates an average of dynamic propagation for a selected period of time and keeps the record of mail trash with her. In case someone needs it, it is available for download.
I’m going to create a small table extension with all the attributes that I have used and the values were back testing. AverageSpread previously mentioned FXT files using Dukascopy differentials. Once again, ECN spreads are the maximum standard FXOpen extends from the Asian session, while STP spreads are the average markets extends to the Asian session.