The EA was written to allow it to run continually for the week. I spent all last week coding a new modified system that changes the MM to add a lot for a win and subtract for a loss. It only allows trading up to 5 lots and sets the TP for each lot 1/4 of the magic number greater to allow for long runs. It will also adjust the SL of multiple winning lots to just behind the TP of one that closes, kind of like a moving SL. This EA will also not cancel the opposite pending order if the “short term trend” is against the opening. For example: If yesterday’s close was lower then the day before, it is termed a short term down trend. If the buy pair is opened in this case, the sell stop pair is not canceled. Also, I adjusted the opening and TP points so a pair in the direction of the trend takes less to open and the opposite takes a little more. I basically halved the TP points as well to try to allow for a greater number of wins without being stopped out by retraces. This EA can be run indefinitely as it does not allow for trading on Sundays. I would still probably stop it after Friday and restart for Monday.
I’m busily testing some other strategies including some very interesting and promising Point and Figure charting stuff, so I’ve let the straddle sit for now. Here’s a link to the improved (hopefully) EA if you’d like to test it this week. Let me know how it goes.
[I]Right click on the link and “Save As”. Place in the Experts folder.[/I]
Had a few technical hitches last Sunday while trying to set this up. I wanted to run it on the UK Alpari broker MT4 platform. I could not get going, they have 5 decimals not sure if that has anything to do with, was then out of town during the week, will try figure it out to give it another go next week.
first: I’m a really FX newbie. I’m normally an option trader, but these times, with volatility in the sky, theta has become quite uninfluent over delta, and selling time premium is quite risky! Therefore, a so “directional” market has pushed me towards the FX, at least for a while or for a part of my trading time/interests.
This also means that, used to ToS, where you have any kind of orders available, though not a scripting language, I found myself a bit disoriented into IBFX, with a very nice analytical platform, and scripting possibilities, but nothing like a simple OCO!
This thread is really nice, because there is sound “forward- testing” (I trust it much more than “back”-testing), and a method-system with solid math-stats grounds.
Thanks a lot especially to bkr1969 for his scripts, which are wonderful (though the last one is really very complicate). I’m still testing the “previous one”, which is posted on brk’s web site.
[ol]
[li]As I am on the other side of the ocean, and waking up every night at 2:00 o’clock for running the scripts manually is quite unpractical, I wanted to let the EA running alone without manual confirmation, and therefore I went back to the backtesting! Brk: the results are quite depressing, with any of the currency pairs tested. Is this due to the imperfect MT4 historical data storage or is there some other reason?
[/li][li]This was true especially if I applied the mm mechanism proposed. From my option’s exp, actually the mm algorithm scares me quite a lot. We had a TV show called “let it or double it”, and this is not the way I’m used to trade… Therefore I tried to “cannibalize” brk’s scripts and let the lot size always constant. But again the results were not good (although a bit better…)
[/li][li]What I’m also trying to do is to modify the scripts for applying the system to the four hours time frame. The method is a quite rational breakout strategy, and, except for very short time frames, it should work. Of course the magic number will be smaller, and therefore profits (but also losses!). But in a day I can make 3-4 cycles if I run it manually, or 6, if it runs without supervision Let me know what do you think about.
[/li][li]Finally I want now to go through the last, much more complicate, brk’s code. Would it be possible changing it in order to:
[/li][/ol]
[ul]
[li]-use the 4hrs time frame
[/li][li]- let the mm method a bit more free or isolated, so that it will be possible applying another one or let simply the lot size =1?
[/li][/ul]
I will try, but of course if there is some more expert programmer giving other contributions it would be very nice.
Sorry for the lenght! Hope this will anyway contribute…
First of all, you are are right about the back testing results. Only one pair didn’t blow an entire account in the past year and that one still lost some. I don’t know if there’s an imperfection in the testing parameters or a huge difference in market conditions or what. That is one of the reasons I’ve been looking at a few different strategies.
As far as re-configuring the latest EA, it should be no problem. The “magic” number needs to be changed to reflect 4 hr instead of daily. Change this part:
The initial lot size can be set when the EA runs, but the code could easily be modified for different MM styles. I have a friend who’s been testing this using straight single lots and has taken his practice account from $3000 to over $11000 this month, so there are obviously different options.
Play with the code and make any modifications you’d like (it’s fairly well commented). If you have any questions, just ask.
Thanks brk,
I already made the modifications you suggested in the code. Modifying it for the mm looked also very easy, but I should have made some mistakes because it gives some strange errors. I want to give a better look to the final, long and complex, and then eventually come back.
I’m anyway quite scared by the backtesting… as you are.
Is the new system you’re working on the other that PIP CHASER posted under “simple day system”?
