Hi JCL, what happened was that I looked at the trade station platform and set up the charts. The system should have opened a trade and Zorro didn't open it. I've actually had an mt4 ea coded for a system I developed with Zorro and it was profitable on both so I'm happy that Zorro is useful to develop mt4 ea's. It's just that for me I'd prefer to run things on mt4 for the reasons stated above.
Yes, but that's exactly the problem. The false trade signal will also happen in real trading. It will then open a position where Zorro would not. This imposes an additional risk to your strategy.
If the MT4 backtest returned about the same profit as the Zorro backtest, it's probably ok - small differences can be caused by the slippage simulation. Still, a difference between testing and trading is always a potential danger. You should compare the signal values and the trade logs. There can be many simple reasons for such differences, but do not ignore them unless you know what's happening.
Even a profitable strategy gives you only a small edge. Any little mistake can take it away.
Re: Plugin for IBFX
[Re: jcl]
#410284 10/31/1215:4210/31/1215:42
No, it was definitely a trade that should have been opened according to the rules of the strategy but it just didn't open.
What I am doing now is developing the strategies with Zorro, coding them into mt4 and then doing several back tests on different instruments and with different brokers. This I feel will be the closest thing I can do to be certain that I have a positive expectancy.
Because I had FXCM trade station open with the chart and it was obvious that a trade should have been opened according to the rules of the script that I was attempting to trade live.
Yes, I understood that already. But the obvious question is who knows better when to open a trade: the chart of your FXCM trading station, or your Zorro script?
I'm just trying to tell you that what you're doing is high risk. You're going to trade a strategy that behaves differently on different platforms, based on a script that did not open trades when it should, for a reason that you did not understand. This sounds like a recipe for disaster. 10% wrong trades can already turn a winning system into a losing one. If you can not find out why your script does not trade as it should, open a thread about it in the script forum, post the code, and you'll get help.
Re: Plugin for IBFX
[Re: jcl]
#410291 10/31/1218:3110/31/1218:31
I thought Zorro was based on the FXCM TS feed? So one should follow the other. It didn't open the trade because there was something wrong with the Zorro platform, not the script/system.
I fully intend to back test everything thoroughly before I trade live so it is no problem. If something doesn't back test right on mt4 then I will not be trading it live. In most occasions I have found that a profitable system produced with the help of Zorro trades well with mt4 as well, which is encouraging.
When Zorro is more robust for trading I will probably make the switch.
Only your script determines when to trade, not the feed or a "Zorro robustness". But anyway, I can do not more than offer my help - after all, it's your money .
Re: Plugin for IBFX
[Re: jcl]
#410301 10/31/1220:4010/31/1220:40
Then you find out why, in a systematic debugging process.
Programmers do that all the time. Beginners are not used to it, so it's difficult to them. That's why I suggested that you post your problem in a separate thread, and I'll help finding the reason, step by step. This way other users also can learn from it. You're not alone: anyone writing scripts will encounter such problems sooner or later.