Please find attached my first version of the EA for the BAT Method.
This is a canonical version. No, that has nothing to do with war. It means, that it observes the actual rules slaveishly, not offering anything to play with. It shall serve for back- and fronttesting.
As you can see I publish it only in compiled form. Fear not, that I shall try to suddenly sell the canonical EA. The method used is intellectual property of TeamAlphid and I have a job which pays my bills already. My main reason for doing that, is that I had some annoying experiences (people messing up my code and publishing it, people selling my EA without my approval, people Complaining that my EA that someone else sold them without my approval does not bring the promised BIG BUCKS etc. etc.).
Since there are not many features there is not much to explain. You can change the Lotsize and the Slippage. That is it.When you backtest and especially if you want to compare your results, with these discussed in the main thread, make sure that you have the Alpari data and 90% modelquality.
I should mention, that the EA does not rely on the BAT Indicator. I have put it in a function, which imitates the indicator, to speed up backtesting (factor ~ 10) and lower the systemload when actually running the EA (fronttesting).
There is one issue everyone, who wants to use this EA should know about. The banal expression "recent or significant low(high)" is laden with problems for a EA. The trained human eye is much more fit to find this low/high, then a robot. Now, being a fan of DeMark this should be done by fractals. After some backtesting I found the best is to look for the most recent 2nd level fractal (I tried 1st to 3rd level and the max out of 1-3 fractals, but the aforementioned seems the best). You might nevertheless sometimes disagree with the EA.
Post 55 for BAT EA v1.3 10 Feb 07
This is a canonical version. No, that has nothing to do with war. It means, that it observes the actual rules slaveishly, not offering anything to play with. It shall serve for back- and fronttesting.
As you can see I publish it only in compiled form. Fear not, that I shall try to suddenly sell the canonical EA. The method used is intellectual property of TeamAlphid and I have a job which pays my bills already. My main reason for doing that, is that I had some annoying experiences (people messing up my code and publishing it, people selling my EA without my approval, people Complaining that my EA that someone else sold them without my approval does not bring the promised BIG BUCKS etc. etc.).
Since there are not many features there is not much to explain. You can change the Lotsize and the Slippage. That is it.When you backtest and especially if you want to compare your results, with these discussed in the main thread, make sure that you have the Alpari data and 90% modelquality.
I should mention, that the EA does not rely on the BAT Indicator. I have put it in a function, which imitates the indicator, to speed up backtesting (factor ~ 10) and lower the systemload when actually running the EA (fronttesting).
There is one issue everyone, who wants to use this EA should know about. The banal expression "recent or significant low(high)" is laden with problems for a EA. The trained human eye is much more fit to find this low/high, then a robot. Now, being a fan of DeMark this should be done by fractals. After some backtesting I found the best is to look for the most recent 2nd level fractal (I tried 1st to 3rd level and the max out of 1-3 fractals, but the aforementioned seems the best). You might nevertheless sometimes disagree with the EA.
Post 55 for BAT EA v1.3 10 Feb 07
Attached File(s)
BAT EA V 1.0.zip
21 KB
|
3,610 downloads