I understand where you coming from holyguy7 but...what can I say here that will not seem like a personal attack at you...it just seem that many of us have helped testing different settings of PG 2.6,2.7,3.2.4 etc,,,and at the end you are keeping for yourself. Nice, I feel warm all over.
Anyway, can't blame you. I will blame it on human nature
Please tell us which version you are using?
I've been following this tread and PG for a while now.
I even had another programmer code the PG in TradeStation 2000i to get accurate back test results. no good!!
Still no one has found the magic settings. If you have good settings then please share it.
Everyone on these threads contributed in some way to your success and finding those settings.
We are all trying find good EAs. a 100 people in this forum will not take 1.7 trillion from the market and leave nothing for you!!! Don't be so mean !
I tried to optimize the settings of this EA (attached) for H1 timeframe.
The settings (pre-set files) and backtesting results attached.
I did it with every tick, 90%, since end of 2004 till April 2006.
I don't think that it will be profiable every year. But it was very good for some pairs since 08/11/2004 till 26/04/2006.
Preset file for AUDUSD (attached).
0.1 lot size.
Every tick, modeling quality 90%,
since 08/11/2004 till 26/04/2006,
Average Profit per 0.1 lots = 111.26
Average Loss per 0.1 lots = 209.67
Maximal drawdown (%): 10.7
Profit factor 1.59.
Total trades: 44.
Total net profit: 1365.34
Preset file for EURCHF (attached).
Average Profit per 0.1 lots = 74.41
Average Loss per 0.1 lots = 27.50
Maximal drawdown (%): 0.5
Profit factor 6.76.
Total trades: 14.
Total net profit: 634.08
Preset file for EURJPY (attached).
Average Profit per 0.1 lots = 57.87
Average Loss per 0.1 lots = 74.05
Maximal drawdown (%): 7.8
Profit factor 1.11.
Total trades: 148.
Total net profit: 517.86
Preset file for EURUSD (attached).
Average Profit per 0.1 lots = 67.74
Average Loss per 0.1 lots = 92.66
Maximal drawdown (%): 9.6
Profit factor 1.24.
Total trades: 100.
Total net profit: 839.20
Preset file for GBPJPY (attached).
Average Profit per 0.1 lots = 43.75
Average Loss per 0.1 lots = 42.31
Maximal drawdown (%): 2.5
Profit factor 1.69.
Total trades: 142.
Total net profit: 1565.43
Preset file for GBPUSD (attached).
Average Profit per 0.1 lots = 67.08
Average Loss per 0.1 lots = 37.90
Maximal drawdown (%): 3.0
Profit factor 1.33.
Total trades: 128.
Total net profit: 922.87.
Preset file for USDCAD (attached).
Average Profit per 0.1 lots = 54.31
Average Loss per 0.1 lots = 45.61
Maximal drawdown (%): 2.4
Profit factor 2.62.
Total trades: 64.
Total net profit: 1477.65
Average Profit per 0.1 lots = 51.92
Average Loss per 0.1 lots = 67.76
Maximal drawdown (%): 4.7
Profit factor 1.34.
Total trades: 96.
Total net profit: 795.52
Thank you newdigital
Thank you newdigital for this report.
90% modeling quality results
My cousin and I have been revising the SuperFXAnt that I posted earlier and have found some settings that work well. We modified the original SuperFXAnt to have a few new features for customizability. Now, I don't really want this EA to end up like the profit generator with so many settings that you can't figure out a good combo. If you do have a good suggestion or change, then make it. I like learning from other people. More minds make for better results. I wanted to leave a complete 1 year report but I couldn't get it to upload. So the yeargraf.gif is the graph of one year ran from 5-1-05 to 5-10-06 with the same settings that you can see in the one month that I ran so that I didn't have to put all the settings here. Just look at the top of the report and you will see my settings. They were both ran on the Eur/$. That's all that I've really worked on for now. As you can see my settings were for the aggressive version.
We have made it so that you can choose either the aggressive version or the more conservative version. I don't really want to get into what the difference is, but if you can understand that code just a little, you will be able to figure it out. Anyway, we live in Utah so we want to use InterbankFX for our broker since it is just down the street but they don't allow scalping so we had to as the scalp control feature. Interbank said that as long as the trades last more than 3 minutes it's not considered scalping. So we built in a time delay so that we could avoid having problems with Interbank. So the leeway option you see in the settings is just the stoploss and takeprofit during the scalping delay to keep from hitting the s/l or t/p before the scalp delay time is up. I also added the feature Maxbarspread to be able to stop the EA from trading when the bar gets to long because I noticed that when the bars are long it usually indicates a strong movement in one direction which kills the aggressive style. Oh, and by the way, the Maxbarspread only effects the EA when the aggressive style is chosen. The conservative style really doesn't need that feature because it won't usually make trades on a strong movement bar.
By the way, the TF is 30min.
Let me know what you think.
p.s. I haven't worked much on the conservative method's settings, so I'll post those when I have them.
p.s.s. I'm currently forward trading the settings that were used in the examples I've attached.
Oh, I forgot to mention that the above result from the back tester were done using the every tick method.
Just in the last couple of minutes since my last post I've been messing with the TF 1H and the results are better. Try using the same settins as I used above only changing the Maxbarspread to 40. Leave all the settings the same and the results are nearly double that of the TF30 min. Before I had the 90% modeling quality the TF 1H wasn't all the great, but I think I'm going to start a forward test using the TF 1H instead of the TF 30min.
I appreciate all your input and comments even if they are negative. Thanks all.
will forward test and let ya know
Backtesting on the same bar is very dangerous.
Wreid this is my backtest resoults on 90% quality as well.
What does backtesting on the same bar mean?
Here are my result with the exact same settings and time frame as yours above.
wreid completly different resoults.
I used FXDD for the backtest
Backtest resoults do funny things when tester opens and closes deals on the same bar