Interesting GBP/USD MA Crossover system | Page 6 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Need help with MA crossover system 15 replies

Interesting GBP System 77 replies

A "different" crossover system - 600 pips on Friday alone in EUR and GBP 48 replies

Interesting Crossover - Am I seeing things? 11 replies

Scalping with crossover system 0 replies

Attachments: Interesting GBP/USD MA Crossover system
Exit Attachments

Interesting GBP/USD MA Crossover system

  • #101
  • Mar 2, 2007 1:36am Mar 2, 2007 1:36am
Quoting OwenT
are you THE man who will do the perfect EA? ...I'm so curious!
very great work btw
thank you thank you
  • #102
  • Edited 5:46pm Mar 2, 2007 5:32pm | Edited 5:46pm
I've just noticed something very ominous. According to my strategy, the core components being a 48 EMA, 72 EMA, and a couple momentum indicators, as of this Friday's close, the long daily trend on GBP/USD which started October of 2006 and at 1.8740 is OVER. You can see this illustrated by the plunge from the 48 EMA through the 72 EMA in one solid candle.

Further more, for the first time since April of 2006 when the even larger daily trend began at 1.7520, my system is calling for a strong sell on the daily charts.

Now fundamentally on the daily time frame, I'm not sure why the trend is over since everyone seems to be losing faith in the dollar, maybe this is just indicating all faith is not lost yet.
  • #103
  • Mar 3, 2007 5:11am Mar 3, 2007 5:11am
Quoting aicccia
the long daily trend on GBP/USD which started October of 2006 and at 1.8740 is OVER.
A very interesting observation. The price broke yesterday out of the daily SHI channel border (1.9469), and now sits right on what I identify as a support line (1.9433), and right below what DownRiverTrader here calls "zone -1" line (1.9443). So, Monday should be an interesting day...
sans peur et sans reproche
  • #104
  • Mar 3, 2007 8:39am Mar 3, 2007 8:39am
Quoting aicccia
The 4th version will be out within 12 hours.
Any progress with it? Let me know if you need help.
Thanks for sharing your strategy and thoughts with us.
  • #105
  • Mar 3, 2007 11:52pm Mar 3, 2007 11:52pm
yes I'm making progress. slowly but surely. If you want to help, tell me a good way to determine if the market is trending or ranging, besides looking at the width between two MA lines.
  • #106
  • Mar 4, 2007 2:53pm Mar 4, 2007 2:53pm
I know "0" about programming, so don't know if something like this is possible to program. Each pair has an average daily range. If it is trading at 50% of that range (variable) then it is ranging. Try different % figures like 50, 60 70, 75 to see how much difference it makes in # of trades and profit. Also maybe difference in daily closes could indentify if market is trending! If it is so many percent above/ below previous day it is trending. Just thoughts, Paul
  • #107
  • Edited 3:53pm Mar 4, 2007 3:40pm | Edited 3:53pm
Quoting plind
I know "0" about programming, so don't know if something like this is possible to program. Each pair has an average daily range. If it is trading at 50% of that range (variable) then it is ranging. Try different % figures like 50, 60 70, 75 to see how much difference it makes in # of trades and profit. Also maybe difference in daily closes could indentify if market is trending! If it is so many percent above/ below previous day it is trending. Just thoughts, Paul
Exactly what I was going to write. It's somehow connected to the daily range. The bad thing is that it won't tell you the future. For example if yesterday's range was only 50pips on cable, it's not a guarantee that it won't go 300pips in the same direction...
It's hard to do this, but if you figure out, you'll get rich in no time )

EDIT: Maybe you could use a volatility indicator! If that is in a downtrend, your EA is slowly switching to range-market but if it's starting to rise, you're switching back to trending mechanism...

EDIT: You could use the Bollinger Bands, and measure it's width. If it's ranging it's narrow, if it's trending, it's broad.
Or something.
  • #108
  • Mar 4, 2007 5:29pm Mar 4, 2007 5:29pm
aicccia, have been unable to get any buys during backtesting. Could you post the settings for 03. Thanks lj
  • #109
  • Mar 4, 2007 7:32pm Mar 4, 2007 7:32pm
Quoting aicccia
I've just noticed something very ominous. According to my strategy, the core components being a 48 EMA, 72 EMA, and a couple momentum indicators, as of this Friday's close, the long daily trend on GBP/USD which started October of 2006 and at 1.8740 is OVER. You can see this illustrated by the plunge from the 48 EMA through the 72 EMA in one solid candle.

Further more, for the first time since April of 2006 when the even larger daily trend began at 1.7520, my system is calling for a strong sell on the daily charts.

Now fundamentally on the daily time frame, I'm not sure why the trend is over since everyone seems to be losing faith in the dollar, maybe this is just indicating all faith is not lost yet.
nice call hmm? it probably will retrace tomorrow some.

On another note, I'm basically done with the code for this version. Right now I'm optimizing it again(its been a while since I've done that) and doubling everything for ranging markets on the 15 minute charts. As of now, the EA is 19kb where as the previous version was 13kb. Since it's grown and evolved so much, once I feel real good about this version I'll start a new thread with a more appropriate title in the Systems folder.
  • #110
  • Mar 5, 2007 6:03pm Mar 5, 2007 6:03pm
I thought I would give an update since things keep taking longer than expected.

The trending part of the EA is done. Right now it's optimized over the last 6 months and with 20% invested per trade, its taking $10 to around $17,000 with around a 5% drawdown(one bad trade out of 18 or 19). So the trending code does quite well. And this is not curve-fitting. Out of the 3000 passes during the optimization, at least half are giving me profits over $3000(remember, from just $10). Even I'm a bit amazed.

As I've said before, the ranging side is basically the trending code on the 15 minute timeframe. So today I doubled everything, marking everything with either an "H" for the hourly trades or "M" for the 15 minute trades. Then I ran into some problems. Apparently though you can run an EA with parameters working on different periods, you can't backtest an EA like that. This confused me and I was kind of overwhelmed. The EA is huge, it's only 600 lines but it feels like 6000. So anyway, I pulled myself together and realized I'm going to optimize all this on two separate EAs, trending on one, ranging on another, then stick them together at the very end. Hopefully it will be as easy as it sounds. This means I'll never be able to backtest the whole thing as one, which would be very nice so I could make sure the ranging trades which turn into trend trades end up getting processed as trend trades. We'll have to forward test the EA on a demo for a few weeks to make sure everything is working as planned.

Hopefully I'll be finished with this monster tomorrow.

Oh yeah and the trending has a profit factor of over a 1000. Beat that! lol
  • #111
  • Mar 6, 2007 2:40pm Mar 6, 2007 2:40pm
Aicccia,

Bravo!! I have been reading this thread anxiously. I must take my hat off to you ! You and the other members here are doing excellent work!

When you post the latest EA, I will tests it as well and post what results I get to contribute to this thread.

Keep up the great work!

Many Thanks from all of us..

Theo
  • #112
  • Mar 6, 2007 8:01pm Mar 6, 2007 8:01pm
Hey guys, this is the trending part of the MoneyMaker system, only the trending part. Right now I'm getting 24 trades, optimized over 6 months.

ROR seems to be around 75% to %100 per month on my standard account with 1:100 leverage, so for all you guys with higher leverage, your returns will be much higher. That is for about 20% invested per trade with 10% drawdowns.

More to come later. Test it yourself and tell everyone what kind of results your getting. A word of caution, on mini accounts on the strategy tester, some brokers(at least Interbank) put some kind of limit on the max trade size. And once it reaches this limit, the EA won't make anymore trades. So don't use too high of your account per trade. So when backtesting on mini accounts, test it out first on lower percentages, no higher than 5%, to see how many trades you should be getting.

And remember, you need the Osama indicator to help make the magic happen.
Attached File(s)
File Type: mq4 MoneyMaker-Trending-04.mq4   14 KB | 600 downloads
File Type: mq4 OSMAMA.mq4   3 KB | 592 downloads
  • #113
  • Mar 6, 2007 8:36pm Mar 6, 2007 8:36pm
I'm not happy though with this version. It misses too many trades.
  • #114
  • Edited Mar 8, 2007 1:54am Mar 7, 2007 2:55am | Edited Mar 8, 2007 1:54am
I have made some modification to it:

1. Add dynamic magic numbering, let it co-exist with other EA, and may attach multiple same EA with difference pair and timeframe in the same time.

2. Add trade comment to show program name and version number within trade record.

3. Add NormalizeDouble() function to OrderSend() parameter, to avoid be jammed when backtest with some un-normalized tick data. (as someone happened in page 7)

4. Add fixlots external variable, if set fixlots=0, the lot number strategy will follow the original design. When you set fixlots as non-zero such as: 1, 0.2, 1.5,....etc, the system will use this number as a fixed trading lot, and ignore the original risk_percent calculation. This will avoid to be jammed in backtest if the risk_percent become too large when equity going lower. (as someone happened in page 7)

I also found this EA perform good since 2006-Nov to now, but looks bad since 2004-Jul to 2006-Oct.

------- Updated
New version -dsy1.1 released as following:

1. Bug fix of dynamic magic numbering
2. Add deinit() function to release resource after EA end. without this function, this EA will eat up all of your memory when running optimization.

Next version to do: speed improvement.

------- Updated
New version -dsy1.2 released as following:

1. Speed improved up to 500% (50sec vs original 240sec test from 2004-Jul)
2. Add Normalization to OrderClose function.
3. Add TF external variable for more flexible trial.
4. Change exist trade searching from ascending to descending. Since ascending search will produce a bug if there are multiple trades generated by this EA.
Attached File(s)
File Type: zip MoneyMaker-GBPUSD.zip   13 KB | 524 downloads
File Type: mq4 MoneyMaker-Trending-04-dsy1.2.mq4   15 KB | 512 downloads
- Ray Tracy - Novice forex trader from GMT+8
  • #115
  • Mar 7, 2007 2:57am Mar 7, 2007 2:57am
Quoting raytracy
I have made some modification to it:

1. Add dynamic magic numbering, let it co-exist with other EA, and may attach multiple same EA with difference pair and timeframe.

2. Add trade comment to show program name and version number.

3. Add NormalizeDouble() function within OrderSend, to avoid be jammed when backtest with some un-normalized tick data.

4. Add fixlots external, if setup fixlots=0, all lots will follow the original design. But you setup fixlots as some number such as: 1, 0.2, 1.5,....etc, the system will use this number as trading lot, and ignore the original risk_percent calculation. This will avoid to be jammed in backtest if the risk_percen are too large vs. qeuite.

I also found this EA perform good since 2006-Dec to now, but looks bad since 2004-Jul to 2006-Nov .
cool, thanks for the interesting additions. It was only optimized over the last 6 months, so if you have good data, you should have excellent results since september.
  • #116
  • Mar 7, 2007 3:17am Mar 7, 2007 3:17am
Quoting aicccia
cool, thanks for the interesting additions. It was only optimized over the last 6 months, so if you have good data, you should have excellent results since september.
Yes, I believe it is good from last 6 months. But I would like to see it more adaptive to difference market condition from last 2 year. I am trying to optimize some external variable as you provided, and will post my result when it is finished.
- Ray Tracy - Novice forex trader from GMT+8
  • #117
  • Mar 7, 2007 3:31am Mar 7, 2007 3:31am
the market conditions of this last spring were entirely different from the market conditions of this past fall, so I doubt you'll be able to have it perform well with the same parameters over a years time. The EA is designed to fit to the current market conditions like a glove, and with optimizing once a week, the EA should always fit the market snugly. I believe 4 months is a good time period to optimize it to, but just to optimize over a few more trades, it was optimized over 6 months.
  • #118
  • Mar 7, 2007 3:38am Mar 7, 2007 3:38am
Also remember that this is a trending EA so it can be very harder to fit it to ranging markets, no matter how much you optimize it. The final EA for this system will work for ranging and trending markets.
  • #119
  • Mar 7, 2007 3:56am Mar 7, 2007 3:56am
Quoting aicccia
Also remember that this is a trending EA so it can be very harder to fit it to ranging markets, no matter how much you optimize it. The final EA for this system will work for ranging and trending markets.
Hmmm.....I think that would be better to wait for your complete system which include both ranging and trending condition......
- Ray Tracy - Novice forex trader from GMT+8
  • #120
  • Mar 7, 2007 8:43am Mar 7, 2007 8:43am
Aicccia/Ray tracy

Since some of us maybe new to Optimizing EAs, Is there any way you can provide instructions or provide a link for where we might find info on learning how to optimize an EA. Since this EA needs to be optimized weekly might be a good excersise if we all know the Optimization procedure. I know that their is an optimization feature in MT4 but dont know how to invoke optimized settings and then save the new optimized version.

This will be helpful so as more of us have this skill set, we all can work more collectively in Optimizing and Testing this EA.

Thanks
Theo
0 traders viewing now