Interesting GBP/USD MA Crossover system | Page 5 | 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

  • #81
  • Feb 19, 2007 12:18pm Feb 19, 2007 12:18pm
Quoting OwenT
I have still the problem that the EA doesn't want to reopen a trade after a losing one. On the picture I think that I would have a Sell after the first one (losing). Or maybe I missed something in the rules?
According the strategy that I've got in my head, you're right it should have opened a trade there. But it's very hard to get an EA to think like us and see when it should and shouldn't open a trade. A chart is like a crime scene and if you look hard enough you can always find clues to pointing to what's going to happen next. We as people analyze many different parameters without even knowing it when we glance at charts, like the the steepness of all the ma lines, the steepness of the osama lines, where all the lines are crossing 0, candlestick price action, trend lines that may be draw on on the chart, knowledge of news coming out, ect. But it's very hard to step back and recognize what exactly is telling you the trade should have been made at a certain place. It's never just one or two variables, it's always at least 4 and usually more.

I know that specific time period well because that missed trade and then that point where it should have taken a trade, but didn't, has been bugging me too. It isn't taking the trade the second time because of some parameter isn't met that normally is required for a good trade but wasn't here. I'm still trying to figure out what to change so the EA will see the charts as I do and realize it should be placing a trade there.
  • #82
  • Feb 19, 2007 12:19pm Feb 19, 2007 12:19pm
Quoting aicccia
We are go for breakout!

Open Trade #3
Febuary 19th, 1:00 PM ET
BUY @1.9527
it better be a breakout....

whoever is selling don't make me come find you.
  • #83
  • Feb 19, 2007 12:21pm Feb 19, 2007 12:21pm
breakouts can happen this late in the day right?
  • #84
  • Feb 19, 2007 12:25pm Feb 19, 2007 12:25pm
Quoting aicccia
We are go for breakout!

Open Trade #3
Febuary 19th, 1:00 PM ET
BUY @1.9527
I personally have already closed that out. But we can pretend I didn't and see where it goes. Looks like its going down. lol. This Sucks!
  • #85
  • Feb 19, 2007 12:56pm Feb 19, 2007 12:56pm
Aicccia, I have bookmarked your thread. For those that are not believers of the starter of the thread or that MA systems work, this is from a contest that Aicccia is in.

aicccia2969,969,804-5.820.3057.09
The 57 is for % gain on the account in just about 3 weeks.

Dude, you are rocking! I know MM isnt the same on the demo, but still. When I get a free day I will certainly read through it. My single % returns monthly aren't worth the time.
The market is my nation. Traders, my family. Hello, brothers and sisters!
  • #86
  • Feb 19, 2007 4:08pm Feb 19, 2007 4:08pm
hmm 57% most have been from a few days ago. In the contest the trades are more discretionary while I'm just relying on the system for assurance that I'm trading in the right direction. Like I said in the other post, it's many things, not just one like MAs, that are influencing my decisions. For example, I closed out my last trade just a few minutes after I opened it because it was clear the breakout wasn't going to happen. But there is some GBP favorable news coming out in the next 24 hours and since the price seems to be slowly wanderer acroingss the trendline instead of shooting up it, I'll probably buy again before then. Momentum is still up, at least right now it is. My ROR in the contest as of today should be around 92%. But that's with putting the whole account behind each trade, which I don't do with my real account. But thanks for the complement.
  • #87
  • Feb 19, 2007 4:53pm Feb 19, 2007 4:53pm
Quoting aicccia
This is the next version. I've added:

1)An equity checker that will close the trade if you lose a certain percentage of the max equity. For example if things go well and you're up 200 pips, and then something happens and it drops to 130 pips and stays there for a few candles, and if you had your max equity to lose set at 75%, then the trade would be close. Keep in mind it measures all of the pip movements off a 3 EMA, so to keep spikes from messing this up.
2)Now the EA looks back up to 6 candles to see if there was a jump across the two MA lines.

thats it for now, but there might be a new one coming out tomorrow night.

keep in mind this EA does not get every move, it does not even get every big move, but when it does make a trade, there are several conditions saying "yes, yes, open now" so it's very unsual for things to go wrong in a big way. It is sensitive to ranging markets and it's possible to go 4-5 days without a trade if ranging is what it is detecting. The bottom line is the EA opens trades only in perfect conditions.

The EA was developed with the GBP/USD in min so I haven't tried it at all on other charts but if you want to optimize it for other currencies be my guest.
Make sure you check "genetic algorithm", have 90% quality data, and test at least 3-4 months. Optimize the parameters in two sets. The first set being: start stop
fastma 6 92
slowma 6 144
currency_momentum 6 92
trend_momentum 6 144
maxpipsfromEMA 1 50
trending 1 50
SpikeHeight 25 150

and the second set:

buy_maxhoursNeg 3 60
buy_maxpipsNeg 20 100
sell_maxhoursNeg 3 60
sell_maxpipsNeg 20 100
startcountpips 30 150
startcounthours 3 48
maxEquitytoLose .2 .8

when optimizing the first set, disable the stop settings and save equity settings but setting their values to false.

For the pairs that move very fast such as GBP/JPY, you might want to increase the numbers to stop optimizing to.
Why are there two set of settings for each variable? Am I suppose to have two charts open with group 1 settings, and the other with group 2.

ET
  • #88
  • Feb 19, 2007 5:00pm Feb 19, 2007 5:00pm
Quoting et_phonehome_2
Why are there two set of settings for each variable? Am I suppose to have two charts open with group 1 settings, and the other with group 2.

ET
no no those two sets are just if you want to optimize the EA on your own, on maybe another currency. The numbers are the two numbers you would let the optimizer optimize in between. For example, the optimizer would look for the best value for fastma between 6 and 92.
  • #89
  • Feb 19, 2007 5:10pm Feb 19, 2007 5:10pm
Here's the next EA version.

1)Added support for standard accounts and leverages of all shapes and sizes. Just set "mini_account" to true or false depending if your account is mini or not and set "leverage_1_to_" to the leverage amount. For example I have a mini account with 1:200 leverage so I'd set "mini_acount" to true and "leverage_1_to_" to 200.
2)the equity save doesn't seem to be helping save money any, in fact when I let it run, my profits are about $1000 lower. Since this seems counter-intuitive, I'm going to keep checking the code and try to get it to work better. So for now just leave it on false, so the EA won't run it, till I figure out whether to just delete it or how to make it better.
3)All the values for the parameters are the updated optimized parameters for this week. So all you need to change is the account setting to suit you account

I will be out of town most of this week so I won't get much work done on the EA
Attached File(s)
File Type: mq4 MoneyMaker_03.mq4   13 KB | 660 downloads
  • #90
  • Feb 19, 2007 5:18pm Feb 19, 2007 5:18pm
I only see the EA taking half of the trades that it should be taking so it will be getting much better.
  • #91
  • Edited 8:44pm Feb 19, 2007 8:18pm | Edited 8:44pm
Quoting aicccia
I personally have already closed that out. But we can pretend I didn't and see where it goes. Looks like its going down. lol. This Sucks!
it looks like this trade is going to turn out okay after all
  • #92
  • Feb 19, 2007 11:55pm Feb 19, 2007 11:55pm
Can not to make MA open 0.1 lot while backtesting. May be it would be better
to make changeable lot size in variables?
  • #93
  • Feb 20, 2007 12:14am Feb 20, 2007 12:14am
Quoting forig
Can not to make MA open 0.1 lot while backtesting. May be it would be better
to make changeable lot size in variables?
forgive me forig but I'm not sure what you mean. The size of the lot the EA opens is whatever % of your account that you wish. Look for "risk_percent", that's the parameter you'd edit to change trade sizes.
  • #94
  • Feb 20, 2007 12:34am Feb 20, 2007 12:34am
Quoting aicccia
forgive me forig but I'm not sure what you mean. The size of the lot the EA opens is whatever % of your account that you wish. Look for "risk_percent", that's the parameter you'd edit to change trade sizes.
OK, then try make backtest for 3 month with 0.1 lot. EA opens only 1 trade in november and that all. I made % = 0.01 and leverage = 100.
  • #95
  • Feb 28, 2007 12:40am Feb 28, 2007 12:40am
How do you back test it. When I do it I get no results.


Quoting aicccia
here is a EA that works with this strategy as the baseline strategy, but it includes a sophisticated spike detector and a stop loss min-system plus many other entry conditions to make it more precise. the EA as a whole is very complicated as a whole as you can see if you take a look at it but it's very logical in the sense of price action once you work it all out. It has 12 optimized parameters, and the whole EA should be re-optimized every weekend over the last 4 months. I'll post the best parameters here every Sunday. With the default parameters, which are the current ones for this week, with 10% invested each trade, it turns 200 into around 4600, with just a 7.9% max drawdown. You may think this must be curve fitting, but I've done that before and this is nothing like it. When you curve fit, you change the parameters so you get the highest balance, but the parameters don't make sense from a logical, price action point of view. These parameters make sense, which you'll see if you can figure out the EA, and these are not the only parameters that give me good results. When genetically optimizing, out of the 10,000 possibilities tried, 1/4 to 1/3 gave me over a 2000 profit from 200, and around 1/10, gave me a balance of over 3000. So even if these parameters were the best for the last 4 months, but not for this next week, they'll still work out very well.

I've posted a few systems before that therotically gave me very high profits, but this is the first time I've turned one into an EA. I also am not done with this EA and I'm improving it daily. The one big fault I have with this systems reliability is that all of this is based on just 16 trades. But they're long, logical trending trades so I think it's okay.

And I almost forgot, this EA and system use a custom indicator I call OSAMA that is just an Osma indicator turned into two ma lines, averaged out to the period you choose, so you can see if the general trend of the osma is up or down. It's an excellent indicator for momentum and that's what I use it for in my EA.
  • #96
  • Feb 28, 2007 1:36am Feb 28, 2007 1:36am
Quoting hccgtim
How do you back test it. When I do it I get no results.
try my next version tomorrow, the previous versions had problems with adjusting to different accounts.

I've found a pretty damn good way to tell the difference between ranging and trending markets so I've added that, so now my trending system takes it's trending trades only in trending markets, which made it a whole lot better since 75% of the losses came from attempted trades in stagnant, ranging markets.

I was in NYC all last week but I'm back now so sorry for the absence.

My goal from here is to 1)tweak the trending aspect of the EA but more importantly 2)create a satisfactory ranging system so I can switch back and forth depending on the market's behavior. My idea for the ranging markets is that ranging is just trending on a smaller scale, so it will basically be the same system that I use for the trending markets, just on a 1/3, 1/4 or 1/5 scale with quicker close mechanisms.

The 4th version will be out within 12 hours.
  • #97
  • Mar 1, 2007 2:04pm Mar 1, 2007 2:04pm
Thanks Aicccia, waiting your new version.
  • #98
  • Edited 3:30pm Mar 1, 2007 3:19pm | Edited 3:30pm
lol boys and girls, let me explain something. The EA has two main parts, one part that trades when it's trending and one when it's ranging. They're almost the exact same code, just the ranging code is all on 15 minute timeframes instead of hourly timeframes. This "Code" is made up of 3 different types of entries. I call them the jump trades, the swing trades and the waterfall trades. The Jump trades signal the beginning of a new trend when the price jumps two MA lines and changes momentum at the same time. Swing trade are when the price in a trend retraces and then resumes the trend. When the trend resumes, a swing trade is placed. The waterfall trades signal the beginning of a new trend also but the price, instead of jumping MA lines and changing momentum all at the same time, slowly slowly changes momentum and like a waterfall or see-saw the price slowly falls or rises into the new trend.

Right now I have those three types of trades under true/false conditions so I can watch the EA perform opening only one type of order and tweak each of the different order settings without confusing distractions. Now here's the really exciting part. I believe I've got the jump trades working pretty well, maybe 90% complete, so they see through all the noise confusion and skip the bad trades and grab the good ones. From end of august of last year to end of november, 3 months, with my standard 20% per trade, all wins, and a 3600% ROR!! And that's therotically with the EA performing at only 1/6 capacity! Pumped?? Good, so am I! lol
  • #99
  • Mar 1, 2007 3:27pm Mar 1, 2007 3:27pm
granted we did have that huge 1000 pip uptrend in the last half of November, but still, I think I'm making progress.
  • #100
  • Mar 2, 2007 1:13am Mar 2, 2007 1:13am
are you THE man who will do the perfect EA? ...I'm so curious!
very great work btw
1 trader viewing now