Backtesting this EA gives me problems for some reason. On a backtest going back to January of last year, it only shows one trade, the first one of the year. It only shows one trade no matter what period I start with.
I mean are you just looking for red above blue for longs and vice versa? Or are you reading anything into the angle, etc?
Ignored
no, it nowhere near that simple. It's complicated but if it follow this on the charts it should make a little sense.
To open a buy for a new trend:
1)the open of the 2nd to last candle must have been below fasst, which is right now 44.
2)the close of the last candle must have been above sloww, which right now is 79
3)there can be no spikes detected in the last 3 candles.
4)fasst(44) must be below sloww(79)
5)currency momentum(1,44) must be above 0
6)the currency momentum(1,44) three candles ago must have been below 0 OR the slow trend momentum(129,237) as of now must be greater then the slow trend momentum(129,237) three candles ago
7)the absolute difference between the ASK and sloww(79) must be less than 2* the difference between slow(79) and fastt(44) OR the absolute difference between the ASK and sloww(79) must be less than maxpipsfromEMA(26 as of now)
8)the absolute value of the slow trend momentum(129,237) must be greater than trending(14 as of now)
all 8 of these conditions must be met for a BUY for a new trend.
for a sell, they're just reversed.
Backtesting this EA gives me problems for some reason. On a backtest going back to January of last year, it only shows one trade, the first one of the year. It only shows one trade no matter what period I start with.
Ignored
it only works going back 4 months at a time. If you want to test it for the first four months of last year, you'd have to first optimize all the settings for the mood and characteristics of the market back then. And the system is just designed for the 1-hour period. The market changes too much every few months in speed, and size of waves, choppyness, that you can't have one set of settings that will give you good results for over 6-8 months. After 6-8 months if your system is strictly mechanical, your profts will go down and your drawdowns will go up. I've found that the last 3-4 months is the optimal period to optimize your EA on.
If you doubt what I say, you can optimize this EA for the last 4 months preciding any time you wish(it will take about 24 hours though, I'd recommend turning the stop setting off and just optimizing the general and spike settings first, and the stop settings after the others have been optimized) and you'll find once optimized it does very nicely for that period of time.
Hi aicccia, thanks for sharing this EA, I will forward test on demo and will let you know results. I have a problem with the EA, it gives me a sound alert everytime the GBPUSD moves a pip! It is definately the EA since the sound alerts stop if I take it off. Any thoughts?
Ignored
try disabling the alerts on the EA pop-up window...besides that I have no idea, it's only supposed to alert you once when it opens or closes a trade, and thats all it's done for me.
the EA is set up for a mini account with 1-200 leverage...so maybe it's trying to place a trade but can't because your account isn't a mini or 1-200 leverage. But I doubt thats the problem.
So what is most important here? For instance, on a long, you say to look for a candle that opens below the 48 and closes above the 72, correct?
What happens if a candle opens above the 48 but still closes above the 72? Is that a non trade? Looking at your long entry today at 9512, the candle opened about 1 pip higher (at least on my eSignal charts) than the 48 ema but you still entered the trade.
Could you just reduce this to using the 72 ema for similar/same results?
I had the 'one trade only' problem also. I started December 1st. and ran up to last week. The journal showed no errors. It took a quick trade for a profit and that was it.
So what is most important here? For instance, on a long, you say to look for a candle that opens below the 48 and closes above the 72, correct?
What happens if a candle opens above the 48 but still closes above the 72? Is that a non trade? Looking at your long entry today at 9512, the candle opened about 1 pip higher (at least on my eSignal charts) than the 48 ema but you still entered the trade.
Could you just reduce this to using the 72 ema for similar/same results?
Ignored
No for that exact reason it looks two candles back, so basically if the gap is crossed in two, not just one candle, then so far it's a trade.
I had the 'one trade only' problem also. I started December 1st. and ran up to last week. The journal showed no errors. It took a quick trade for a profit and that was it.
Ignored
December 1st 2006?
I don't know what to tell you except make sure you have 90% modeling quality for GBP-USD, make sure it's one hour period, and that these are the values you have for the parameters.
extern string General_Settings = "---------- General Settings";
extern int fastma=44;
extern int slowma=79;
extern int currency_momentum=43;
extern int trend_momentum=28;
extern int maxpipfromEMA=26;
extern int trending=14;
extern double risk_percent=0.10;
extern string Stop_Settings = "---------- Stop Settings";
extern bool use_stop=true;
extern double buy_maxhoursNeg=53;
extern double buy_maxpipsNeg=68;
extern double sell_maxhoursNeg=68;
extern double sell_maxpipsNeg=87;
extern string Spike_Settings = "---------- Spike Settings";
extern bool use_spike=true;
extern int SpikeHeight=71;
here is the most recent strategy tester report for the most recent EA which you guys have. For some reason I accidentally gave you the report for an older version of the EA. The main difference is the max drawdown has been cut from around 21% to 8%.
Well I found out 2 things at least on my platforms. It backtests on an IBFX mini account and gets several trades in January of 07. On MetaQuotes and Alpari it only takes 1 trade. Also Risk_Percent has to be set at .01 on a full size account. With a 10,000 balance and risk set at .1 it tries to open 20 lots.
Well I found out 2 things at least on my platforms. It backtests on an IBFX mini account and gets several trades in January of 07. On MetaQuotes and Alpari it only takes 1 trade. Also Risk_Percent has to be set at .01 on a full size account. With a 10,000 balance and risk set at .1 it tries to open 20 lots.
Ignored
yes....it is set up for a mino account with 1 to 200 leverage...I'll add options to change all that tonight. sorry.
it only works going back 4 months at a time. If you want to test it for the first four months of last year, you'd have to first optimize all the settings for the mood and characteristics of the market back then. And the system is just designed for the 1-hour period. The market changes too much every few months in speed, and size of waves, choppyness, that you can't have one set of settings that will give you good results for over 6-8 months. After 6-8 months if your system is strictly mechanical, your profts will go down and your drawdowns will go up. I've found that the last 3-4 months is the optimal period to optimize your EA on.
If you doubt what I say, you can optimize this EA for the last 4 months preciding any time you wish(it will take about 24 hours though, I'd recommend turning the stop setting off and just optimizing the general and spike settings first, and the stop settings after the others have been optimized) and you'll find once optimized it does very nicely for that period of time.
Ignored
I understand that it's optimized for the last few months. That isn't what I was referring to. Even if I start the backtest starting in January of this year, it shows one trade on January 11 to the 16th and that's it. If I started the backtest in December of 06, it only shows one trade on December 5th that's closed on the 11th. Obviously, more trades were generated than that.
By just eyeballing the chart, I believe one period of time where this system might have suffered large drawdowns was August of last year. Can someone optimize for April through July and use those results to backtest August? I would do it myself, but the backtester isn't working correctly. Trying to figure out what's wrong with it now.
I understand that it's optimized for the last few months. That isn't what I was referring to. Even if I start the backtest starting in January of this year, it shows one trade on January 11 to the 16th and that's it. If I started the backtest in December of 06, it only shows one trade on December 5th that's closed on the 11th. Obviously, more trades were generated than that.
Ignored
it's set up for a mini account with 1 to 200 leverage and I think thats causing the problem, if your account is different
Nicholas could you please tell me what indicators need to be set up on the charts with the moneymaker01 EA to get it to work properly? I was got one of your other posts on another thread and I had the idea that all that was need was the EA activated and the Osama indicator on a 30 min chart with candlesticks. Is this right? Thanks. Appreciate the effort and the thread.
by genetically optimizing them with metatrader but I'm wanting to find a way for the EA to find all the values on it's own, based off if the pair is trending or randing.
this system works great when GBP/USD is trending on a daily period...but if it starts ranging bad, like within 100 pips for a few days, and then down 200 and then range for another 100 pips for a few days, and if you do this for a month more like from February 2006 to july 2006, this system has problems. I was thinking ranging is just trending on a smaller scale, but during that period the price was doing some pretty crazy things. For example I found a 4 day period during april 2006 where the GBP/USD had a range of around 40 pips. And the candles were all much smaller than they are now. It was like a completely different currency. Nowadays we think a range of 100 is tight. So I'm trying to think of a way to make money out of price action like that, so I can have that mini-system going on in the background always making some change, while the main trending part makes the real money when the GBP/USD trends.
by genetically optimizing them with metatrader but I'm wanting to find a way for the EA to find all the values on it's own, based off if the pair is trending or randing.
Ignored
So you reoptimize every week using data from the past 4 months, or you reoptimize every 4 months?
