Great EA in backtest! - page 109
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
This morning is not nearly so fun. The demo took and won two positions. the live account took and lost one position...the slippage settings and defaults are the same on both accounts. That is not good.
This code is still kicking my butt too...I just don't know how to get it to do what I need it to do. This is a different approach to the same problem.
// These loops cycles TotalOMatches thru the OpenHistogramLevels looking for identical price levels and creates histogram of all unique open levels and their associated matches
int ct5=0,i5=NumberOfBars,level=NumberOfBars;
bool levelnotfound=false;
//for(ct5=NumberOfBars;ct5>0;ct5--)
{
// We loop thru the OpenHistogramLevels index looking for the new price level
while(Opens) // We loop thru the new prices
{
if(i5==SIZE) {continue;}//moves past the array zero index
while(OpenHistogramLevels[level]) // We loop thru the Histogram Index
{
if(level==SIZE) {continue;}//moves past the array zero index
// We augment the match values of each level to reflect the highest # of matches
if(OpenHistogramLevels[level] == Opens)//identifies matching price value in OpenHistogramLevels
{
// If we ARE working with a price level which is already in the histogram we...
if(OpenHistogramMatches[level] < TotalOMatches)//compares matches at this level
{
OpenHistogramMatches[level] = TotalOMatches;//increase match value if new value is larger
}
}
else
{
levelnotfound=true;
}
level--;
if(level>SIZE*(-1) || level<0)//this only allows the loop to cycle from 0 to array "SIZE"
{
break;
}
}
if(levelnotfound)// We append the new level to the histogram
{
OpenHistogramLevels=Opens;
levelnotfound=false;
//Print("Levels: ",OpenHistogramLevels," Matches: ",OpenHistogramMatches);
}
i5--;
if(i5>SIZE*(-1) || i5<0)//this only allows the loop to cycle from 0 to array "SIZE"
{
break;
}
}
// Print("Levels: ",OpenHistogramLevels," Matches: ",OpenHistogramMatches);
}This morning is not nearly so fun. The demo took and won two positions. the live account took and lost one position...the slippage settings and defaults are the same on both accounts. That is not good.
I hear you Aragorn. I find myself in the same boat. I opened two more demo account past sunday, just before the market opened. One at FXDD, the other at IBFX. Both accounts have identical settings down to the last byte.
As of today, the FXDD account is down and the IBFX is slightly up.
My mini live with IBFX traded differently as well.
I ask myself why doing all the work with forward/backward testing, tweaking code etc when at the end the live account acts completely different that the demos.
AZBOfin
I hear you Aragorn. I find myself in the same boat. I opened two more demo account past sunday, just before the market opened. One at FXDD, the other at IBFX. Both accounts have identical settings down to the last byte.
As of today, the FXDD account is down and the IBFX is slightly up.
My mini live with IBFX traded differently as well.
I ask myself why doing all the work with forward/backward testing, tweaking code etc when at the end the live account acts completely different that the demos.
AZBOfinI suppose we do it because the solution is in the approximation values. We really have no other choice but to approximate as best we can and use what we have to do the job. It's an inprecise messy job at best, at least the best I've been able to do is that way so far.
for anyone following my support resistance development I just explained it to someone else on another thread..maybe this will help you understand my goals with the project.
https://www.mql5.com/en/forum/175257/page17
I just reduced the risk to half of what it was and darn if it didn't just win a position while my back was turned and put nearly all the loss of the last trade back in the account..if I had left it alone I would have won the other half...
the more I run this the more I get the message don't interfere with it, let it go...I have a hard time letting go of control. But almost every time I get in the way of this EA it's a mistake on my part and costs me.
ok the last two days has been totally crappy on the euro. Has anyone else had a streak of losers like I have recently? I have got to get the support resistance filter going.
ok the last two days has been totally crappy on the euro. Has anyone else had a streak of losers like I have recently? I have got to get the support resistance filter going.
i'll do some damage control over the weekend, but you're right, it doesn't look very good. since i started with your euro-version this past tuesday, i'm down -23 pips with 4 winning and 3 loosing trades. ouch.
AZBOfin
PS: traded on IBFX live mini account with standard settings
i'll do some damage control over the weekend, but you're right, it doesn't look very good. since i started with your euro-version this past tuesday, i'm down -23 pips with 4 winning and 3 loosing trades. ouch.
AZBOfin
PS: traded on IBFX live mini account with standard settingsI just don't know how to interpret these things. I've watched the demo account take a hit too but it has won more trades than the live acct. I don't know how to trust a system that makes just %72 win/loss..seems like I don't know when that nasty 28% is going to hit and I want a system I can trust enough to leverage. I end up leveraging just in time to hit the skids with this.
What I want to know and still don't know is what kills this system. What market conditions undermine it. I think it's time to scrutinize the backtest more closely on those losing times and figure out what it was about the market when it was losing that made it lose and what about the market makes it win. Until we can really answer those fundamental questions about this ea it's going to be realy hard to design an appropriate filter. We've just got to know more about what works and doesn't work for this system.
my support an resistance project is progressing and I've found someone with more experience programming who'se willing to assist me now. It's all good. Until then keep the lot sizes small and the risks down if you decide to let it go live. That's all I can think right now. That is unless someone else wants to post a statement showing otherwise. All is not lost but then neither are all hopes realized. Looks like time for more persistence.
I tried to purchase CT, cannot get them to respond to my email as I had a question about thier purchase process, no customer service means no sale, I'd buy it in a heartbeat if I could get a response...
ok see this?
These are all losing trades. They occur within the same hour. These occur when the CCI allows the trades and the logic says a reverse is here so it enters.
I've watched this occur in live trading and demo. It's not just backtester that does this.
I suggest for starters we make a subroutein that only allows one trade per bar for starters. The program's logic isn't built for more than one trade per bar and obviously it takes some big drawdowns because of it.
I'm not exactly sure how to do that but that's a good starting point. How about it any other developers have a handy function that will only allow one trade per bar?
ok see this?
These are all losing trades. They occur within the same hour. These occur when the CCI allows the trades and the logic says a reverse is here so it enters.
I've watched this occur in live trading and demo. It's not just backtester that does this.
I suggest for starters we make a subroutein that only allows one trade per bar for starters. The program's logic isn't built for more than one trade per bar and obviously it takes some big drawdowns because of it.
I'm not exactly sure how to do that but that's a good starting point. How about it any other developers have a handy function that will only allow one trade per bar?thats a great idea. I took 5 losses in a row on a bar just like that. Not even news.
You can try to set cci to only sell below 0 and only buy above 0. You can try taking out that range between 50 and -50 where it can buy and sell.
but 1 order per bar is brilliant
this weeks observations
ok see this?
These are all losing trades. They occur within the same hour. These occur when the CCI allows the trades and the logic says a reverse is here so it enters.
I've watched this occur in live trading and demo. It's not just backtester that does this.
I suggest for starters we make a subroutein that only allows one trade per bar for starters. The program's logic isn't built for more than one trade per bar and obviously it takes some big drawdowns because of it.
I'm not exactly sure how to do that but that's a good starting point. How about it any other developers have a handy function that will only allow one trade per bar?I also have noticed E.A does not like to trade more than once in the hour creating heavy loss when it does.This e.a man, it kicks the proverbial...... but not the way the programmer intended. I have found that trading reverse is profitable not only during times disabled but also during normal hours. The relationship between t.p and s/l makes it so. I have experienced situation this week and last where the E.A will win eight trades in a row and then suffer two losses (in my case wins to claw it back). Another small observation i found which supports a statement David made a while ago concerns the t.p cyberia aims for. In the beginning CYBERIA (no mm and fixed lots) appears to aim for a t.p of exclusively. After a week the t.p went up to 10 and 11 per trade. Not only that but it,s success rate improved markedly. I don,t know if this is possible but if you could have a function where the E.A for the first week went through its calculations but did not actually place the trades, and then placed trades from the second week onwards (or when it leapt to 10 tp). I think for those of you who have yet to run E.A in live account you would see a big difference (less drawdown if any at all).
My standard settings using davids e.a on usdjpy is at $193 profit. Start date October 16 (72% profitable trades). i ,m really tempted to put $5k in this just to satisfy my curiosity and get things moving. But my common sense is beating back my nonsense.For now...
reverse decision trades for the week $1959.24 profit (53.8% profitability). On tuesday alone E.A took $600. Biggest drawdown i have seen is $179 (8 trades in a row). NFP should be intersting today hopfully with volatility will achieve $2.5k
I,m taking some of the profit to get reverse decision in code. I offer first refusal to Aragorn since he has done a most admirable job.