Great EA in backtest! - page 37

 
xxDavidxSxx:
you guys are a trip. You are making this way more complicated than it should be.

Here are all my back tests. from 80-90%.

I did run into a problem when starting from sept 1st 05. It dumped my account by the 9th.

Simple soulution.....read the manual......All I did is change "enable reverse detector" to true.Edit: reverse detector isn't in the manual. I just changed one thing at a time. Using sept 05 for the model time since it was the worse time for the ea. When one thing didn't work I put it back and changed another. Untill I got through sept 05 in profits. Then started the 1 years worth of back testing. I noticed alot of the down draws occured in news. But also alot of the big profits did too.

Now, if you look closely at all my results I have 1 year back test on both 4 hr and 1 hr. each one brought 2mil profits over the corce of the year. When it reached 680,000 the fxdd platform started rejecting the lot size. So I started over with 500$ from the last date of accepted trade.

Notice there was always a drawdown. Trade times were far apart. Sometimes it was in a drawdown for 2-3 months. Forex requires patiences. Lot of it.

But it always ended up in huge profits....

1hr charts produced more profits. On both 1 and 4hr charts january-may of 2006 was a bad time for the EA But a patient man would have still ended up with 600k+ eventually. And remember this was ran through news and all. 24/7.

I will open a new account with 500$ I have a seperate PC that I will set up for continous run of this program. I might even spend the 500$ for the pro version.

you wanted 90% model accuracy? well here it is.

edit: sry the 4hr charts are 79.66% one of the 1 hr charts is 75. But the rest of the 1 hr charts are 90%.

Dave

What version you test?

 
Oligarh:
What version you test?

I am forced to test the 1st version as the others won't trade on FXDD. But I am going to ask the admin at there site if there is a setting I can change so it will. I need the feature of adding time not to trade.

I am agressively backtesting different times of the last year. so far january is the worst for a drawdown. If I start with 500$ on the first the draw down is too great and lot size gets too small for it to trade.

It is getting raped on news because in january news was the only thing that moved the market. By the 19th it is down to 100$. I need it to skip trading 8am - 12 noon EST. I tried to copy paste the feature from other versions but didn't work. I probly missed a small symbol or something.

I know that back test isn't the same as real live trading but gives a very good simulation.

some months produce 100k easily but with my luck that won't be the time I first give it a real money try.

By the way...it made 400$ on the news this morning. wipsaw news is great for it but if price goes streight and never looks back it will experiance a draw down and if there is back to back news days it can draw down as much as 94%.

The paid for version makes adjustments all on its own.

Dave

 
furious_angel:
well, the 1_185f version i was running on 1 hour, but i did go back and re-read some of the first posts and i think you are to run the original version on 1 minute time frame. i could be wrong, so please double check some of the first posts.

1 min time didn't produce any profits on back test. neither did 5 min.

As you can see in my compairson the profits got smaller on each new version. Assuming you didn't change them and there both on default.

The first one is the only one that would run on FXDD with only 500$ starting.

My suggestion to all is use the first one, and on 1hr time frame and up.. But thats just me.

Dave

 

I just scanned through this thread all the way. Didn't do that befor...

FXSpeedster, Openstorm hasn't been on here in a while?

Help with getting this to run on FXDD please.

 
xxDavidxSxx:
I just scanned through this thread all the way. Didn't do that befor...

FXSpeedster, Openstorm hasn't been on here in a while?

Help with getting this to run on FXDD please.

OK i am an idiot. I got the newest version working for me. Won't say what it was to save further embarracement.

I am a technical trader not a programmer or developer. But I'll continue to get this thing figured out.

 
xxDavidxSxx:
I am forced to test the 1st version as the others won't trade on FXDD. But I am going to ask the admin at there site if there is a setting I can change so it will. I need the feature of adding time not to trade.

I am agressively backtesting different times of the last year. so far january is the worst for a drawdown. If I start with 500$ on the first the draw down is too great and lot size gets too small for it to trade.

It is getting raped on news because in january news was the only thing that moved the market. By the 19th it is down to 100$. I need it to skip trading 8am - 12 noon EST. I tried to copy paste the feature from other versions but didn't work. I probly missed a small symbol or something.

I know that back test isn't the same as real live trading but gives a very good simulation.

some months produce 100k easily but with my luck that won't be the time I first give it a real money try.

By the way...it made 400$ on the news this morning. wipsaw news is great for it but if price goes streight and never looks back it will experiance a draw down and if there is back to back news days it can draw down as much as 94%.

The paid for version makes adjustments all on its own.

Dave

Strange. At me an adviser did nothing. Not a single transaction. true I put him only in 9:00 GMT v1.85f.

 

CT newsflash

Hello to everyone,

I have been busy testing some enhancements to CT but still version 185f offerrs the best trading and profits. Simpler is better in this case.

The EA makes really good profit on M5 EUR and JPY. However to find a broker that will take the real action is impossible at least on M5. CrownFX, CFGTrader, FXDD do not want the business of this EA and if you start trading there on M5 they will shut you down after 24 hrs. So the only feasible option is to go to H1 which places a lot less trades. Time will tell if the brokers will allow it on H1 anyway. If the EA does not trade at whatever broker you have it set up, rename the EA to name of your choice, and reload it on the charts. It will be allowed to trade until (perhaps) the broker decides not to allow trading by it. I am now aware that Metatrader4 allows for ban of EAs with certain names. Doing this option will allow the EA to trade for a while.

One question I have: does anybody know a broker that has confirmed that they can clear CT at all and they will allow it? Otherwise having a good EA is of no use since you can only demo trade it. Please comment on this.

 

I find the response from the bugtrack people to be more about denial than responsive to actual bugs in the strategy tester. I agree that the way the data is processed appears to be unreliable. Using several different EA's now, the gogetter EA, the multiscalper EA, they both give differing reports over identical data sets and date ranges. So now they have a new, 'visual' function in the strategy tester...ok that's nice. now we can see better but that's not changing the instabilities of how it's processing data with exterpolations at all. It's more of a strategy teaser than a strategy tester. It's like a roulette game, will test the data this time or some cockeyed extrapolation of the data this time? It's anyone guess.

Using identical settings I see vastly different reports and all the bug track people say is that I must be using different data. Well, if there is differing data being used it's not because I tell the strategy tester to use different data. So maybe they are right and it is using different data....one must wonder why? Why does it decide to do this when it's not given any human direction to do so? What makes 'IT' think it should use different data? They never address this question. They dismiss it and then add a new 'visual' feature to the tester....

sure yea by all means send them tons of complaints...if you think it will do any good...

personally...I think they are all too drunk to care. (based on my experience with them)

That being the case...make the best of the crappy tool that you can but don't expect any real response or actual bugtracking from the support team.

They believe it's all just fine and as long as THEY believe so it's unlikely they will hear anything else from anyone, even hundreds of anyones. But who knows. Maybe there is a chapter of egotistical russian programmers anonymous or something that will find them and they will find recovery? Stranger things have happened.

It's been proven to me that they see what they want to see, hear what they want to hear, and believe what they want to believe. So be it.

 
fxspeedster:
Hello to everyone,

I have been busy testing some enhancements to CT but still version 185f offerrs the best trading and profits. Simpler is better in this case.

The EA makes really good profit on M5 EUR and JPY. However to find a broker that will take the real action is impossible at least on M5. CrownFX, CFGTrader, FXDD do not want the business of this EA and if you start trading there on M5 they will shut you down after 24 hrs. So the only feasible option is to go to H1 which places a lot less trades. Time will tell if the brokers will allow it on H1 anyway. If the EA does not trade at whatever broker you have it set up, rename the EA to name of your choice, and reload it on the charts. It will be allowed to trade until (perhaps) the broker decides not to allow trading by it. I am now aware that Metatrader4 allows for ban of EAs with certain names. Doing this option will allow the EA to trade for a while.

One question I have: does anybody know a broker that has confirmed that they can clear CT at all and they will allow it? Otherwise having a good EA is of no use since you can only demo trade it. Please comment on this.

how do you know fxdd will shut down this ea? Has it happened? and to who.

I need to know this soon befor I trade live. I am almost ready with it. I have spend the last week or so tweeking and testing 24/7 I have had like no sleep in the past week Probly invested the equevilant of 6 months time for some one who would work on it in evenings and weekends.

How do they know what your using? And how do I change the name?

I am working on 1hr time frame. I have it trading very conservitivly it will go days or even a week between trades. Thats fine because there is no "get rich quick holy grail" What you all have to remember is patience and don't look at the $ ammount of profits but the % increase of account. And keeping the max draw down to a minimum. I have testing at average of 63% accuracy. 11 pip s/l. And 35% max draw down. testing 2 years back, all together and also 1 month at a time. Taking the worst months and the worst times of the worst months to do my tweeking. "ITS ALL A MATER OF BALLANCE BETWEEN THE BEST MONTHS AND THE WORST"

I know that back testing is not live real money but it does serve as a great model.

Ignore the back tests I posted because I realized that there was 94% draw down on any given month. That was totaly unacceptable.

Dave

Please give me more info on FXDD and the statement you made.

Edit:Also, what peroid is the CCI indacater set to in this and how can I change it? is it 13? If so I have answered my own question.

 

here is what a max draw down month looks like for my testing.

38.17% draw down. but ending in 16% account profit.

do you recognise this draw down %?.....(38.2 fibo retracement) I have seen a couple of months like this in testing. I belive this tells me I am on the right track. Not done testing yet, I just thought this was interesting.

Dave

edit: I just noticed something. it was 83$ profit on 500$. Thats 16.6%.... thats only .2 away from 16.8 (.168, 1.68, 16.8, "golden ratio").....And I am a firm beliver in fibonacci and golden ratio being found in every thing in the universe. I make a living and so do many others based on these numbers.

I belive you should consider this in your developement of this EA. I have when tweeking the way its trading.

Once I have tested this on live money and have good results I'll let every one in on my settings. But untill its proven on real money I will wait.