THV MTF EA beta testing | Page 65 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

THV system, final edition 26,001 replies

MTF Stochastics into MTF Stoch Histogram 8 replies

THV ind 8 replies

My Journal - Based on THV V3.0 by Cobraforex 30 replies

THV RSI_V2 with Alarms --- need help! 6 replies

Attachments: THV MTF EA beta testing
Exit Attachments

THV MTF EA beta testing

  • #1,281
  • Jul 8, 2009 11:34am Jul 8, 2009 11:34am
Quoting kaizasoze
Has anyone able to close a few profitable trades using patterns so far? I have enabled pinbar, doubletopbottom and harmonics on M30 with previousbars =5. No trades so far. Lets c how it goes!
Kaz
can't help you there. but i think there's more to it than just enabling the settings and let it run. the settings have to be in perfect harmonie imo.
  • #1,282
  • Jul 8, 2009 11:38am Jul 8, 2009 11:38am
Quoting cobraforex
Hi Nick, good to see you.
Folks, Nick is one of the very first members of the THV EA Team.
Here in FF it's the first post but he's NOT a newbee and please treat him with respect as he's a helping hand always.
@Nick
Nice meeting you. Looking forward for your expertise.

AZBOfin
  • #1,283
  • Jul 8, 2009 11:54am Jul 8, 2009 11:54am
@ Cobra - thanks for the warm welcome back. For sure I can help out; however, it will take some time to familiarize myself with current version of EA - looks like the beasty has evolved somewhat.

Also not sure of Karun testing plans - what has been completed to date and what remain to be done.
  • #1,284
  • Jul 8, 2009 11:57am Jul 8, 2009 11:57am
Take your time Nick, we're aren't in any hurry.

Maybe you wanna drop an email to Karun, that way he can explain the advances we had lately.


Quoting getdown71
@ Cobra - thanks for the warm welcome back. For sure I can help out; however, it will take some time to familiarize myself with current version of EA - looks like the beasty has evolved somewhat.

Also not sure of Karun testing plans - what has been completed to date and what remain to be done.
  • #1,285
  • Jul 8, 2009 12:06pm Jul 8, 2009 12:06pm
Quoting AZBOfin
can't help you there. but i think there's more to it than just enabling the settings and let it run. the settings have to be in perfect harmonie imo.
sure thing mate..I asked Xux a while back and he said they play a support role, we cannot trade just by themselves. What I am trying hard to find is settings that enable them. No joy so far...but hunt is ON!

I was just asking in case any of our fellow testers have had any success with it, to get some insight.

Cheers
Kaz
  • #1,286
  • Jul 8, 2009 2:17pm Jul 8, 2009 2:17pm
Hi all, I've run some backtests on Cobra 30min 0.7v settings for year 2009. I don't know if it's any use to you because modelling quality is not very good.
Attached Image (click to enlarge)
Click to Enlarge Name: BackTest_2009_Cobra30min_v07.gif Size: 13 KB
Attached File(s)
File Type: zip BacktestCobra30min_V0_7.zip   46 KB | 437 downloads
  • #1,287
  • Jul 8, 2009 2:23pm Jul 8, 2009 2:23pm
@ AZBOfin Thanks for the welcome.

Looking forward to getting up to speed again on the latest version.
  • #1,288
  • Jul 8, 2009 2:25pm Jul 8, 2009 2:25pm
Thanks netropolis
  • #1,289
  • Jul 8, 2009 3:57pm Jul 8, 2009 3:57pm
Welcome Nick been testing the ea for a while now, and I am very happy with the progress.Hope I can also contribute to the testing.I am testing something this week that might be considered not to do with a real account due to the danger of drawdown, but will see how it goes and post more towards end of week.One thing that I am finding hard is to find a good exit point.Enter points seem to be not the problem even in manual trading, but the greed factor plus fear factor always worried me in when to place an exit or how much to let the trade breath?
Have a nice day everyone!
Also thanks Tom81 for the backtest
Vince
  • #1,290
  • Jul 8, 2009 4:01pm Jul 8, 2009 4:01pm
Hi Letizia - whats the largest drawdowns you have experienced to date ?
  • #1,291
  • Jul 8, 2009 4:12pm Jul 8, 2009 4:12pm
Quoting Tom81
Hi all, I've run some backtests on Cobra 30min 0.7v settings for year 2009. I don't know if it's any use to you because modelling quality is not very good.
pretty bad set ...

one thing I have noticed that kills my results is the LotMultiplier. I think it should be disabled. Chasing the losers was never a good idea
  • #1,292
  • Jul 8, 2009 4:40pm Jul 8, 2009 4:40pm
I have tried a kizzillion times with a tri-zillion settings to "optimize" the ea v6 and v7 both. I have never actually got one to go through. I can't even get good results on just a back test.

One, they take forever and two, there are bunch of mis-matched chart errors and 60% modeling quality. So after waiting 48 hours to process, I get zilch.

Is there something about this I'm not aware of? Thanks for any direction you can give me.
The Flying Dutchman
  • #1,293
  • Jul 8, 2009 4:53pm Jul 8, 2009 4:53pm
Quoting getdown71
Hi Letizia - whats the largest drawdowns you have experienced to date ?
I think I should just delete user Letizia and get a new one.I just don't want to start on FF as new.I am using username Letizia out of respect for my mother she died about 7 months ago.Anyway back to business.
What I am doing now is I have five sessions of eu open using the same set file, but using slightly different pip entry for the MinCrosslIchimoku I also have closeifprofit off and OppositeSignal exit on.My thinking was that the trade should close on oppositeSignal?
I am also using different risk from 1% to 5%
The set file is the same with all except with the changes I mentioned above.What I found strange with the setfile is that it has the pattern set and usually you have patterns turned on for higher tf.
As far as optimizing since there are so many options I find it hard to pick which one to optimize so I have spent constant hours trying to change different settings.Also I did not use the every tick option or it would take my computer 100 years (:
Attached Image (click to enlarge)
Click to Enlarge Name: eurusd.gif Size: 49 KB

Attached File(s)
File Type: doc eu1.set.doc   4 KB | 352 downloads

Vince
  • #1,294
  • Jul 8, 2009 5:03pm Jul 8, 2009 5:03pm
Have been running many backtests using lsweeney's 1hr set file. results are coming out very nice. One thing I did miss in his settings is that he turned martingale on. I just saw a nice big stop hit during the backtest, so looks like I will have to run it all over again with martingale off (I don't agree with using it).
I first ran every pair for the past week, saw some incredible results, so now I am rerunning the best ones for 3 months.
And yes, it does suck that it takes so damned long. Looks like it's going to take about half a day or more to run a single 3 month backtest.
Ah well, if the results come in as nice as I think they will, it'll be a small price to pay.
  • #1,295
  • Jul 8, 2009 5:08pm Jul 8, 2009 5:08pm
Quoting tazmaan
...it does suck that it takes so damned long. Looks like it's going to take about half a day or more to run a single 3 month backtest.
Ah well, if the results come in as nice as I think they will, it'll be a small price to pay.
It's nuts that it should take that long to run the data. I'm not sure what MT4 is doing, but computers should be able to process this data in a flash. I had Visible On and it took a 3 month backtest over 30 hours to process. That's just crazy.
The Flying Dutchman
  • #1,296
  • Jul 8, 2009 5:15pm Jul 8, 2009 5:15pm
Quoting letizia
I think I should just delete user Letizia and get a new one.I just don't want to start on FF as new.
Hi Vince
Write to the admin of FF. He should be able to roll you over to a new user account.

Good luck
AZBOfin
  • #1,297
  • Jul 8, 2009 5:20pm Jul 8, 2009 5:20pm
Quoting flyingnutjob
I have tried a kizzillion times with a tri-zillion settings to "optimize" the ea v6 and v7 both. I have never actually got one to go through. I can't even get good results on just a back test.

One, they take forever and two, there are bunch of mis-matched chart errors and 60% modeling quality. So after waiting 48 hours to process, I get zilch.

Is there something about this I'm not aware of? Thanks for any direction you can give me.
it all boils down to a good set of tick data from your broker. then you need to convert it to the other TF's. without this you never get 90%+ modeling quality. don't know if it's worth the effort though. personally, i don't believe in backtests.

AZBOfin
  • #1,298
  • Jul 8, 2009 5:29pm Jul 8, 2009 5:29pm
Quoting letizia
As far as optimizing since there are so many options I find it hard to pick which one to optimize so I have spent constant hours trying to change different settings.
imo this is one of the biggest problems during this phase of testing. you should stick to a specific setup for a longer period of time - don't change it just because you had a bad trade. by changing the setup constantly you will never get a real picture of the success of your strategy.
i would recommend to let the shorter TF's run for at least a week or two, then make some adjustments. longer TF's such as 1H and 4H you should let untouched for at least a month.

i know, it's easier said that done

AZBOfin
  • #1,299
  • Jul 8, 2009 5:32pm Jul 8, 2009 5:32pm
Quoting tazmaan
Have been running many backtests using lsweeney's 1hr set file....
you should send your tick data to the dutch man (flyingnutjob), he'll thank you for that
  • #1,300
  • Jul 8, 2009 5:49pm Jul 8, 2009 5:49pm
You just have to do some work prior to running any backtests.
I use fxdd, and I make sure to go to the history center and download all the 1hr data I can before running the 1hr backtests (download m30 data if you are running a m30 backtest etc).
Doesnt matter what broker you use (aside from ibfx maybe) as long as you download the entire history for your pair/timeframe ahead of time.
8 traders viewing now