THV MTF EA beta testing | Page 120 | 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

  • #2,381
  • Apr 28, 2010 11:03pm Apr 28, 2010 11:03pm
Hi Cobra team

I am presently doing forward tests on this EA. Could you please advise what the difference is between iMaxCross and FastTrixDirection. What does iMax check for? I understand that iMaxCross corresponds to the colour change on the fast trix line, but what criteria does it check?

Thx in anticipation

msood
  • #2,382
  • Apr 29, 2010 5:13am Apr 29, 2010 5:13am
Quoting rengel
After having run hundreds forward tests, some with a short success to turn into disaster and the rest all disastrous, I'm starting to believe that it will not be possible to create a working EA that will generate benefits with security.
What do you think??
Me too, I've been testing it for 3 weeks, I was having a few success in the beginning, then found it usually entered the trend too late. Then I made 4 different settings based on different method, each have 3 set-ups for m1, m5 and m15. They all turned out to be losing money somehow.
Only my most conservative set on m1 can make a slight profit overall.
I guess only Ricmi knows the trick.

But the testing is not totally a bad thing for me, I got inspiration from this system and came up with a new idea. I'm now working out the indicator.
  • #2,383
  • Apr 29, 2010 5:20am Apr 29, 2010 5:20am
Quoting msood
Hi Cobra team

I am presently doing forward tests on this EA. Could you please advise what the difference is between iMaxCross and FastTrixDirection. What does iMax check for? I understand that iMaxCross corresponds to the colour change on the fast trix line, but what criteria does it check?

Thx in anticipation

msood
iMax has nothing to deal with the Trix.
iMax is the two lines very close to the HA bars, (1 aqua 1 red by default)
Fast/Slow Trix Direction corresponds to the colour change of the trix lines.
You may not want to close orders based on them.
  • #2,384
  • Apr 29, 2010 6:01am Apr 29, 2010 6:01am
I've been testing it for 3 weeks. made 4 settings, each with specific settings for M1, M5 and M15. Results are not pleasing at all. Only following cobraforex's rule on M1 timeframe can make a slight profit over time. Manual live trading can overcome many noises which EAs can't.

Entry and exit based on Fast Trix direction can be deceptive since fast trix changes direction quite often, let alone iMax, which is extremely sensitive to price fluctuation. Slow trix is a good one but it responses too slow.
The pip difference between the high/low and the bar which slow trix direction change is confirmed can rise up to 20 pips or more, (for a general 50-80pip trend on M5, and you have to remember that you always entered late into the trend).

I've never used Ichimoku cloud because to me it's just forward-shifting the recent prices, it makes no indication on future price movement.

I always wonder why ricmi can achieve 0 losses. At least I know he's taking a very strict rule for entry. My best M15 setting can make 2-5 trades every day with slight overall loss. Comparing to ricmi's recent backtest result. You can see he did the backtest on M15, but there are many days where his setting didn't make a trade at all. Therefore I would say that he must be referring to a high timeframe, daily or even weekly. Moreover, he only set his t/p at 40 pips, and you see most of the trades were ended by s/l with very low profit. I can tell he isn't that optimistic even on trends based on high timeframes. If you are looking for a consistent winning setting, don't hurry yourself.
  • #2,385
  • Apr 29, 2010 6:41am Apr 29, 2010 6:41am
Hi lmyyks

Thx for that info.

You must be right about the EA because I don't see much activity/ buzz here as would be evident if folks were getting great results.

The thread on the manual system is more active.

Oh well, another EA bites the dust I suppose...

BTW, in your experience, have you found any EAs that work to give consistent profit > loss over time?

I have forward tested quite a few, and none could do this consistently. Some do a decent job on some time windows, but fail miserably at other times.

Nothing that I could bet my last dollar on....
  • #2,386
  • Apr 29, 2010 9:19am Apr 29, 2010 9:19am
Quoting msood
Hi lmyyks

Thx for that info.

You must be right about the EA because I don't see much activity/ buzz here as would be evident if folks were getting great results.

The thread on the manual system is more active.

Oh well, another EA bites the dust I suppose...

BTW, in your experience, have you found any EAs that work to give consistent profit > loss over time?

I have forward tested quite a few, and none could do this consistently. Some do a decent job on some time windows, but fail miserably at other times.

Nothing that I could bet my last dollar...
One problem I think is that the THV system enter the trend rather late, I always have loss trades because the trend reverses soon after an order is opened. And you won't dare to place orders too early because a trend hasn't been formed. Therefore I would say this EA is not that satisfactory in catching trend.

I tried the bombard v1.45 a few days ago, it was amazing
http://www.forexfactory.com/showthread.php?t=90649
however it only runs for demo account,
I'm not going to use that anyway because it magnified lots ridiculously, simply too dangerous for me.

And now i'm working on my own indicator, part of the idea actually came from my experience with the THV. I hope to catch the major trends while entering early enough so that s/l won't be hit that easily by noises. The primary model is ok so far.
Attached Image (click to enlarge)
Click to Enlarge Name: 123.jpg Size: 138 KB
  • #2,387
  • May 3, 2010 7:51am May 3, 2010 7:51am
I have a hunch that this EA could do better on higher time frames, i.e., H1, H4, daily as I can see that the trix are fairly good in giving an overall feel of the long-term trend. However, in such cases, the use of an EA may be an over kill because it only requires a person a few minutes per day to trade with such high frameworks, I guess it beats the purpose of having an EA in the first place because they are best for shorter term periods and for managing trades (a good example for trade managing EA is the popular 'Swiss Army' EA, which many manual traders use).
  • #2,388
  • May 4, 2010 3:13am May 4, 2010 3:13am
Quoting lmyyyks
One problem I think is that the THV system enter the trend rather late, I always have loss trades because the trend reverses soon after an order is opened. And you won't dare to place orders too early because a trend hasn't been formed. Therefore I would say this EA is not that satisfactory in catching trend.


That is the most stupid thing to say. Theres hundreds of different settings you can put in to EA options. You can change them to make the EA suit for your needs and catch what you call a "trend" as early or as late as you need to. Why dont you just change indicator settings rather than complaining how slow this EA is.

Im not saying this EA is perfect, but its the best EA i have ever seen for sure. I have seen hundreds of them.
The difficult part is that you need to work out good settings by yourself and it takes weeks and months time. If you want an EA which works without changing anything, buy yourself FapTurbo or whatever you like.
This however beats any commercial EA by the mile! Thats if you have patience to look inside it and work with it.

Ok enought of it.

Now few suggestions what that EA needs.

- Multiple orders option.
- Timeframe chosing option for Coral and possible for Kinko Hyo.
- Adding additional TMA indicator would be great.

Thats pretty much it. Adding those thoughts would surely make it superior.
Also maybe its possible to make it backtest a little faster. Atm its painfully slow even with high end PC running the backtest.



BIG THANKS to the makers!

K.
  • #2,389
  • May 4, 2010 5:25am May 4, 2010 5:25am
Quoting Supranium
That is the most stupid thing to say. Theres hundreds of different settings you can put in to EA options. You can change them to make the EA suit for your needs and catch what you call a "trend" as early or as late as you need to. Why dont you just change indicator settings rather than complaining how slow this EA is.

Im not saying this EA is perfect, but its the best EA i have ever seen for sure. I have seen hundreds of them.
The difficult part is that you need to work out good settings by yourself and it takes weeks and months time. If you...
Sorry I have to correct my comment, what I mean is that the original entry condition practised by cobraforex is not that good as a setting for this EA. As a new comer for forex, I truly knew nothing but breakout strategies. The THV is the first EA i've tested and what I've learnt from it is more than any other trading threads in this forum. The creaters have done a good job undoubtedly.

I've made a number of settings, one is, of course, cobraforex's setting. I've ran this setting for 3 weeks. However the overall result is not pleasing. I took screenshots for every trade for the first week and studied how it caught the wrong direction. Then modified it continually. Somehow great results were not consistent. I'm not the only one failed to produce a consistently profitable setting. You see ricmi is the only successful tester so far and everyone is asking for his settings.

I never questioned the success of the THV system, I tried it manually and it did bring me some demo money. However, when the original setting is put into this EA, it doesn't live up to expectation. I think everyone who tests this EA is because they believe in the system but not they like the the additional options of the EA itself. Those options only make it more complicated and harder for testers to figure out profitable settings. My best setting is not even THV because it is simply too different from the original setting, I'm just using this EA to run my own setting.
  • #2,390
  • May 5, 2010 2:36am May 5, 2010 2:36am
Hello all. My first post to this thread. Have put in a lot of effort trying to automate the THV system. Read this thread completely. Read enough of the original thread to get what I needed to know. Tested EAv1.0, and wrote my own EA with THV4 logic. After a months worth of effort, I have come to the conclusion that this EA will never be reliable for the simple reason that the THV3/4 indicator re-paints the last two bars of Trix information. This re-paint issue is addressed a little in the original thread, primarially by MissPips. But it is dismissed as "not a big problem". That might be the case when trading manually, but it will kill an EA. I have attached two pictures to illustrate the re-paint problem. In the pictures, the indicator at the bottom simply reflects a THV4 logic state. A green bar means that Trix is moving up, 7 out of 8 MTF_Trix dots are green, coral is honored, and two MTF_Cloud dots are green. A red bar means that Trix is moving down, 7 out of 8 MTF_Trix dots are red, coral is honored, and two MTF_Cloud dots are red. On the first picture, a sell position is opened after the close of a first red bar because THV4 primary conditions are believed to have been met. After the close of the second red bar, both red bars re-paint and become yellow, stating that THV4_Trix conditions are not met. The red bars re-paint because THV4_Trix has re-painted. The re-painting doesn't happen all the time, but it happens frequently enough produce frequent losers.

That is my conclusion. My only question mark is why Ricmi has not encountered this problem? Perhaps he has selected a small window of time during the European session when the market is moving and re-paint isn't an issue?? I would like to hear Ricmi thoughts on this. If I am wrong, I have come to like the taste of crow. I have made this, my only post to this thread, to try and caution those of you who are continuing to put effort into what, from a technical perspective, is a losing cause.
Attached Image(s) (click to enlarge)
Click to Enlarge Name: indi_signal_1.jpg Size: 83 KB
Click to Enlarge Name: indi_signal_2.jpg Size: 83 KB
  • #2,391
  • May 5, 2010 3:26am May 5, 2010 3:26am
Quoting EasyRyder
Hello all....
different brokers' server have different opening and closing time where prices will continue to change, there's a big gap for your monday open, but that is not necessarily apply to other brokers' mt4 platform.
  • #2,392
  • May 5, 2010 12:18pm May 5, 2010 12:18pm
The Monday gap has nothing to do with the fact that the two bars changed state The bars change state because THV3/4_Trix changes its value on the past two closed candle. This change of value is called "re-painting" and it is fatal in automated trading. Run a few hours of data slowly through a visual backtester so that you can see what is happening on successive ticks within a bar. After a bar closes, you will sometimes see its THV3/4_Trix state change. By this I mean that the value of Trix will change for a closed bar on one or more timeframes. When Trix changes on a closed timeframe, then MTF_Trix dots change colors and the trade you opened based upon the colors of the dots is no longer valid. And then you have a loser.

Many people are wondering why EA1.0 doesn't give consistent results -- I am suggesting why. Go to some expert EA programmers and ask them what they think about indicators that re-paint. I wish I were wrong, because so much effort has gone into this EA. I have spent a lot of my own time with this EA. At the moment, I am trying to find a trend/momentum indicator that does not re-paint to replace Trix. So far only marginal success.
  • #2,393
  • May 5, 2010 1:54pm May 5, 2010 1:54pm
In an attempt to clearify the re-paint issue, I have used the visual backtester to produce a repaint example using the THV4-Trix indicator.
I have attached two charts that you will recognize, showing THV4-Trix repainting. You will see that the 1:10 candle on the megenta verticle line repaints on the 1:15 candle. When the 1:15 candle opens, the 1:10 candle is green. When the 1:15 candle closes, the 1:10 candle is has changed to red. A repaint. When one is combining Trix signals on multiple timeframes, this re-painting can happen quite frequently. If an EA opens a trade long position at the opening of the 1:15 candle based upon the green 1:10 Trix signal (or combined MTF-Trix signal), the position becomes invalid when the Trix state of the 1:10 candle is repainted.

Trix is not a good indicator for EA work.
  • #2,394
  • May 5, 2010 1:56pm May 5, 2010 1:56pm
The pictures for my last message are attached.
Attached Image(s) (click to enlarge)
Click to Enlarge Name: Trix Repaint 1.jpg Size: 212 KB
Click to Enlarge Name: Trix Repaint 2.jpg Size: 191 KB
  • #2,395
  • May 5, 2010 3:23pm May 5, 2010 3:23pm
Quoting cobraforex
We still have no set files to present as this last version was finished today, I'm working on a 5M version which I will upload when finished.

The backtesting takes a lot of time as the EA is very CPU heavy as you might imagen.
After lurking for some time, this is my first time posting to this thread. I've attached 5 M1 set files. They are basic - (less is more) and they all work. All are the same except for the magic number. If you care to use these set files, be sure to check off "Only long positions" in the Common tab.
Attached File(s)
File Type: zip Set_Files.zip   7 KB | 504 downloads
If you're scared get a dog.
  • #2,396
  • May 5, 2010 6:09pm May 5, 2010 6:09pm
Quoting Boxxie
After lurking for some time, this is my first time posting to this thread. I've attached 5 M1 set files. They are basic - (less is more) and they all work. All are the same except for the magic number. If you care to use these set files, be sure to check off "Only long positions" in the Common tab.
AUDCHF performing well
If you're scared get a dog.
  • #2,397
  • May 5, 2010 7:05pm May 5, 2010 7:05pm
Quoting EasyRyder
The Monday gap has nothing to do with the fact that the two bars changed state...
I'm not actually talking about the repaint problem, but the opening gap. what i'm suggesting is that the trix gives false signals no matter it repaints or not because a part of the price action is not known by the server but the trix still does the calculation. Trix may not be a good indicator for you but more important is that you should avoid trading within certain period after the monday start because the gap affects the accuracy of all your indicators. Let's say if you are using a 5M chart and your are referring to trix of both periods 20 and 35, then you should set monday start at 0255 or later.
  • #2,398
  • May 6, 2010 2:31am May 6, 2010 2:31am
Quoting Supranium

Now few suggestions what that EA needs.

- Multiple orders option.
- Timeframe chosing option for Coral and possible for Kinko Hyo.
- Adding additional TMA indicator would be great.

Thats pretty much it. Adding those thoughts would surely make it superior.
Also maybe its possible to make it backtest a little faster. Atm its painfully slow even with high end PC running the backtest.

BIG THANKS to the makers!

K.
Hi Supranium you sound fairly experienced with the EA, have you reached some level of trading success with settings??
  • #2,399
  • May 6, 2010 2:40am May 6, 2010 2:40am
Quoting EasyRyder

Trix is not a good indicator for EA work.
I believe most traders familiar with thv would be aware that the trix repaints at least one bar, but having it affect 2 bars is concerning. all moving averages are limited by this flaw.

i find the best way to mitigate this and reduce probability of trix repaints is to use mtf trix. mtf direction alignment provides stronger chance of trend.
  • #2,400
  • May 6, 2010 4:40pm May 6, 2010 4:40pm
Quoting Boxxie
After lurking for some time, this is my first time posting to this thread. I've attached 5 M1 set files. They are basic - (less is more) and they all work. All are the same except for the magic number. If you care to use these set files, be sure to check off "Only long positions" in the Common tab.
It's 5:39pm EST now. I took long and shorts today for the five pairs and had an excellent day.
If you're scared get a dog.
3 traders viewing now