Supernova GBP/JPY Mini Trend Catcher | Page 238 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Forex Trend Catcher EA 721 replies

trend catcher fib ea 5 replies

Early Trend Catcher Part2 46 replies

Early Trend Catcher 31 replies

Trend Catcher 5 replies

Attachments: Supernova GBP/JPY Mini Trend Catcher
Exit Attachments

Supernova GBP/JPY Mini Trend Catcher

  • #4,741
  • Edited 2:22pm Jan 31, 2008 1:54pm | Edited 2:22pm
EDIT: Just nevermind what I wrote in this post before the edit.

Now my question is, how does CCI work for entries? Isn't it enough for it to be on the right side of the treshold but it must also have crossed it after last QQE cross?
I know that it has to for exits but I'm not so sure if it's so good for entries.

http://upl.vs-hs.com/i/8d.gif


Very late entry here to the right.
http://upl.vs-hs.com/i/491.gif
  • #4,742
  • Jan 31, 2008 3:06pm Jan 31, 2008 3:06pm
Rick,

From my own forward tests, i am getting many information to improve our EA...

I want to list my suggestions...

1. Multiple confirmations for exits. 2 of 4 criteria must be met for exit. Or 3 of 5 criteria must be met for exit.

2. Same for entries. 3 of 5 criterias must be met for valid entry.

3. Your money management strategies are inneed. We may take partial profits. I think this will increase our longterm profits dramatically.

4. We have to fix "trade context busy" trouble ASAP. The EA must try to send order again and again. Not wait for next candle to close.

5. I am trying to determine best trading times for each pair. We should be able to tell EA that "trade between these hours and do not trade between these hours" for each pair.

Once we have these 5 stuff in our EA, we will really run into very profitable system...

I advice you to get it ready before next week, of course if possible for you.

Good luck,
Enes
  • #4,743
  • Jan 31, 2008 4:02pm Jan 31, 2008 4:02pm
1. Not only shall we be able to take QQE cross signals before the candle has closed and the cross in confirmed. I think this should be possible for all indicators which it is possible with. Especially for exits. I've noticed when we get 1 long candle in the wrong direction and it waits for the close of this untill it closes the trade we have went into a loss. Sometimes trades are entered very late also because it waits for a very long to close that goes in our direction. Examples of closes:


http://upl.vs-hs.com/i/070.gif


2. A new idea is that we could choose to use different exit methods depending on how long we have been into a trade or how many pips it has moved. Maybe just a good MM would work better but at least here's another idea. This is getting advanced . DMI is very good to stay in trades for a long time.
http://upl.vs-hs.com/i/f8312aba1f4ba...80d0270cf3.gif

But, if the trade goes against us early, more reactive indicators like qqe or or cci would have gotten us out of the trade. Here the CCI or QQE would have gotten us out in a more timely manner in both long trades. At least if it could react before the bar closes.
http://upl.vs-hs.com/i/7d.gif


Or are there any indicators that will take care of all this for us?
  • #4,744
  • Jan 31, 2008 4:15pm Jan 31, 2008 4:15pm
Hey testers!!

I wanted to suggest that all testers and anyone who is interested join together in a chat room somewhere to bounce ideas off after each week.

Maybe just to talk about our findings and changes that we think will better the EA. This is so that all ideas are put together and we can get a better feedback of each persons test results.



What do you guys think?
  • #4,745
  • Jan 31, 2008 5:29pm Jan 31, 2008 5:29pm
Quoting Gnat
Hallo Rick and Tester!

My initial trades on GBPUSD and EURJPY are still there and are key to my entertainment. It seems that the Lag combo as we test it here always finds a good reason to stay in the trade. As you can see in the snapshot down there, the GBPUSD trade never came into winning territory and is by now quite far away from getting there. Still my only exit tool, the Laguerre is in good faith. Now as it seems it is closing in on the lines to maybe finally close the GBPUSD down.
Rick, could you tell me what you want me to do with that? It seems that this slow reactivity is good for slow moving pairs as stated yesterday, but naturally the Supernova rather calls for the more nimble pairs and here the Lag seems to be t o o s l o w hehehe!
Maybe Wolfster has a good idea how to hammer in more funk into the LaG and I will try to lure FX_Ninja over here, as he is in my eyes Mr. LaGuerre around here.

Best Wishes! :
Matt

I'm surprised to see this as I'm a firm believer in Laguerre. I love it because one can adjust it in so many ways, it will do just what you want- though it is becoming apparent that it may be more suitable for shorter time frames. (perhaps)

What bothers me here is that is didn't exit when it was supposed to, with the settings you have. I'm assuming they are the defaults, right? If that's the case then we need to be looking at why the ea didn't take the Laguerre signal. I did a blow-up of the time frame when yours should have closed; there are clearly three crosses of the line where it should have closed out. That make me wonder if your indis match the ea settings, or if maybe the ea is not reading it. Maybe a demo feed issue? naa, one of the crosses was good for at least two hours....
Attached Image (click to enlarge)
Click to Enlarge Name: gnat lag.gif Size: 41 KB
  • #4,746
  • Jan 31, 2008 5:36pm Jan 31, 2008 5:36pm
Quoting wolfeman
I'm surprised to see this as I'm a firm believer in Laguerre. I love it because one can adjust it in so many ways, it will do just what you want- though it is becoming apparent that it may be more suitable for shorter time frames. (perhaps)

What bothers me here is that is didn't exit when it was supposed to, with the settings you have. I'm assuming they are the defaults, right? If that's the case then we need to be looking at why the ea didn't take the Laguerre signal. I did a blow-up of the time frame when yours should have closed; there are clearly three crosses of the line where it should have closed out. That make me wonder if your indis match the ea settings, or if maybe the ea is not reading it. Maybe a demo feed issue? naa, one of the crosses was good for at least two hours....

Hey Wolfster!

I checked on the LaG on my chart and the settings of the EA - that should not be the problem. Well, maybe there is a problem with my demo ... I will look at it tomorrow as the bed is calling out loudly for me!

Thanks for looking into the issue - guess we will solve the problem tomorrow! Hope to catch you on early London! Any clue where Rick may be?

See ya!
GMatt
  • #4,747
  • Jan 31, 2008 8:49pm Jan 31, 2008 8:49pm
Hi Matt

First, Rick was flying back from Europe on Thursday afternoon.

As a couple of people have suggested, it'll be worth reviewing our settings for the week and noting any positive or negative aspects. I think almost all accounts have had a tough time, but that's good as it's identified some flaws.

Do you want to consolidate these reviews and then prioritize for the testing for the coming week? I think the QQE entry on the next bar needs to be addressed, as many have reported. Also, some have suggested QQE entry as soon as the cross occurs on a bar - I've not had a lot of time for watching charts this week, and I'm just wondering if that cross can or has happened several times on the same bar?

Should the individual testers post or send comments separately on their observations? What you think?
  • #4,748
  • Jan 31, 2008 10:31pm Jan 31, 2008 10:31pm
Quoting mamajoe
Hey testers!!

I wanted to suggest that all testers and anyone who is interested join together in a chat room somewhere to bounce ideas off after each week.

Maybe just to talk about our findings and changes that we think will better the EA. This is so that all ideas are put together and we can get a better feedback of each persons test results.



What do you guys think?
Well that is pretty much what this thread is doing. It is best if we keep all the communication in here as anyone interested in the Supernova will have all the communication in one place.

I found it extremely valuable to read the entire thread to follow the evolution of this baby and then I was able to get a lot of questions answered from all the question everyone else asked and got answered.

Now it is pretty easy to keep up by checking the thread on a regular basis.

Cheers
Ken Pipsqueaker "See you on the Pipside"
  • #4,749
  • Feb 1, 2008 1:42am Feb 1, 2008 1:42am
Quoting Pipsqueaker
Well that is pretty much what this thread is doing. It is best if we keep all the communication in here as anyone interested in the Supernova will have all the communication in one place.

I found it extremely valuable to read the entire thread to follow the evolution of this baby and then I was able to get a lot of questions answered from all the question everyone else asked and got answered.

Now it is pretty easy to keep up by checking the thread on a regular basis.

Cheers
Couldn't agree with you more Pipsqueak.
  • #4,750
  • Feb 1, 2008 8:24am Feb 1, 2008 8:24am
Quoting Gnat
Hallo Rick and Tester!

My initial trades on GBPUSD and EURJPY are still there and are key to my entertainment. It seems that the LaG combo as we test it here always finds a good reason to stay in the trade. As you can see in the snapshot down there, the GBPUSD trade never came into winning territory and is by now quite far away from getting there. Still my only exit tool, the LaGuerre is in good faith. Now as it seems it is closing in on the lines to maybe finally close the GBPUSD down.
Rick, could you tell me what you want me to do with that? It seems that this slow reactivity is good for slow moving pairs as stated yesterday, but naturally the Supernova rather calls for the more nimble pairs and here the LaG seems to be t o o s l o w hehehe!
Maybe Wolfster has a good idea how to hammer in more funk into the LaG and I will try to lure FX_Ninja over here, as he is in my eyes Mr. LaGuerre around here.

Best Wishes! :
Matt
Hi Matt,

Just getting back online after getting home from a business trip to Helsinki. I was writing a message about being offline when the hotel internet went off My card had run out and it would have cost a lot to get another weekly card for my last night there.

I spent several hours going back through the LaGuerre based charts and I think that what we are talking about as to using the LaGUerre differently makes sense. I am really tired from traveling, but will write up my findings when I can think again

Rick
  • #4,751
  • Feb 1, 2008 8:27am Feb 1, 2008 8:27am
Quoting Gnat
Hallo Rick!

It just had to happen: Now my EURJPY turned positive! Well, look at the chart - it seems that LaGuerre is making an equilibrium trade out of this.

Greetz!
Matt

Eidt: Looking at the single tests it is amazing that simplicity and "initial system components" seem to make the week. ( here: #4571 ) I am also amazed that my frustrating setup is in surplus while others I would have preferred have lost. Please have a look and let your ideas flow!

Yes, this is what we need to see. I really feel that all the frustration this week has been worth it. We have a much better understanding of the individual indicators which will allow us to tune the system intelligently rather than by guessing.

Really great analysis, Matt. Thanks!

Rick
  • #4,752
  • Feb 1, 2008 8:29am Feb 1, 2008 8:29am
Quoting Enes
Rick,

SuperNova Stats is doing really great job for me... It is showing me lots of useful results..

Thanks for it.

Enes
Hi Enes,

Great! Any recommendations for a next version would be greatly appreciated.

Rick
  • #4,753
  • Feb 1, 2008 8:48am Feb 1, 2008 8:48am
A new idea, hope you are not getting tired of my ideas
We should make it able to trade only in the direction of the longer trend. We should be able to choose type of moving average and the period for it. So if it's moving up only take long trades.
And a third option would be that the price has to be below the MA to go long when in uptrend and vice versa when in downtrend.
  • #4,754
  • Feb 1, 2008 8:51am Feb 1, 2008 8:51am
Quoting Enes
Rick,

From my own forward tests, i am getting many information to improve our EA...

I want to list my suggestions...

1. Multiple confirmations for exits. 2 of 4 criteria must be met for exit. Or 3 of 5 criteria must be met for exit.

2. Same for entries. 3 of 5 criterias must be met for valid entry.

3. Your money management strategies are inneed. We may take partial profits. I think this will increase our longterm profits dramatically.

4. We have to fix "trade context busy" trouble ASAP. The EA must try to send order again and again. Not wait for next candle to close.

5. I am trying to determine best trading times for each pair. We should be able to tell EA that "trade between these hours and do not trade between these hours" for each pair.

Once we have these 5 stuff in our EA, we will really run into very profitable system...

I advice you to get it ready before next week, of course if possible for you.

Good luck,
Enes
Hi Enes,

Over the next few weeks of testing I will be implementing some of the suggestions and collecting the others to propose to the whole group. That way we can all have a say on the priority for enhancements.

Your ideas are really helpful. Thanks!

Rick
  • #4,755
  • Feb 1, 2008 8:53am Feb 1, 2008 8:53am
Quoting SimonJ
1. Not only shall we be able to take QQE cross signals before the candle has closed and the cross in confirmed. I think this should be possible for all indicators which it is possible with. Especially for exits. I've noticed when we get 1 long candle in the wrong direction and it waits for the close of this untill it closes the trade we have went into a loss. Sometimes trades are entered very late also because it waits for a very long to close that goes in our direction. Examples of closes:


http://upl.vs-hs.com/i/070.gif


2. A new idea is that we could choose to use different exit methods depending on how long we have been into a trade or how many pips it has moved. Maybe just a good MM would work better but at least here's another idea. This is getting advanced . DMI is very good to stay in trades for a long time.
http://upl.vs-hs.com/i/f8312aba1f4ba...80d0270cf3.gif

But, if the trade goes against us early, more reactive indicators like qqe or or cci would have gotten us out of the trade. Here the CCI or QQE would have gotten us out in a more timely manner in both long trades. At least if it could react before the bar closes.
http://upl.vs-hs.com/i/7d.gif


Or are there any indicators that will take care of all this for us?
Yes, I agree wit your assessment on the reactive indicators. I have a test bed with QQE taking entries and exits at the moment of the cross, not waiting for the new bar. In SUperNova, I can change the EA to do this for any and all indicators with just a few minor changes.

Rick
  • #4,756
  • Feb 1, 2008 9:27am Feb 1, 2008 9:27am
Are we to close out all positions today or just let the testing run?
Ken Pipsqueaker "See you on the Pipside"
  • #4,757
  • Feb 1, 2008 9:35am Feb 1, 2008 9:35am
Quoting SimonJ
A new idea, hope you are not getting tired of my ideas
We should make it able to trade only in the direction of the longer trend. We should be able to choose type of moving average and the period for it. So if it's moving up only take long trades.
And a third option would be that the price has to be below the MA to go long when in uptrend and vice versa when in downtrend.
It already has the option to trade in the direction of the trend

As for the suggestion about the MA, that is exactly the filter I am testing on the reactive QQE and the results have been very promising. It does well filtering out fast moving markets and markets that are "too expensive". I am also testing an envelope of a %std deviation of the MA. One of these 2 methods will be inclused in the reactive QQE release.

Rick
  • #4,758
  • Feb 1, 2008 9:55am Feb 1, 2008 9:55am
Hi there,

How many lot does the preset will open once signal is generated?

I've tried to change mine, but it always open with 1.30 lots. Any ideas on this?

Also... The post that you showed us is some of the open trades. What happen when the trades are closed? Are most of them in profit?

Thanks..

Quoting SimonJ
Here you go!

I have also found even better settings for the EA, backtested and does profit for the whole of 2007 except december, GBP/JPY 30m, seems to work well for any pair on this timeframe, use on live account on your own risk. Much better than the settings I shared yesterday. Pretty funny is that whenever I use any stoploss it always makes less profit!
http://upl.vs-hs.com/info_imgs/56.jpg
  • #4,759
  • Feb 1, 2008 10:01am Feb 1, 2008 10:01am
Quoting rmyers
It already has the option to trade in the direction of the trend
Ah nice, does it have it in 1.7 beta6 or in your latest working version you havent released yet?
If it has it in 1.7 beta6 which setting is it? The only one I can think of is CheckConfirmingTF, which I'm using and it gives good results.
I think CheckConfirmingTF checks so that all indicators also agree on the ConfirmingTimefrime before entering a trade, correct? Not really like trading in the trend in the way I thought of it though.
Please enlighten

Btw I'm currently reading the MQL4 docs so that maybe I can help with the coding later on. I know javascript, php and c++ since earlier, but there's so many new functions and variables to learn here.

Thanks
  • #4,760
  • Feb 1, 2008 10:05am Feb 1, 2008 10:05am
Quoting mainsaham
Hi there,

How many lot does the preset will open once signal is generated?

I've tried to change mine, but it always open with 1.30 lots. Any ideas on this?

Also... The post that you showed us is some of the open trades. What happen when the trades are closed? Are most of them in profit?

Thanks..
I guess MM is on on 3%. So you have to turn that off.
I didn't let the EA close the trades but I looked back at the exit indicator and those that where in profit would have been closed in profit also.
The biggest problem with all settings I can come up with is that it opens too many trades that immediately goes into loss, because it opens them too late often, or when the market is ranging too much.
20 traders viewing now