- #1,902
- Dec 8, 2007 3:43pm Dec 8, 2007 3:43pm
- Joined Jun 2007 | Status: Unconventionable Conventionist | 2,768 Posts
After the "recess" maybe?
The mega compilation I did back then was the mega Excel thing I sent you and cryten a few months ago. I believe I had put each pair in a different colour etc.
- #1,903
- Dec 8, 2007 3:44pm Dec 8, 2007 3:44pm
- Joined Jun 2007 | Status: Unconventionable Conventionist | 2,768 Posts
- #1,905
- Dec 8, 2007 3:53pm Dec 8, 2007 3:53pm
- Joined Jun 2007 | Status: Unconventionable Conventionist | 2,768 Posts
http://i15.tinypic.com/8fvsy0y.jpg
LOL we're alll alone now.
Let's talk about the REAL SERIOUS & SECRET stuff while no one can hear
:
LOL we're alll alone now.
Let's talk about the REAL SERIOUS & SECRET stuff while no one can hear
:
Hehehe... I really just stumbled over this like: "How could I have missed that?" LOL
But yes -read it- -been there- -done it-!! http://www.jubailforums.com/vb35/ima...s/canadian.gif (you see I even have a Canadian now!)
Would you mind to set me up with some information by EMail later on? Only about how you see your exposure to work in the nearer future and whether there is something for the team to take off your immense shoulders and so on?)
We really only want to help, but dunno how and what and where! You just leave some instructions in the bee-cage for example and we will give it a go!
http://www.jubailforums.com/vb35/ima.../gathering.gif A group effort! So you can get rid of boss#2!! Hehehe!
Matter
Hehehe I kinda "answered" you with my questions in the previous post :surprised
- #1,908
- Dec 8, 2007 4:02pm Dec 8, 2007 4:02pm
- Joined Jun 2007 | Status: Unconventionable Conventionist | 2,768 Posts
ntagMt,
what have you been looking at / testing regarding the ea(s) lately? Any brilliant findings, genial lightbulbs, outstanding discovery?? (or even just an improvement of some sort?)
Well, I only found out that the CCI/PSAR link may not be working as described in the long conversation about firefighters you will remember and I contacted Herb about it. I guess the roboto may be good if this is ironed out and the settings are finished!
And what's cryten been up to?
Obiwan had his own little Waterloo, but Blücher came in time and he is very well indeed and all are in high spirits and ready to go. There was a bit of a crampy situation about the thread and all for some time until recently. Maybe you have read the restart post!?
Herbert and I have been emailing about bot bugs and he's working hard on trying to get everything ironed out (not in the main ea, in the CloseAll).
And I'm still testing............ of course
Herb needs direction on what to do with the "main" EA now and is absolutely felxibel and funny as ever. Sweetened our time with some hilarious ads he sent by EMail!
And what about Finn?
Finny has coded the LinearQ and is still developing that as a means to complement the Supernova QQEA and make it a trading tool for the manual side. It may also find its way into a EA later on! Very nice ideas there in the land of moose and Nokia! And we need cars as the guys on Top Gear! That is important for him as he drives a lot!!
Ersma!! ROTFL
Gnatler
- #1,909
- Dec 8, 2007 4:09pm Dec 8, 2007 4:09pm
- Joined Jun 2007 | Status: Unconventionable Conventionist | 2,768 Posts
Hehehe I kinda "answered" you with my questions in the previous post :surprised
Maybe my sensors are spoiled there a bit... http://www.jubailforums.com/vb35/ima...helpsmilie.gif
hihi..
Heheheh "DA BOSS NUMBER TWO", sounds like a scary movie where the plot is centered around murdering this guy wooooooooo scary.
Now, to answer your question, I will think of something which I will structure and systematize. Something like:
With such and such set-up in MT4 tester, test:
Pair/timeframe
Insert data for dmi
Insert data for mom
etc.
Now test SL from x to y and give me the total.
Sure this would accelerate things.
By the way, I did ask you, cryten and Herb to do exactly this last week or the week before on GBPJPY settings I had sent you. Only Herb actually carried out the test and sent me the results.......
Anyway, the problem with this is the data used and broker used. I need to have RELIABLE results (ie reproductible). Now depending on where you get your historical data and what broker you use (how they calculate what a lot is worth, for instance, and their spreads) will have an impact on the results and they won't necessarily correspond to what I would have gotten had I done it myself.
The only alternative I see at the moment, is that you copy what I do (maybe not on all pairs, let's start with one). This way, we could at least verify if we come close percentage wise to the same results and that would strengthen our view for a possibility of cooperation.
ya know what I mean there?
(I'm eating peeeedzah right now and it's quite good)
So, very simple, set your MT4 tester to TOOLS CHART no of bars 10000000, download the historical data for GBPJPY, do a quick test to see if it takes the trades from jan 1 to oct 27 (you need to set that up first of course) and check if you get 90% modelling quality.
Then checkmark OPTIMIZATION and start with the first parameter, in this case DMI from 1 to 160. Collect the results, put em in a speadsheet or anywhere you like. Then put DMI to (your best result) and proceed to MOM from 1 to 160 and so forth. You can also mix and match. I sometimes do TS before SL or TP before TS etc. But I first get the best DMI/MOM combination as a foundation. The best TF for GBPJPY is 4H.
Any questions?
EDIT: I forgot to include the "table":
DMI: 1-160
MOM: 1-160
SL: 40-150
TS: 40-125
TP: 50-140
Now, to answer your question, I will think of something which I will structure and systematize. Something like:
With such and such set-up in MT4 tester, test:
Pair/timeframe
Insert data for dmi
Insert data for mom
etc.
Now test SL from x to y and give me the total.
Sure this would accelerate things.
By the way, I did ask you, cryten and Herb to do exactly this last week or the week before on GBPJPY settings I had sent you. Only Herb actually carried out the test and sent me the results.......
Anyway, the problem with this is the data used and broker used. I need to have RELIABLE results (ie reproductible). Now depending on where you get your historical data and what broker you use (how they calculate what a lot is worth, for instance, and their spreads) will have an impact on the results and they won't necessarily correspond to what I would have gotten had I done it myself.
The only alternative I see at the moment, is that you copy what I do (maybe not on all pairs, let's start with one). This way, we could at least verify if we come close percentage wise to the same results and that would strengthen our view for a possibility of cooperation.
ya know what I mean there?
(I'm eating peeeedzah right now and it's quite good)
So, very simple, set your MT4 tester to TOOLS CHART no of bars 10000000, download the historical data for GBPJPY, do a quick test to see if it takes the trades from jan 1 to oct 27 (you need to set that up first of course) and check if you get 90% modelling quality.
Then checkmark OPTIMIZATION and start with the first parameter, in this case DMI from 1 to 160. Collect the results, put em in a speadsheet or anywhere you like. Then put DMI to (your best result) and proceed to MOM from 1 to 160 and so forth. You can also mix and match. I sometimes do TS before SL or TP before TS etc. But I first get the best DMI/MOM combination as a foundation. The best TF for GBPJPY is 4H.
Any questions?
EDIT: I forgot to include the "table":
DMI: 1-160
MOM: 1-160
SL: 40-150
TS: 40-125
TP: 50-140
- #1,913
- Dec 8, 2007 4:30pm Dec 8, 2007 4:30pm
- Joined Jun 2007 | Status: Unconventionable Conventionist | 2,768 Posts
Heheheh "DA BOSS NUMBER TWO", sounds like a scary movie where the plot is centered around murdering this guy wooooooooo scary.
Now, to answer your question, I will think of something which I will structure and systematize. Something like:
With such and such set-up in MT4 tester, test:
Pair/timeframe
Insert data for dmi
Insert data for mom
etc.
Now test SL from x to y and give me the total.
Sure this would accelerate things.
By the way, I did ask you, cryten and Herb to do exactly this last week or the week before on GBPJPY settings I had sent you. Only Herb actually carried out the test and sent me the results.......
Anyway, the problem with this is the data used and broker used. I need to have RELIABLE results (ie reproductible). Now depending on where you get your historical data and what broker you use (how they calculate what a lot is worth, for instance, and their spreads) will have an impact on the results and they won't necessarily correspond to what I would have gotten had I done it myself.
The only alternative I see at the moment, is that you copy what I do (maybe not on all pairs, let's start with one). This way, we could at least verify if we come close percentage wise to the same results and that would strengthen our view for a possibility of cooperation.
ya know what I mean there?
(I'm eating peeeedzah right now and it's quite good)
So, very simple, set your MT4 tester to TOOLS CHART no of bars 10000000, download the historical data for GBPJPY, do a quick test to see if it takes the trades from jan 1 to oct 27 (you need to set that up first of course) and check if you get 90% modelling quality.
Then checkmark OPTIMIZATION and start with the first parameter, in this case DMI from 1 to 160. Collect the results, put em in a speadsheet or anywhere you like. Then put DMI to (your best result) and proceed to MOM from 1 to 60 and so forth. You can also mix and match. I sometimes do TS before SL or TP before TS etc. But I first get the best DMI/MOM combination as a foundation. The best TF for GBPJPY is 4H.
Any questions?
Mail me some Pizza!! http://www.jubailforums.com/vb35/ima...lies/pizza.gif
LOL... then remember who told you about the bars and settings there
: (Sorry... hihihi)The results from testing will probably only deviate due to the spread between brokers as on MT4 you normally use the Metatrader data. If not, then possibly Alpari or IBFX-Tick-Data. Nevertheless the candles are nearly always identical. The brokers screw you on the longer TFs within the candles and not at the beginning or the end in order not to be caught that easily and as the re-check would be too easy.
So what may be convenient to do is to take your settings (as a file or Tableau as usual) and check them with our many brokers. Then optimization may run from there and might only lead to tiny deviations in the matrix of settings.
Furthermore the setting already found may be used to test the different settings for QQEA and, if fixed, for the CCI&PSAR combo, that may not be doing exactly what it was intended for.
Trendchaser pointed out, that as the EA is only taking trades at the end (=beginning of the next) candle, the modelling quality does not play a great role, as the opening and the closing price is not question there. So it may be a test to try and compare the modelling quality low to high. He once did that in the thread and showed exactly the same equity curve.
So with the settings we have we may easily optimize for the different brokers. Amazingly anyone in the group seems to be using another pirate broker! (And I am still looking for an ECN for our group and I guess I have one! WITH MT4... psssst!)
http://www.jubailforums.com/vb35/ima...lies/clown.gif Jaja ... deine Mudder!
Mats
- #1,914
- Dec 8, 2007 4:32pm Dec 8, 2007 4:32pm
- Joined Jun 2007 | Status: Unconventionable Conventionist | 2,768 Posts
I edited. Please back-up
- #1,916
- Dec 8, 2007 4:47pm Dec 8, 2007 4:47pm
- Joined Jun 2007 | Status: Unconventionable Conventionist | 2,768 Posts
Shame on me;-)
#1850
Well, I will try to find the actual post again too... but maybe not tonight! http://www.jubailforums.com/vb35/ima...lies/peace.gif
But I guees you can wrap your mind around the fact that candle opening and closing is contained in any set of data you may find on a tester especially above 5M data.
GnatFly
No problem at all... I just looked through the journal as I remembered the Trendchaser thingy being mentioned a few ... timesspans... or so before:
#1850
Well, I will try to find the actual post again too... but maybe not tonight! http://www.jubailforums.com/vb35/ima...lies/peace.gif
But I guees you can wrap your mind around the fact that candle opening and closing is contained in any set of data you may find on a tester especially above 5M data.
GnatFly
- #1,919
- Dec 8, 2007 4:59pm Dec 8, 2007 4:59pm
- Joined Jun 2007 | Status: Unconventionable Conventionist | 2,768 Posts
- #1,920
- Dec 8, 2007 5:01pm Dec 8, 2007 5:01pm
- Joined Jun 2007 | Status: Unconventionable Conventionist | 2,768 Posts
Yeah I did go search for that post when you first mentioned it (last week?) and never found trendchaser speaking about this specifically. You might have read it in another thread maybe?
