Forex Factory (https://www.forexfactory.com/forum.php)
-   Trading Systems (https://www.forexfactory.com/forumdisplay.php?f=71)
-   -   Negative And positive correlation divergence (https://www.forexfactory.com/showthread.php?t=911103)

dagoods Apr 29, 2019 12:33pm | Post# 41

I noticed one thing, the correlated pairs take turns in leading. Sometimes one closes above/below BB lines first sometime the other. In order to see which is leading which I think we have to go back and find a common ground where both are at the same line.

Yep. If you could predict which one will follow the other one, you could break the banks. The focus of these strategies should be predicting the timing.... But how? the million dollar question. You would have to start with a dashboard I'd imagine. .... or you could basket them somehow perhaps to take advantage of the market as a whole with multiple pairs diverging and converging.... if you guys figure it out remember who gave you the idea and send me a copy of the code!

PS have you tried non repainting TMA bands instead of bollinger bands? Curious if it would improve results or not.

hibra68 Apr 29, 2019 3:16pm | Post# 42

{quote} Yep. If you could predict which one will follow the other one, you could break the banks. The focus of these strategies should be predicting the timing.... But how? the million dollar question. You would have to start with a dashboard I'd imagine. .... or you could basket them somehow perhaps to take advantage of the market as a whole with multiple pairs diverging and converging.... if you guys figure it out remember who gave you the idea and send me a copy of the code! PS have you tried non repainting TMA bands instead of bollinger...
I dont know much about Tma bands but the idea is to show divergence of correlation, so if it shows that, than why not.
If for example positively correlated pairs both broke and closed above bb at the same time, this would be your starting point, later when one of them closes below low bb first and the other is lagging behind than you sell the lagging pair untill it touches lower bb.
now you have a leading and laggin pair. but roles might change so you have to keep track.

yalgaar Apr 29, 2019 3:28pm | Post# 43

{quote} If for example positively correlated pairs both broke and closed above bb at the same time, this would be your starting point, later when one of them closes below low bb first and the other is lagging behind than you sell the lagging pair untill it touches lower bb. now you have a leading and laggin pair. but roles might change so you have to keep track.
When one of them closes below low BB and the other one does not....it could also mean that the one that closed below was false break....and it may now come up back to where the perceived lagging one is! Just a thought!

hibra68 Apr 29, 2019 4:09pm | Post# 44

{quote} When one of them closes below low BB and the other one does not....it could also mean that the one that closed below was false break....and it may now come up back to where the perceived lagging one is! Just a thought!
true but you have to start somewhere and take a chance. the bollinger bands job is to keep price in, so the price bounces around it most of the time. Therefore if price is getting close to top or bottom bb, most of the time it will at least touch it.

dokopy Apr 29, 2019 6:20pm | Post# 45

{quote} Yep. If you could predict which one will follow the other one, you could break the banks. The focus of these strategies should be predicting the timing.... But how? the million dollar question. You would have to start with a dashboard I'd imagine. .... or you could basket them somehow perhaps to take advantage of the market as a whole with multiple pairs diverging and converging.... if you guys figure it out remember who gave you the idea and send me a copy of the code! PS have you tried non repainting TMA bands instead of bollinger...
It is possible to select 3 to 5 correlated pairs and trade a pair that differs by divergence.

hibra68 Apr 29, 2019 9:48pm | Post# 46

{quote} It is possible to select 3 to 5 correlated pairs and trade a pair that differs by divergence.
I wonder if its possible to connect to one of the correlation table's data via mt4 indicator to asses divergence?

Mataf has a correlation table on their website updated constantly so does Oanda, there might be others.

If the indicator can get data from the correlation table it can than track correlated pairs over 80% as an example, if one of the over 80% correlated pairs drops dramatically while the others remain intact it would give an alert stating that pair x has dropped to for example 65% correlation in relation to other pairs. It would also determine where the price of pair x is on the bollinger bands in comparison to other pairs . If for example all the pairs are near the top BB and pair x is near mid BB or bottom BB it would determine how far away it is from top BB and give us a target.

Further things could be added to this idea..

dokopy Apr 30, 2019 12:53am | Post# 47

1 Attachment(s)
{quote} I wonder if its possible to connect to one of the correlation table's data via mt4 indicator to asses divergence? Mataf has a correlation table on their website updated constantly so does Oanda, there might be others. If the indicator can get data from the correlation table it can than track correlated pairs over 80% as an example, if one of the over 80% correlated pairs drops dramatically while the others remain intact it would give an alert stating that pair x has dropped to for example 65% correlation in relation to other pairs. It would...
I just started it manually. The results are still good. I choose more than 80% for H1 and H4. On site: https://www.myfxbook.com/forex-market/correlation
Click to Enlarge

Name: example.png
Size: 132 KB

hibra68 May 1, 2019 7:25pm | Post# 48

2 Attachment(s)
Here is how these 2 pairs take turns leading when it comes to closing above/below top/bottom Bollinger bands

First 2 pictures.

Nzdjpy closes below bottom BB while gbpnzd only closed above mid BB. So we expect gbpnzd to close above top BB at least.
Click to Enlarge

Name: gbpnzd1a.png
Size: 106 KB
Click to Enlarge

Name: NZDJPYH1a.png
Size: 100 KB

hibra68 May 1, 2019 7:28pm | Post# 49

2 Attachment(s)
Here is how these 2 pairs take turns leading when it comes to closing above/below top/bottom Bollinger bands First 2 pictures. Nzdjpy closes below bottom BB while gbpnzd only closed above mid BB. So we expect gbpnzd to close above top BB at least. {image} {image}
The second set of pictures.

Now the Gbpnzd closed below bottom BB first and nzdjpy just closed above mid BB, now we expect nzdjpy to close above top BB
Click to Enlarge

Name: gbpnzd2.png
Size: 106 KB
Click to Enlarge

Name: nzdjpy2.png
Size: 100 KB

hibra68 May 1, 2019 7:31pm | Post# 50

2 Attachment(s)
{quote} The second set of pictures. Now the Gbpnzd closed below bottom BB first and nzdjpy just closed above mid BB, now we expect nzdjpy to close above top BB {image} {image}
In the last set.

Nzdjpy closes below bottom BB first while gbpnzd is between mid BB and top BB. this one is not quiet good a trade but there is still 18 pips between price and top BB and later on the price keeps going up.
Click to Enlarge

Name: gbpnzd3.png
Size: 102 KB
Click to Enlarge

Name: nzdjpy3.png
Size: 109 KB

hibra68 May 1, 2019 7:50pm | Post# 51

2 Attachment(s)
Here is 15 min

Nzdjpy closed below bottom BB while Gbpnzd is still near mid BB
Click to Enlarge

Name: gbpnzd15min.png
Size: 96 KB
Click to Enlarge

Name: nzdjpy15min.png
Size: 87 KB

hibra68 May 1, 2019 7:52pm | Post# 52

Dont forget guys closing above/below top/bottom BB shows MOMENTUM most of the time.

So price will continue most of the time.

morne72 May 2, 2019 6:09am | Post# 53

{quote} In the last set. Nzdjpy closes below bottom BB first while gbpnzd is between mid BB and top BB. this one is not quiet good a trade but there is still 18 pips between price and top BB and later on the price keeps going up. {image} {image}
Hibra68 ,was the trade taken when correlation was above 80% on both 1h and 4h ?

yalgaar May 2, 2019 6:55am | Post# 54

Hibra68-

I guess you have explained the entries clearly. I still had the following questions for clarifications:

1) Have you decided the exit of these trades? What is TP/SL once the trade is taken?
2) Which pairs are you suggesting for this trading?
3) Any visual testing done so far. Would be nice to see some results of visual testing?

hibra68 May 2, 2019 4:49pm | Post# 55

hibra68- i guess you have explained the entries clearly. I still had the following questions for clarifications: 1) have you decided the exit of these trades? What is tp/sl once the trade is taken? 2) which pairs are you suggesting for this trading? 3) any visual testing done so far. Would be nice to see some results of visual testing?
1) tp is touch of bb low or high depends on direction. Sl is equal to tp. Alternative is to bring sl to breakeven when it touches the outer bands than trail.
2) any pair that is correlated naturaly like usdcad/cadjpy. Or. Eruaud/audchf as long as they have a history of high correlation over a long period of time.
3) i tested nzdjpy/gbpnzd extensively but didnt record results because i didnt really have a tp or sl in mind at the time. But while i was testing i noticed the price reached outer bands and beyond almost all the time.

yalgaar May 2, 2019 7:03pm | Post# 56

{quote} 1) tp is touch of bb low or high depends on direction. Sl is equal to tp. Alternative is to bring sl to breakeven when it touches the outer bands than trail. 2) any pair that is correlated naturaly like usdcad/cadjpy. Or. Eruaud/audchf as long as they have a history of high correlation over a long period of time. 3) i tested nzdjpy/gbpnzd extensively but didnt record results because i didnt really have a tp or sl in mind at the time. But while i was testing i noticed the price reached outer bands and beyond almost all the time.
In your visual testing, how many trades are created approximately per week? Would be nice to do a visual test for at least 1 week and document all results following fixed set of rules. If that shows profitable, I am sure many coders will be interested to code it. What do you think? I am happy to do this for you if you can define all rules of trades.

hibra68 May 2, 2019 8:18pm | Post# 57

{quote} In your visual testing, how many trades are created approximately per week? Would be nice to do a visual test for at least 1 week and document all results following fixed set of rules. If that shows profitable, I am sure many coders will be interested to code it. What do you think? I am happy to do this for you if you can define all rules of trades.
I will getback to you.

hibra68 May 2, 2019 10:24pm | Post# 58

from 4/15 until today

gbpnzd vs nzd jpy 1hr


Start: 4/15 13.00 mutual start( means one has closed above the other has closed below BB.
net pips won /// max drawdown while in trade
7 4
5 2
12 12
7 14
62 10
total won 93 pips

SL= if price crosses and closes on the other side of BB. It never did during the 5 trades. But the maximum drawdown was 14 pips as you can see from above.

SL/ TP can be varied. i.e. You can use trailing because price can go far when it breaks above/below BB

I will test different pairs when I get chance. Especially positvely correlated pairs.

Mojotrades May 7, 2019 8:35am | Post# 59

Hello Hibra,

Is there a telegram group where discussions are taking place?
This strategy is an excellent one and one that i will be trialing extensively.

To confirm a few things, are you taking the data from the Mataf dashboard? (When 1H & 4H show 80+ or - %)
Thanks for sharing, and your continued support.
How do you go abouts of ensuring that you do not miss any trades when high correlation takes place?
(Please continue to post)

Thank you,

hibra68 May 7, 2019 2:34pm | Post# 60

Hello Hibra, Is there a telegram group where discussions are taking place? This strategy is an excellent one and one that i will be trialing extensively. To confirm a few things, are you taking the data from the Mataf dashboard? (When 1H & 4H show 80+ or - %) Thanks for sharing, and your continued support. How do you go abouts of ensuring that you do not miss any trades when high correlation takes place? (Please continue to post) Thank you,
Tnx for your post. I was hoping for someone to make an EA or indicator for this, but so far didnt happen. It is very hard for me to trade the system live because I work full time and the system requires constant checking, unless you use 4hr and above tf. But i like 1hr to 15m tf for this system.
No i dont have telegram channel for this.
Yes i use mataf for correlation.
But here is a little tip, if the pairs are correlated traditionaly and naturaly , you dont need to check mataf.
Tnx.


© Forex Factory