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)

hibra68 Apr 20, 2019 9:24am | Post# 1

Negative And positive correlation divergence
 
1 Attachment(s)
Hello!
I love using Bollinger Bands, so I keep thinking how to make use of it all the time.

I wanted to check out negatively correlated pairs and see how they react around top , mid and bottom bb.

Normally if one pair closes above top BB the other one should close at bottom bb and vice versa.

For example if you would look at EURUSD VS USDEUR than this would be true every single time, meaning that one would be above top BB and the other below bottom BB every single time.

So this would not help us traders at all.

Basically I am looking for divergence in the sense that when one pair is top BB the other one is at top BB as well. The closer they are to each other in price when it comes to BB lines the better.
I wanted to use 1hr tf so that I have enough signals and also enough meat to trade. I dont want 4 hr or more because of big SL. But you can try any tf you like.
The obvious choice was EURUSD VS USDCHF . However, I found that there wasnt enough pip value in 1hr. I mean its not bad but I found a better pair.

NZDJPY VS GBPNZD Feel free to find other pairs that work great.

Now for some reason I picked NZDJPY as the leading pair. I dont think this matters though, just go with the pair that breaks and closes top or bottom BB first.

So I have a chart attached with 2 trades on it. Normally its not easy to find but they are quality signals.
So in the chart you will see that NZDJPY breaks and closes top BB while GBPNZD is still above mid BB (since they are negatively correlated the GBPNZD should at least be close to bottom BB at this point. But its above mid BB so we take advantage of this) So we sell GBPNZD here hoping that it goes down to bottom BB like its suppose to.

Second trade NZDJPY CLOSES BOTTOM BB while GBPNZD is still below mid bb.

For stop loss in the case of GBPNZD I made a rule to use 60 pips as max(you can do what u want) and 30 pips min. So if the last low/high is bigger than max or smaller than min we go with max and min other wise we place sl at top/bottom of last high/low.

Take profit is 1:2 or 1:1 up to individual . I'm using 1:2 with breakeven at 1:1 with half taken off.

Other idea about TP is when the leading pair (NZDJPY) closes above, below or mid BB line we exit.

Let me know what you guys think and if you find any examples with other negatively correlated pairs please post your charts.

Thanks,
Click to Enlarge

Name: TurnkeyFX.png
Size: 130 KB

ChicagoRob Apr 20, 2019 11:09am | Post# 2

Very interesting. Can you post some more examples?

hibra68 Apr 20, 2019 1:17pm | Post# 3

Very interesting. Can you post some more examples?
i will when i have chance.

hibra68 Apr 20, 2019 3:04pm | Post# 4

1 Attachment(s)
Ok this is for opening tomorrow.

But this is just for show only because there is only 6 pips to top BB and as you know at sunday open the spreads are huge.

The SL at bottom of the low in this case the bottom of the candle is at our minimum around 32 pips.

The Tp would be 64 pips bring sl to breakeven when it goes 32 and take half out.

This is not exactly per our rules since the candle has closed above mid BB. But its very close to mid BB.
Click to Enlarge

Name: TurnkeyFX.png
Size: 103 KB

hibra68 Apr 20, 2019 3:42pm | Post# 5

1 Attachment(s)
Here is one more example
Click to Enlarge

Name: TurnkeyFX.png
Size: 113 KB

hibra68 Apr 20, 2019 3:46pm | Post# 6

1 Attachment(s)
This one failed, but the price went above mid BB with 35 pips.
We get stopped out at 32 pips but price only went against us for 36 pips.

So it was close.
Click to Enlarge

Name: TurnkeyFX.png
Size: 106 KB

hibra68 Apr 20, 2019 3:57pm | Post# 7

1 Attachment(s)
It works with EURUSD VS USDCHF

17 PIP SL 34 TP
Click to Enlarge

Name: TurnkeyFX.png
Size: 113 KB

hibra68 Apr 20, 2019 10:48pm | Post# 8

1 Attachment(s)
I changed the name to positive and negative correlation divergence because, why not?

In this example we have 2 positvely correlated pairs. I checked out for 1hr and 4hr correlation of over 80%. Below is a list of negatively correlated pairs. I only have for now chfjpy and nzdjpy as positive correlation.

usdcad - cadjpy

usdchf - eurusd gbpusd nzdusd chfjpy

audnzd - nzdusd

gbpnzd - nzdusd, nzdjpy

nzdcad - audnzd

nzdjpy + chfjpy
Click to Enlarge

Name: TurnkeyFX.png
Size: 121 KB

foff00 Apr 21, 2019 2:36am | Post# 9

Could you try gold/silver or usdjpy/ usdcad?
Is the concept if pairs correlating, sometimes correlation gets weaker, you expect resuming the same level of correlation?Peoblem is that sometimes does not.
Also, is on what bases you decide which pair you buy or sell?You do not know which pair will get stronger or weaker..
Regardless it is a nice idea, an easy way if proven to be working. A professional coder easily could make an EA. of it

ChicagoRob Apr 21, 2019 3:33am | Post# 10

This needs to be tested out. The problem with indicators is that they lag. However, from your examples, it appears that using correlation and divergence with BBs could be a leading indicator - you’re betting that one currency pair will “catch up” with the other pair and resume being “in sync.”

I’ve looked at a bunch of charts and have discovered these setups are far and few between, even on lower TFs. Trading several matching pairs would help a lot in this regard. For whatever reason, I find it a lot easier to work with positive correlation on a visual basis. It shouldn’t take too long to manually backtest a few pairs, to see how predictive the strategy is.

hibra68 Apr 21, 2019 9:04am | Post# 11

Could you try gold/silver or usdjpy/ usdcad? Is the concept if pairs correlating, sometimes correlation gets weaker, you expect resuming the same level of correlation?Peoblem is that sometimes does not. Also, is on what bases you decide which pair you buy or sell?You do not know which pair will get stronger or weaker.. Regardless it is a nice idea, an easy way if proven to be working. A professional coder easily could make an EA. of it
I have looked at "mataf currency correlation" table. I picked pairs that are negatively correlated more than 80% for 1hr and 4hr. You could pair up more currency pairs if you only look at 1hr or less if you look at 1hr 4hr daily. If you want even more pairs lower the correlation percentage, etc.

hibra68 Apr 21, 2019 9:08am | Post# 12

This needs to be tested out. The problem with indicators is that they lag. However, from your examples, it appears that using correlation and divergence with BBs could be a leading indicator - you’re betting that one currency pair will “catch up” with the other pair and resume being “in sync.” I’ve looked at a bunch of charts and have discovered these setups are far and few between, even on lower TFs. Trading several matching pairs would help a lot in this regard. For whatever reason, I find it a lot easier to work with positive correlation on a...
I agree, I have to take time and find more positively correlated pairs. There arent many over 80% for 1hr and 4hr. Maybe better to check out above 70% Correlation.
Anyhow I definetly think an Indicator is needed to troll different timeframes for different paired groups for signals.

MQLTrade Apr 21, 2019 9:24am | Post# 13

1 Attachment(s)
on 1H i see many good correlations
Click to Enlarge

Name: 1.png
Size: 46 KB

hibra68 Apr 21, 2019 9:24am | Post# 14

To get more signals I will look at mid BB to see if one pair has closed above/below it and the other is still far away.

MQLTrade Apr 21, 2019 9:36am | Post# 15

1 Attachment(s)
4h correlations
Click to Enlarge

Name: 1.png
Size: 43 KB

hibra68 Apr 21, 2019 9:47am | Post# 16

4h correlations {image}
Mataf correlation table shows different values.

For example Audjpy/Audchf is 91% on the 1hour in your chart

Mataf 1hr shows Audjpy/Audchf as -15% and 4hr shows 66.9%

I am looking for when both 1hour and 4hour are above 80% for the moment.

But if you want more signals just look at 1hr.

hibra68 Apr 21, 2019 9:49am | Post# 17

1 Attachment(s)
Here is an example of price crossing mid BB instead of top or bottom BB
Click to Enlarge

Name: TurnkeyFX.png
Size: 117 KB

MQLTrade Apr 21, 2019 9:54am | Post# 18

{quote} Mataf correlation table shows different values. For example Audjpy/Audchf is 91% on the 1hour in your chart Mataf 1hr shows Audjpy/Audchf as -15% and 4hr shows 66.9% I am looking for when both 1hour and 4hour are above 80% for the moment. But if you want more signals just look at 1hr.
Do u know how many bars Mataf use for correlation calculation? i Use 120 bars

hibra68 Apr 21, 2019 10:00am | Post# 19

{quote} Do u know how many bars Mataf use for correlation calculation? i Use 120 bars
It does not have that information.

but if you think 120 bars is sufficient than thats ok. I mean it wont hurt to try.

Look at pairs where both 1hr and 4hr are above 80% than if you dont have enough pairs lower the percentage to 70% and so on.

From your chart I see that usdchf and audchf have 90 % for both 1hr and 4hr its a good 1 to test out.

aelimian Apr 21, 2019 10:03am | Post# 20

Are you using the standard BB? Thanks


© Forex Factory