Hedge and Correlation Strategy | Page 111 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Correlation + Momentum Strategy? 6 replies

MT4 MTF Correlation table and Correlation oscillator indicators 155 replies

InsomiaFX Correlation Double Hedge EA 66 replies

Looking for EA for Correlation Strategy 6 replies

Hedge strategy (not a hedge fund) 13 replies

Attachments: Hedge and Correlation Strategy
Exit Attachments

Hedge and Correlation Strategy

  • #2,201
  • Feb 3, 2011 12:47pm Feb 3, 2011 12:47pm
Thank you! Just as I suspected, our internet is having problems and giving false data. Mine was 94! Yikes!
  • #2,202
  • Feb 3, 2011 12:49pm Feb 3, 2011 12:49pm
Quoting Dreamliner
Thank you! Just as I suspected, our internet is having problems and giving false data. Mine was 94! Yikes!
At 94 I would dive in Ha ha.. but I think you should wait the 34 looks correct.
Over confidence can be a moral hazard.... not to mention a financial one!
  • #2,203
  • Feb 3, 2011 1:21pm Feb 3, 2011 1:21pm
Quoting roundrock
pl take this EA, i fixed a bug on digits issue with some pairs
Roundrock - I used your EA to garner some nice profits. After your update the behavior started to change. The EA will open the trades as before but close them right away. It does this (it appears) until the diff drops below 80.

If you have not seen this is there some info I could gather to help issolate the issue?

I'd love to have this running again.

Best
T
  • #2,204
  • Feb 3, 2011 1:54pm Feb 3, 2011 1:54pm
Quoting Kruspe
Iīm not sure if is someone considering a lot size (lot ratio between open pairs).. maybe I didnīt noticed it here, thread is going really fast (itīs great).. anyway..

I think, itīs not good to open trade with the same lot size (every pair has a different average daily volatility)... if you trade this pairs with the same lot size it could happend, that the slower pair canīt "meet" the faster one.. pairs with small volatility have no chance against pairs with higher volatility.
For example 1.pair has 50pips daily range...
Great suggestion Kruspe, I mentioned this a couple of days ago when I noticed I had losses that involved pairs with big ATR differentials and was trying to work out a solution based on the ATR, but wasn't really getting anywhere with something easy to execute. Your idea is excellent, very easy to use on the fly. The simplest solutions are usually the best.....
  • #2,205
  • Feb 3, 2011 2:22pm Feb 3, 2011 2:22pm
Quoting Dreamliner
My son keeps assuring me, just wait dad, they will come back into correlation...
He is right.

  • #2,206
  • Feb 3, 2011 3:17pm Feb 3, 2011 3:17pm
Quoting Dreamliner
Wow man you're a pipster, a pipmeister! Is your favorite pizza PIParoni?

Anyway, you're doing all this the day before NFP? All my stochs have gone home and gone to bed for the day because of NFP.
Hi DL

So are you suspending trading around NFP or keep it going and jump into the gaps?

Is this strategy not exploiting the temporary volatility of such events or is this too much of a risk?

Thanks
  • #2,207
  • Feb 3, 2011 3:41pm Feb 3, 2011 3:41pm
I'm definitely doing NFP (a few minutes after); it's just that everything seems to go to sleep the day before, in anticipation I expect.

Quoting Zr-fx
Hi DL

So are you suspending trading around NFP or keep it going and jump into the gaps?

Is this strategy not exploiting the temporary volatility of such events or is this too much of a risk?

Thanks
  • #2,208
  • Feb 3, 2011 4:34pm Feb 3, 2011 4:34pm
The essence of a pm I received a few minutes ago is this:
"I was wondering if it would be possible to change your Going Home Sending Script to allow the entry of a different lot size for each pair?"

The attached allows for this.

Attached File(s)
File Type: mq4 Going Home trade sending script.mq4   2 KB | 613 downloads
  • #2,209
  • Feb 3, 2011 5:22pm Feb 3, 2011 5:22pm
Quoting SteveHopwood
The essence of a pm I received a few minutes ago is this:
"I was wondering if it would be possible to change your Going Home Sending Script to allow the entry of a different lot size for each pair?"

The attached allows for this.

Thanks alot Steve. That allows us to use Kruspe suggestion to enter different lot sizes for pairs which have different ATRs, pip values. I also find that this script ensures that the trades are entered simultaneously, so they are hedged straight away whereas entering each trade manually does have a delay which can end in one pair moving before the other is entered. Great
  • #2,210
  • Feb 3, 2011 8:17pm Feb 3, 2011 8:17pm
Hey just wondering what brokers are you guys using for the different lot size trades?
Because the smallest lot size my broker allows is 0.1 so i am unable to test how effective this is.
  • #2,211
  • Edited Feb 4, 2011 12:24am Feb 3, 2011 9:33pm | Edited Feb 4, 2011 12:24am
Quoting Kruspe
Iīm not sure if is someone considering a lot size (lot ratio between open pairs).. maybe I didnīt noticed it here, thread is going really fast (itīs great).. anyway..

....So, when you open (buy USD/JPY and sell GBP/JPY) and both pairs will go in the same direction (because of correlation) itīs good (you hit the direction of one pair- itīs ok), but there is a problem because GBP/JPY is round 1.6 times faster than USD/JPY. So you have to parlay 1.6x time lot size of USD/JPY....
Hi Kruspe. Where did you get your figure of 1.6x for UJ/GJ from?

I've been thinking about this a bit more. GJ has an average daily range of 151 pips (per my chart indi) at the moment and UJ is 70 pips. So the UJ needs to move 2.1 times as far as GJ to keep theoretically the combined value of trades equal when a news event happens (after a 60 pip spike yesterday thats a risk I want to control using triples), and to equalise the problem of one pair moving faster than another.

Comparing pair prices doesn't always work eg GBPCAD has range of 135 pips and a price of 1.6, GBPJPY range of 151 and a price of 131.6 or 1.31 using Kruspe's method. In this case the lower range pair has a higher price, so a different calc is needed.

SMJ is going to hate me for this .... Say we are taking a GJ/UJ triple trade. So if we wish to risk 1 one lot we do the following: Average Daily Range GJ / ADR UJ (150 pips/70 pips) = gives a ratio of 1:2.1 .So the GJ lot size = 1/3.1 (1+2.1) = .32 lots and the UJ = 2.1/3.1 = .67 lots
I've tried to do this manually at entry time, but sometimes by the time I dig up the ADRs and do the calc, the stoch diff has dropped back below 90, and the trade is gone. Or make I make a calc error (fat fingers or I get an attack of the stupids).

But if Tradable Correlations indicator showed the calculated lot size for the pairs as above it would speed things up alot. Even better would be an option to enter a user defined lot size that the indicator would split between the pairs and display and then one could just enter these into SteveHopwood's Going Home Sending Script. Now that would be super (if there aren't any flaws in my thinking).

Really sorry to suggest this one SMJ . I've no idea how much work is involved in this (sounds alot with an ATR calc) so if it's too much please don't worry about it, it's no problem at all. I really appreciate the work you've done to date.

[Update: Ckp has said in post 2213 that he has found this doesn't work as while you might have daily ADR for calculation purposes there are fluctuations at a lower timeframes that negate any positves. His experience is that even lots work ok....so no need to bother with this SMJ]
  • #2,212
  • Feb 3, 2011 11:28pm Feb 3, 2011 11:28pm
  •  Ckp
  • | Joined Jul 2010 | Status: Master ur EXIT. Entry is secondary. | 29 Posts
Quoting smjones
Well that is one way to do it, but a better way would be to set the timeframe to days and select 0 as the periods, that way it would look at every closing daily bar you have in your history.

Or you could change the max history bars in the tools section of mt4. In any event by putting a 0 in the periods field, the program looks at all of the bars
Thanks SMJ

I would like to better understand how the logic of your correlation co-efficient works. I'm refining some correlation strategies which involve a whole lot of number crunching. Conceptually I know what im looking for but practically i don't know how it is calculated. Could you suggest where i might go for further reading material. I would like to educate myself on this. Any pointers would be useful. Thank you.
  • #2,213
  • Feb 3, 2011 11:59pm Feb 3, 2011 11:59pm
  •  Ckp
  • | Joined Jul 2010 | Status: Master ur EXIT. Entry is secondary. | 29 Posts
Quoting occamsrazors
Hi Kruspe. Where did you get your figure of 1.6x for UJ/GJ from?

I've been thinking about this a bit more. GJ has an average daily range of 151 pips (per my chart indi) at the moment and UJ is 70 pips. So the UJ needs to move 2.1 times as far as GJ to keep theoretically the combined value of trades equal when a news event happens (after a 60 pip spike yesterday thats a risk I want to control using triples), and to equalise the problem of one pair moving faster than another.

Comparing pair prices doesn't always work eg GBPCAD has range of 135...

Before you go any further think this through very carefully. I have previously experimented with entering lots at a percentage equal to the daily range of a pair. I abandoned the idea pretty fast. It sounds like a good plan but it's not really. Here's why;

Assuming you use ADR which is Average DAILY Range. The average range/movement of a pair averaged over 1day. If you trade correlation on a D1 timeframe (TF) it should work in theory (Because its averaged over one day). But consider this. There are 4 major currency markets in a day and they all have their own speed. If you trade in M5/M15/M30/H1 using D1 ranges its not going to be accurate. Disaster waiting to happen.

By trading a larger lot size on a slow moving pair you get really screwed over when it does jump big! On M5/M15/M30/H1 it happens often and those are the ones that stop you out. I say this from experience.

I trade equal lots and you still get your pips but without the heavy draw down and inevitable stop-out. But that's just my experience. I'm keen to hear any other comments...
  • #2,214
  • Edited 1:44am Feb 4, 2011 12:16am | Edited 1:44am
Quoting Ckp
Before you go any further think this through very carefully. I have previously experimented with entering lots at a percentage equal to the daily range of a pair. I abandoned the idea pretty fast. It sounds like a good plan but it's not really. Here's why;

Assuming you use ADR which is Average DAILY Range. The average range/movement of a pair averaged over 1day. If you trade correlation on a D1 timeframe (TF) it should work in theory (Because its averaged over one day). But consider this. There are 4 major currency markets in a day...
Thanks for your input Ckp. It's good to hear your opnion. It was an idea floating in my head but I haven't been able to test it properly manually (ie enter set of trades at equal lots and another at proportional lots to compare results as I'm too slow on entering the trades and the prices move)

I'll put a note on about this post in my previous one and tell SMJ not to worry.
  • #2,215
  • Feb 4, 2011 12:23am Feb 4, 2011 12:23am
The exit at the cross of Stoch in lower TF only cut your losses earlier, but profits too......

I think that using for exits the Stochastics Lines is not the better option, we will never know how much pips this could represents.

Iīm going to try a classical way for that, with a fixed TP and SL using MPTM. Now Iīm sure I will never get in any DD.......
  • #2,216
  • Feb 4, 2011 3:39am Feb 4, 2011 3:39am
Quoting occamsrazors
Hi Kruspe. Where did you get your figure of 1.6x for UJ/GJ from?

I've been thinking about this a bit more. GJ has an average daily range of 151 pips (per my chart indi) at the moment and UJ is 70 pips. So the UJ needs to move 2.1 times as far as GJ to keep theoretically the combined value of trades equal when a news event happens (after a 60 pip spike yesterday thats a risk I want to control using triples), and to equalise the problem of one pair moving faster than another.

Comparing pair prices doesn't always work eg GBPCAD has range of 135...
That's ok I would not consider making this change, because it is beyond the scope of what the Tradable Correlations Table is. The indicator is just that an indicator not a trade setup. I have posted the source code and it is well remarked if anyone else would like to. However, I think I am finished on this indicator, because it does everything I had envisioned it to.
  • #2,217
  • Feb 4, 2011 3:53am Feb 4, 2011 3:53am
Quoting Ckp
Thanks SMJ

I would like to better understand how the logic of your correlation co-efficient works. I'm refining some correlation strategies which involve a whole lot of number crunching. Conceptually I know what im looking for but practically i don't know how it is calculated. Could you suggest where i might go for further reading material. I would like to educate myself on this. Any pointers would be useful. Thank you.
The correlation function is taken from a script I found that did not have a name on it to give credit to. However, I remember this is the exact formula as it is described at wikipedia and also investopedia. To research further I would recommend the wiki info, as it is quite concise and accurate.
  • #2,218
  • Feb 4, 2011 5:33am Feb 4, 2011 5:33am
Quoting occamsrazors
Hi Kruspe. Where did you get your figure of 1.6x for UJ/GJ from?
Yes, my fault.. I was not talking about actual prices. I took and little edit my older post to describe the idea.

Iīm gonna look to all new posts in here.. Iīm curious what new you guys did explore
  • #2,219
  • Feb 4, 2011 6:05am Feb 4, 2011 6:05am
Quoting Ckp
Before you go any further think this through very carefully. I have previously experimented with entering lots at a percentage equal to the daily range of a pair. I abandoned the idea pretty fast. It sounds like a good plan but it's not really. Here's why;

Assuming you use ADR which is Average DAILY Range. The average range/movement of a pair averaged over 1day....
Hi Ckp, damn.. good perception ... I really like it how this tread lives.. I canīt wait to new opinions, comments, etc. We make together this strategy perfect
  • #2,220
  • Feb 4, 2011 6:25am Feb 4, 2011 6:25am
The problem with Stochastics is they can be overbought or oversold for a very long time. This makes it difficult to have confidence in the difference between the two stochastics. I have been thinking about this problem, which seems to be the biggest hurdle in this method. I think a have come up with a algorithm to equate two or as many pairs as you want to each other and not have to rely on limiting devices, like stochastic.

I have to go now, but will work on it this weekend and have something to discuss in a day to two. I am actually very excited about this, because the math seems to prove out when I test it.

ps, still working on my correlation of correlations, that has turned into quite the project.
1 trader viewing now