Three Pairs Hedging | Page 33 | Forex Factory

Options

Go to
Permalinks
Print Thread

Related Instruments

EUR/USD, CHF/JPY, EUR/CHF, EUR/GBP, EUR/JPY, GBP/CHF, GBP/JPY, GBP/USD, USD/CHF, USD/JPY

Similar Threads

Hedging all possible currency pairs at once! 13 replies

Hedging synthetic pairs 25 replies

Hedging correlated pairs using pip difference 16 replies

Need help hedging Exotic Pairs 2 replies

Trending Pairs / Ranging Pairs 0 replies

Attachments: Three Pairs Hedging
Exit Attachments

Three Pairs Hedging

  • #641
  • Mar 22, 2013 4:14pm Mar 22, 2013 4:14pm
Quoting milou
Hello ChuckNZ,

I think your great work ... but I think you could improve the entry, change the system spreads. Currently if I understand the method of calculation is that speads the spread of each pair must be less than the amount chosen to open three positions.
example:
spread 2.9
usd / chf 02.09
GBPCHF 2.3
gbpusd 1.8
total 7.0
I think it would be more profitable to modify the conditions of entry by accumulating the sum of the three spreads less than the amount chosen.
spread 2.9
usd / chf 0.7
GBPCHF 1.1
gbpusd 1.1
total 2.9

that many provoquerais...
Hello Milou and thanks for your comments. I agree with you 100%... with the benefit of hindsight. It will be a simple change and I'm sure it's a change for the better. When I change it, however, I will make it optional for those who have finally mastered an understanding of the earlier versions.

I haven't (yet) explained what the EA is doing with SmartEntry. I will do that over the weekend. It is possible, using SmartEntry, that checking the spread isn't even important. I'll explain that later too.

In the mean time, I will change the EA to use a combined spread filter.

Thanks again for pushing me along.
  • #642
  • Mar 22, 2013 4:31pm Mar 22, 2013 4:31pm
ChuckNZ,

Since I started this EA it has the following results;

SETS opened closed
usdjpy/audusd/audjpy 1 0
gbpusd/usdchf/gbpchf 4 3
eurchf/chfjpy/eurjpy 1 0
eurusd/usdchf/eurchf 1 0
eurgbp/gbpusd/eurusd 1 0
gbpusd/usdjpy/gbpjpy 5 4
eurusd/usdjpy/eurjpy 3 2

Only three "Tri-pairs" seem to reach TP....gbpusd/usdchf/gbpchf gbpusd/usdjpy/gbpjpy eurusd/usdjpy/eurjpy
Is this the case with your results too?

Bruce
  • #643
  • Mar 22, 2013 6:41pm Mar 22, 2013 6:41pm
Quoting jeuro
To calculate you need to know the prices of the 3 in same tick.

Current Prices are Eur/usd 1,29476, gbp/usd 1.52053 and eur/gbp 0,85147

Buying 100 000 Pounds (buy gbp/usd) will get you +100 000 pounds and and a debt of -152 053 usd
Buying 100 000 Euros (buy eur/gbp will get you +100 000 euros and a debt of -85 147 pounds
selling 100 000 euros (sell eur/usd) will get you + 129 476 usd and a debt of -100 000 euros

To have a "perfect hedge" one would need to match 1 per one what we own and what we owe.
But is this case we do not.

We own...
Thanks, Jeuro, for taking the time to explain this to me. I just wish that I wasn't exhausted from no sleep for several days while working on this EA.

I understand everything up to where you say that the above BBS deal is essentially long GBPUSD. I get the USD22,577 and the GBP14,854. But, I don't understand the 0.01483 lots.

In any event, what I really want to know... and you have answered that for me.... is which pair we are exposed to.... in this case GBPUSD and which direction (long).

What I plan on doing is deciding whether to use BBS or SSB, based on that exposure and using different methods of predicting the direction of the exposed pair. I've code the function to do the calcs you provided above and I'm tesing it now. As I said, it would help if I didn't fall asleep in the middle of typing!!

Will have a new version by Monday which will be an attempt at predicting the direction of the exposed pair.

Can you confirm that a movement of any consequence in the value of the set can only be attributable to the pair with unbalanced exposure?

If you have any suggestions for predicting the direction, feel free to make suggestions.

Thanks
  • #644
  • Mar 22, 2013 11:32pm Mar 22, 2013 11:32pm
Quoting brucech
ChuckNZ,

Since I started this EA it has the following results;

SETS opened closed
usdjpy/audusd/audjpy 1 0
gbpusd/usdchf/gbpchf 4 3
eurchf/chfjpy/eurjpy 1 0
eurusd/usdchf/eurchf 1 0
eurgbp/gbpusd/eurusd 1 0
gbpusd/usdjpy/gbpjpy 5 4
eurusd/usdjpy/eurjpy 3 2

Only three...
Hi Bruce,
I have similar results with TP reached on gbpusd/usdchf/gbpchf 2 times, gbpusd/usdjpy/gbpjpy 4 times.
I also had stop loss set at 300 and was stopped out on gbpusd/usdjpy/gbpjpy once.

Paul
  • #645
  • Edited 1:49am Mar 23, 2013 12:12am | Edited 1:49am
Hi Chuck,
Quoting ChuckNZ
Thanks, Jeuro, for taking the time to explain this to me. I just wish that I wasn't exhausted from no sleep for several days while working on this EA.

I understand everything up to where you say that the above BBS deal is essentially long GBPUSD. I get the USD22,577 and the GBP14,854. But, I don't understand the 0.01483 lots.
Jeuro is referring to the residual 14,853 GBP that is found in this line:
Quote
We own 100 000 pounds and we have a debt of -85 147 pounds... we own an extra + 14853 pounds

So 14,853 / 100,000 = 0.14853 lots of GBPUSD.


Quote
Can you confirm that a movement of any consequence in the value of the set can only be attributable to the pair with unbalanced exposure?

If you have any suggestions for predicting the direction, feel free to make suggestions.

Thanks

If you open up the same size at the same time on the ring previously specified, you effectively get GU long based on the pairs chosen and the respective prices / exposure.

1) If you change size, you can change net exposure. There are an infinite number of different net exposures you can develop but only one true hedge that is stationary, aka triangular hedge, impeccable hedge, perfect hedge or triangular arbitrage, where all exposures net out to zero.
2) If you stagger time of opening (or closing) each pair, you change the P&L dynamics (but not net exposure).

If you have a net exposure to GU as described above, and open and close all positions on the 3 pairs at once, your P&L is totally driven from the net exposure. In this case the net exposure is GU long. Simultaneously opening and closing all 3 positions severely limits the tools available to you. You are effectively left with only #1 to manage your net exposure but no way to alter the P&L dynamics through your strategy. If you are perfectly hedged and open and close all positions simultaneously, you simply eat spread each time with no chance for directional movement.

IMO #2 is the more accessible approach (changing time of opening/closing each pair) due to minimum size for rounding purposes offered through most MT4 brokers. But, if you can effectively hedge out most of the directional currency risk through evening out the net currency exposure (#1), then you are left with the interesting P&L dynamics you have created via staggered (#2) entry / exit times = financial engineering with no need to predict direction!

For ideas to change the P&L dynamics of the hedge, STUDY this post then read forward a few posts (down to post # 491) and then skip over to Three pairs hedging/Renko, particularly the posts by jeuro who shares a few more tidbits on the original idea before he goes dark.

I prefer not to post my name on these boards for privacy, but since it's already out there on the web, for you Chuck I'm Patrick.

By the way I have taken a cursory look into the concept posted by jeuro and can confirm after spending an evening or three doing some simulations in R that there is "something there". Unfortunately I haven't had time to fully explore or even implement the concept due to all the projects that I must finish first, but it is definitely on the "to do" list.
  • #646
  • Mar 23, 2013 5:41am Mar 23, 2013 5:41am
Quoting ChuckNZ
Thanks, Jeuro, for taking the time to explain this to me. I just wish that I wasn't exhausted from no sleep for several days while working on this EA.

Can you confirm that a movement of any consequence in the value of the set can only be attributable to the pair with unbalanced exposure?

If you have any suggestions for predicting the direction, feel free to make suggestions.

Thanks
Thanks for the detailed explanation FXEZ

Yes, I confirm the profit or loss in equity comes only from the unbalance exposure. There is no magic in placing
3 trades simultaneously. It is only math.

No, I don't have any "regular" suggestion for predicting the direction. That is the eternal quest of us traders.
The only suggestion is stated in that one old post of mine when I started to look for ways of back testing certain concepts.

I know (in a mathematical form) what the price will be in the future of almost any pair. Transforming that knowledge
into a sensible strategy is the hard part.

The main purpose for posting in this type of threads is to somehow prevent that traders/coders waist time and energy.
Specially like you Chuck, that are willing to roll up you sleeve and work hard at it.
I fully identify myself with the exhaustion and no sleeping. I still do the same once in a while.

J.
  • #647
  • Mar 23, 2013 6:24am Mar 23, 2013 6:24am
Quoting jeuro
Thanks for the detailed explanation FXEZ

Yes, I confirm the profit or loss in equity comes only from the unbalance exposure. There is no magic in placing
3 trades simultaneously. It is only math.

No, I don't have any "regular" suggestion for predicting the direction. That is the eternal quest of us traders.
The only suggestion is stated in that one old post of mine when I started to look for ways of back testing certain concepts.

I know (in a mathematical form) what the price will be in the future of almost any pair. Transforming that...
Hi Jeuro and thanks again,

I am testing a new version that calculates the net exposure and trades each set using BBS or SSB depending on whether it "thinks" the net exposure is going to go up or down. It's working nicely (I have tick mill for weekend testing), but the indicator I'm using has an intermittent bug that is driving me crazy. I've been on it for 21 hours, so may have some sleep which usually helps to solve these kinds of problems.
  • #648
  • Mar 23, 2013 11:46am Mar 23, 2013 11:46am
Quoting ChuckNZ

1. Can someone show me how they calculate this?
The author of articles 1 and 2 does a very good job in explaining the arithmetic.
Rich
  • #649
  • Edited 8:56pm Mar 23, 2013 4:57pm | Edited 8:56pm
Major upgrade attached. Important notes follow. Please don't use this new version without reading the following.

1. Thanks to a couple members of this group, I now (almost) understand that trading equal sizes of the three pairs in each set is only a parital hedge. All through this thread, they have been trying to get it across that:

a. Trading equal lot sizes is NOT going to provide a 100% hedge.
b. If it was a 100% hedge, there wouldn't be any potential to make (or lose) money.
c. Any profit (or loss) that we make is due purely to the movement of the residual exposure.

2. Early last week, Murphy's Law was good to anyone trading with this EA. It took dozens of trades that were all profitable. Towards the end of the week, the sets were slow to go into profit and/or are still sitting on DD.

3. I added something called "SmartEntry" that attempted to squeeze a couple of extra pips out of each set. For the most part, it helped. But, two pips isn't enough of an edge.

4. I now understand that trading equal-size lots has exposure to two currencies (not pairs). Long GBP and short USD (for instance).

5. Unless we add some logic to "predict" the direction of the remaining exposure, we have less than a 50/50 chance of making a profit (due to spreads).

6. So, this version is my first attempt at predicting the direction of our exposure in each set.

7. This new feature is activated by setting AutoDirection to true (the default setting).

8. When AutoDirection is set to true, an indicator (CurrencyStrength) is called. So, you will need to have that indicator in your indicator folder.

9. When AutoDirection is set to true, the EA will decide which of the following options to use for each set:

a. No trade (seldom happens).
b. Trade the set using BBS logic.
c. Trade the set using SSB logic.

10. When AutoDirection is set to true, each set will be using a different Magic number than the number used when it is set to false. So, for testing purposes, you can run two instances of the EA. One with AutoDirection set to true and the other set to false. This will tell you (and me) whether the new logic is beneficial.

11. Another important feature of this version is that the OrderComment field now shows the Set number (1-8) and the magic number. This is handy should you decide to quit trades in one set.

12. This version is now using EURGBP, GBPAUD, EURAUD for Set 7. The previous pairs were duplicated in Set 1. Set 1 and Set 7 were previously the same, one using BBS and the other using SSB.

13. Another member of this group has offered an indicator to me for trying as part of this automatic trend detection logic. I will be working with him (or her) as soon as I have another cuppa.

Since it's the weekend, I can't thoroughly test this version. I'm 99% confident, but suggest that you only trade it on demo accounts for the first few hours of trading on Monday (or Sunday for some of you). I will provide a quick update should there be any problems.

Good trading!!
Attached File(s)
File Type: ex4 CurrencyStrength_001.ex4   20 KB | 815 downloads
File Type: ex4 TriangularHedge_021.ex4   77 KB | 869 downloads
  • #650
  • Mar 24, 2013 2:39am Mar 24, 2013 2:39am
If you would like to help with this project, it would be HUGE if you could run two instances of this EA on a demo account for a couple of days this coming week.

Instance #1 with AutoDirection = true.

Instance #2 with AutoDirection = false.

Both on same account. Both applied to a EURUSD M15 chart. Both with the same parameters except for AutoDirection.

When the orders are closed, please have a look at the trades during the test to see which instance put on the trades and make a summry for each instance.

The magic numbers for AutoDirection = true will be between 82608 and 82626.

Magic numbers for AutoDirection = false will be between 12879 and 12896.

Please report any bugs or unexplained behaviour right away so I can fix it.

Thanks for your help.
  • #651
  • Mar 24, 2013 2:43am Mar 24, 2013 2:43am
Chuck,

Count me in...

I will set them up on FXDD and Pepperstone.
Both on same account. Both applied to a EURUSD M15 chart. Both with the same parameters except for AutoDirection. Everything else will be OOTB parameters.

Bruce
  • #652
  • Mar 24, 2013 4:22am Mar 24, 2013 4:22am
Chuck,
I will run it on a demo account to as you requested.

Cheers,
Paul
  • #653
  • Mar 24, 2013 7:09am Mar 24, 2013 7:09am
ill set up test on 2 brokers also.
brian
  • #654
  • Mar 24, 2013 7:25am Mar 24, 2013 7:25am
No offense but there seems to be a real knowledge vacuum in this thread. In examples listed in the first post, there is no hedging being done. You are simply trading with an exposure equivalent to a single pair that is disproportionate to the volume traded. You use more margin, pay more in transaction costs and gain nothing.
  • #655
  • Mar 24, 2013 12:13pm Mar 24, 2013 12:13pm
Quoting saussiche
No offense but there seems to be a real knowledge vacuum in this thread. In examples listed in the first post, there is no hedging being done. You are simply trading with an exposure equivalent to a single pair that is disproportionate to the volume traded. You use more margin, pay more in transaction costs and gain nothing.
Everyone here understands this. We get the mathematics. We understand it is basically trading 1 pair. What you and a couple of other people do not seem to get is this is an attempt to have a robot trade for people. One that has little draw down and will make money with little human intervention. That is the purpose of all Trading Robots. You can add logic, technical's and mathematics to a robot as has been done to all robots and I challenge you to show me one that will work year after year without blowing up an account. (BTW, i have seen a few that work for the creators in back tests and forward tests. but when tested by other people don't work) I also looked back on your previous posts in other threads and see you have given no constructive advice or any strategies or systems for trading. I see nothing posted by you to indicate that you make money trading and you are not obligated to either. That said, you are also not obligated to criticize other peoples hard work to just give a negative comment (although that is your right). Whenever you come to someone with a problem you see, you should also come up with at least 1 solution for the problem. Can you do that? Maybe start a thread with your strategy or system that you make money with?
Cheers,
Len
Think Outside the Box.....
  • #656
  • Mar 24, 2013 4:29pm Mar 24, 2013 4:29pm
Hi Chuck, setup a new demo account at lqdmarkets. They have fixed low spreads. I am running demo with your request. True and false. Will post results once a day or so.
Cheers,
Len
Think Outside the Box.....
  • #657
  • Mar 24, 2013 4:53pm Mar 24, 2013 4:53pm
"Thanks" to anyone testing this EA.

I've only had one trade per instance (autodirection true and false) on all accounts (live and demo) so far. It's interesting that one instance used BBS and the other used SSB for the same pair. Due to the large spread, they both have a small open loss and the loss for the set on AutoDirection is smaller.... FWIW. The trades were for set #5.

Spreads are quite wide, of course, for the first few hours of the trading week.

You may wish to turn off SmartEntry for this test. It really slows down the order generation and may not be worth the effort.

The next version displays the settings for AutoDirection and SmartEntry to make life easier.

I also noticed that set #1 isn't being controlled by AutoDirection. That, too, has been fixed in the next release.

Another trade just opened... on both instances.... for set #2.

And another for set #10.... and #4. Spreads are narrowing.

Good trading!
  • #658
  • Mar 24, 2013 5:15pm Mar 24, 2013 5:15pm
Mine says set #7 is not being traded although i did nothing to change this. Is this a bug? Paris
  • #659
  • Mar 24, 2013 5:27pm Mar 24, 2013 5:27pm
Quoting Paris Hilton
Mine says set #7 is not being traded although i did nothing to change this. Is this a bug? Paris
Hi Paris,

No... it's not a bug.... it's a feature. I'm sure you only see that on the instance running with AutoDirection set to true?

It's basically saying that it doesn't have enough information to decide whether to use BBS or SSB. That will change as the day progresses. With EURUSD gapping down this morning, it will take a while for trends to be detectable.

I also just noticed that the OrderComment() field doesn't contain the set number and magic number that I meant to have working. That, too, is fixed in the next release.

I'll probably have a new version posted here in three or four hours... just trying to incorporate as much as possible into the one release.

I couldn't test actual order generation over the weekend, so I missed a couple of things.

The EA has closed a couple of profitable trades on one of my brokers already. One was using AutoDirection and the other was not. Jury is out as to whether the trend detection indicator I'm using is adding value. I'll need a few hundred trades to make that determination.

I'm also working on four different trend detection methods to build into the EA (all optional, of course).
  • #660
  • Mar 24, 2013 6:42pm Mar 24, 2013 6:42pm
Chuck,


Option request;
For direction testing purposes, could there be an option for forced trade entry of the three pairs in a determined direction?
With that option, we can view the charts and see possible direction or trend and then force entry.

Bruce
3 traders viewing now