Three Pairs Hedging | Page 28 | 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

  • #541
  • Mar 17, 2013 5:02am Mar 17, 2013 5:02am
I'm still sceptical, but I've coded the EA and will test it when the markets open. I've seen a lot of posts that say it is impossible to consistenly make money with this strategy. But, I have to give it a try to satisfy myself.

The new EA will always be in the market as long as the spreads (on entry) are reasonable. Spreads on exit aren't relevant, as the EA is only concerned with the net profit.

It will be on auto pilot when the market opens. Will let you know how testing goes.

Assuming that it performs well, I need to add the ability to trade more pairs and a few more features. At that point, I'm happy to post the EA here.
  • #542
  • Mar 17, 2013 6:21am Mar 17, 2013 6:21am
I've let the following pair run upon opening last Friday morning. Even though the pip cost is not equal, I opened 50K(USD) positions on each

S EURUSD
B GBPUSD
B EURGBP

The net sum went into $35 profit at one point (didn't note the time) and at the NY session close (close for the week) the net sum is -$16. This does work. It won't go positive every trading day, but it is useful to open a triangle hedge, let it run for a week not closing any positions and it will be relatively hedged with a net sum oscillating +/-. Spreads do need to be tight upon opening. I usually wait until 6-8 hrs after the Sydney session opens then.

Let it run manually to see how it works before you put much more time into it. Look for pairs with tight spreads and high ADR.
  • #543
  • Mar 17, 2013 7:27am Mar 17, 2013 7:27am
at a first glance, the ring EURJPY/CHFJPY/EURCHF seems to be the safest
And this is only because of the "fact" that EURCHF will not fall below 1.20
So my choice would be
S EURJPY
L CHFJPY
L EURCHF
when EURCHF is <= 1.23
I would even close the ring partially if the profit of the JPY positions is positive with a certain amount and keep a negative eurchf position knowing my current risk is max 300 pips and 1.23 is a level that isn't too hard to reach again for that pair

So overall it would be a long eurchf "speculation" with the benefit of making some extra pips when eurchf is going down
  • #544
  • Mar 17, 2013 7:52am Mar 17, 2013 7:52am
Thanks for shared and completed with trading performance also. Hedging is one of the good trading strategy if we know the right pairs to trade with. thank you very much
  • #545
  • Mar 17, 2013 9:17am Mar 17, 2013 9:17am
Quoting ramack
I've let the following pair run upon opening last Friday morning. Even though the pip cost is not equal, I opened 50K(USD) positions on each

S EURUSD
B GBPUSD
B EURGBP

The net sum went into $35 profit at one point (didn't note the time) and at the NY session close (close for the week) the net sum is -$16. This does work. It won't go positive every trading day, but it is useful to open a triangle hedge, let it run for a week not closing any positions and it will be relatively hedged with a net sum oscillating +/-. Spreads do need to be tight...
Trading 0.50 lot in each pair your hedge is imperfect. You traded (net) approximately 0.07 lot long gbp/usd . The profit or loss you get will be from the imbalance.

Ck it out. At the time your balance was +35, you probably had the gbp/usd some 50 pips up from entry, and at the close, some 22 pips below entry.

Why don't you trade only long 0.07 lot gbp/usd ... you will save a lot of spread.?

J.
  • #546
  • Mar 17, 2013 12:11pm Mar 17, 2013 12:11pm
What is the math formula to come up with a ratio in excel for 3 pairs so that a max and min range can be established over a given sample size? I have downloaded the price data and just need to plug it in the spreadsheet.

Is the correct?

=(A1*B1/2)/C1

were

A1 = GBPUSD
B1 = EURGPB
C1 = EURUSD

for a buy/buy/sell
  • #547
  • Mar 17, 2013 12:22pm Mar 17, 2013 12:22pm
Quoting jeuro
Trading 0.50 lot in each pair your hedge is imperfect. You traded (net) approximately 0.07 lot long gbp/usd . The profit or loss you get will be from the imbalance.

Ck it out. At the time your balance was +35, you probably had the gbp/usd some 50 pips up from entry, and at the close, some 22 pips below entry.

Why don't you trade only long 0.07 lot gbp/usd ... you will save a lot of spread.?

J.
The combination of pip cost and imperfect correlation is what makes it imperfect, yes, it's the imbalance, making it unhedged is how it's profitable. It was actually +90pips.

The advantages of using this triangle hedge is , no SL needed, so you won't get hit by stop hunters. I can go to sleep at night with several of these setup, let them run and not worry about it. Start watching which pairs are going in the right direction, add a little to the position, hold it longer term and spread isn't an issue. If it goes the wrong way, it's somewhat hedged, more than one position in any of the three pairs.
  • #548
  • Mar 17, 2013 2:23pm Mar 17, 2013 2:23pm
Quoting ramack
The combination of pip cost and imperfect correlation is what makes it imperfect, yes, it's the imbalance, making it unhedged is how it's profitable. It was actually +90pips.

The advantages of using this triangle hedge is , no SL needed, so you won't get hit by stop hunters. I can go to sleep at night with several of these setup, let them run and not worry about it. Start watching which pairs are going in the right direction, add a little to the position, hold it longer term and spread isn't an issue. If it goes the wrong way, it's somewhat...
I see no advantage using the triangle. Your trades net is aprox 0.07 long in the gbp/usd. Just go long 0.07 in the gbp/usd, dont put a SL and wait. Is the same of what you are doing now. If you want to hedge or ad later, just do it with 0.01 or so...

If the GU moves 100 pips up from entry, you will have aprox 70 profit (100 pips x 0.70c=70).
if the GU moves 100 pips down from your entry, you will have 70 loss.

J
  • #549
  • Mar 17, 2013 5:31pm Mar 17, 2013 5:31pm
Ok.... here it is.... the EA that I promised. It's technically working fine. It has already completed two profitable trades, one hour into trading.

As written, it will trade four different triangular "hedges".

If it continues to be profitable, with negligible drawdowns, I will be enhancing it considerably.

Each of the triangular hedges has its own Magic number.
Attached File(s)
File Type: ex4 TriangularHedge_002.ex4   39 KB | 1,149 downloads
  • #550
  • Mar 17, 2013 5:40pm Mar 17, 2013 5:40pm
Quoting ChuckNZ
Ok.... here it is.... the EA that I promised. It's technically working fine. It has already completed two profitable trades, one hour into trading.

As written, it will trade four different triangular "hedges".

If it continues to be profitable, with negligible drawdowns, I will be enhancing it considerably.

Each of the triangular hedges has its own Magic number.
Thanks for your efforts Chuck! Nice one!
  • #551
  • Mar 17, 2013 5:56pm Mar 17, 2013 5:56pm
The parameter settings should be obvious.... but I can see how they might not be.

MaxSpread (default at 5 pips). No trades will be done (for any set of three) if the spread on any of the pairs is more than MaxSpread pips. I envisage setting this to 2 pips (depending on the broker) during U.S. trading hours, perhaps wider at other times. If you set MaxSpread too low, you won't get very many trades. Set it too high and it could take a couple of hours or days to close trades at a profit. The next version will have MaxSpread for each set... as some pairs require a higher spread.

ProfitPerFullLot (default is $100). If you are trading 0.01 lots (with the default setting) all three positions within a set will be closed with a $1 profit. If you are trading 1.0 lots, the EA will quit all three positions when the open profit is $100 or more.

At the moment, the EA won't work on brokers with a suffix on their symbols. Next version will sort this out.

"Lots" should be obvious.

The EA has now done four profitable trades in 1.5 hours. I'm happy (enough) with this EA. I hope it makes some money for you. It appears to be rock solid.... that's the main thing. It closes positions when it needs to.... one thing that a lot of EA's fail to do.

Good trading!!
  • #552
  • Mar 17, 2013 6:46pm Mar 17, 2013 6:46pm
New version of EA....

1. Now handles brokers with suffixes on symbols.

2. Now has five trio combinations (instead of four).

3. Now lets you set MaxSpread for each combination.

By the way, my trading is always on live accounts with several brokers, not on useless demo accounts.

BUT.... I encourage you to try this on demo accounts until you at least get a feel for how it trades.

Currently sitting on four closed trades ($398) with five open combinations with DD of $72. You can count on an immediate DD as soon as trades are open. That's the spreads.

Next version I'll probably let you "dabble" with the pairs. So far, the pairs have been fixed.

You can see the pairs (in this version) where it prompts you for MaxSpread for each combination.
Attached File(s)
File Type: ex4 TriangularHedge_003.ex4   40 KB | 1,051 downloads
  • #553
  • Mar 17, 2013 7:29pm Mar 17, 2013 7:29pm
Quoting txfxtrader
What is the math formula to come up with a ratio in excel for 3 pairs so that a max and min range can be established over a given sample size? I have downloaded the price data and just need to plug it in the spreadsheet.

Is the correct?

=(A1*B1/2)/C1

were

A1 = GBPUSD
B1 = EURGPB
C1 = EURUSD

for a buy/buy/sell
I've read your post about 5 times and I'm still not sure if I understand what you mean correctly. But if you're looking to do a buy buy sell ring of GU EG EU then:

A1 * B2 - C1 gives you the correct triangular relationship and the residual is essentially EU - EU. However as jeuro points out if you size the pairs the same you don't get cancellation of risk. If you are trading equal size for the 3 pairs then the formula is effectively A1 + B1 -C1 which as you can verify will move quite a lot due to the sizes not cancelling. To get the A1*B2 - C1 formula or GU*EG - EU you will need to size the pairs properly. To do this I suggest you read up on the subject as this all relates to triangular arbitrage: Triangular arbitrage lot size
  • #554
  • Mar 17, 2013 7:54pm Mar 17, 2013 7:54pm
Quoting ramack
The combination of pip cost and imperfect correlation is what makes it imperfect, yes, it's the imbalance, making it unhedged is how it's profitable. It was actually +90pips.

The advantages of using this triangle hedge is , no SL needed, so you won't get hit by stop hunters. I can go to sleep at night with several of these setup, let them run and not worry about it. Start watching which pairs are going in the right direction, add a little to the position, hold it longer term and spread isn't an issue. If it goes the wrong way, it's somewhat...
Expanding on what jeuro said with putting his words to a test - is the exposure really effectively long 0.7 lot GU or something else that will make money any day with no stop loss with sound sleeping due to the (mean reverting) imbalance between the pairs? Let's let the numbers speak for themselves. Using the calculations found on the link in my previous post I start with the assumptions based on current (bid) market prices:
Attached Image

S EU 1.29087 -50000 EUR +64544 USD
B GU 1.50848 +50000 GBP -75424 USD
B EG 0.85562 +50000 EUR -42781 GBP

residuals:
EUR -50000 + 50000 = 0
GBP +50000 -42781 = 7219
USD +64544 - 75424 = -10880

So effectively long 7219 GBP, short 10880 USD are the only imbalances.
Working back from the GBP to USD you get 7219 x 1.50848 = 10864 so effectively this is long 0.7219 lots of GU with a residual of 16 USD possibly due to rounding or due to using bid prices for ease of calculation instead of bid/ask prices.

In summary I would say that the sound sleep, no stop loss strategy should (possibly) be reevaluated. <g>
  • #555
  • Mar 17, 2013 9:06pm Mar 17, 2013 9:06pm
Quoting FXEZ
I've read your post about 5 times and I'm still not sure if I understand what you mean correctly. But if you're looking to do a buy buy sell ring of GU EG EU then:

A1 * B2 - C1 gives you the correct triangular relationship and the residual is essentially EU - EU. However as jeuro points out if you size the pairs the same you don't get cancellation of risk. If you are trading equal size for the 3 pairs then the formula is effectively A1 + B1 -C1 which as you can verify will move quite a lot due to the sizes not cancelling. To get the A1*B2 - C1...
Thank you my friend. I was thinking about the triangular relationship.
  • #556
  • Mar 17, 2013 9:11pm Mar 17, 2013 9:11pm
Quoting ChuckNZ
New version of EA....

1. Now handles brokers with suffixes on symbols.

2. Now has five trio combinations (instead of four).

3. Now lets you set MaxSpread for each combination.

By the way, my trading is always on live accounts with several brokers, not on useless demo accounts.

BUT.... I encourage you to try this on demo accounts until you at least get a feel for how it trades.

Currently sitting on four closed trades ($398) with five open combinations with DD of $72. You can count on an immediate DD as soon as trades are open....
Thank you Chuck. Would it be possible to have the EA open a B/B/S & S/S/B at the same time and a re-entry when a set draw down in pips occurs for the losing triangle?
  • #557
  • Mar 17, 2013 10:42pm Mar 17, 2013 10:42pm
Hi TxFxTrader,

The next version will let you put in six sets of three pairs along with what you want to do (BBS, SSB). So, you could set the pairs the same for two sets with one set as BBS and the other as SSB.

Maybe not today, but for sure by tomorrow.
  • #558
  • Mar 17, 2013 11:02pm Mar 17, 2013 11:02pm
Quoting jeuro
Trading 0.50 lot in each pair your hedge is imperfect. You traded (net) approximately 0.07 lot long gbp/usd . The profit or loss you get will be from the imbalance.

Ck it out. At the time your balance was +35, you probably had the gbp/usd some 50 pips up from entry, and at the close, some 22 pips below entry.

Why don't you trade only long 0.07 lot gbp/usd ... you will save a lot of spread.?

J.
hi all, here's my opinion,
as jeuro said this hedge is imperfect, it's true, no gain without fluctuation. however i think we couldn't merely say why don't trade only 0.07lo gbp/usd. hedge eliminates most risk for us also some profits, but at least we know what's the level we are now, i mean i will know the price is too high or too low, so that we can make descion to buy or sell. let me take EUR/CHF EUR/USD and USD/CHF for example, we can calculate the relationship for those three pairs, say we use one-dimensional linear regression equation. now we come up the resid for equation, of course we know the standard deveriation. now we see clearly wehther the price level(std dev) lies in the range. the entry point is when the resid goes out of the range and we leave when it back to range. for this strategy we could gain 80-100 pips. stop loss depends on account capital.
  • #559
  • Mar 18, 2013 12:11am Mar 18, 2013 12:11am
@ChuckNZ

Hey Chuck,

Thanks for the EA. I'm running it on my demo account right now to check it out, but it doesn't seem to open any trades. Is the EA supposed to open trades right after activation (if spread is low enough)? Am I right if I open just 1 chart and add your EA to it? My currency symbols have a suffix, but that shouldn't be a problem, right? I'm also using a 5 decimal broker though. How's the EA dealing with that? I've tried to multiply pip settings by 10 to account for the 5 decimals, but without success. My spread is definitely not the problem as it's very tight and I've tried with using high maxspread values. In my mt4 options I have EA's enabled and allowed for live trading. Anything else I might be overlooking?

Grtz,
Koen.
  • #560
  • Edited 3:41am Mar 18, 2013 3:02am | Edited 3:41am
Quoting satfish
hi all, here's my opinion,
as jeuro said this hedge is imperfect, it's true, no gain without fluctuation. however i think we couldn't merely say why don't trade only 0.07lo gbp/usd. hedge eliminates most risk for us also some profits, but at least we know what's the level we are now, i mean i will know the price is too high or too low, so that we can make descion to buy or sell. let me take EUR/CHF EUR/USD and USD/CHF for example, we can calculate the relationship for those three pairs, say we use one-dimensional linear regression equation. now...
Hi satfish.

The risk incurred is the risk incurred. If by making 3 trades you have a net position of long GU 0.07, your risk/reward is of 0.07 long GU.
If you trade only 0.07 GU long, your risk/reward is of 0.07 long GU.... exactly the same. The cost of making 3 trades of 0.50 is about 22.50
and the cost of doing 1 trade of 0.07 is about 1.05. Plus if the positions are kept for a while you also have the cost of swaps.

I understand that good strategies can be figured out from this triangles of pairs. But all the the details you mention....levels, regressions, equations, etc, can be calculated by just placing 3 horizontal lines in all 3 pairs in the current prices at one point on time. No need to spend money in spreads for it.

J.
0 traders viewing now