Statistical Arb/Pairs trading strategy!

So I fire up my chart with the 2nd pair overlay showing,
then click cross, so if I choose GBP/USD as primary pair, I just hit the cross hair and measure from the top of the most recent candle on GBP/USD until it hits the most recent candle of EUR/USD and that is the difference?

I.e there is a 24 pip difference right now between where the GBP/USD is compared to the EUR/USD but using the GBP/USD pips

You should not use candles nor bars. Use normal line chart, without all the OHLC. What you are looking for is the CLOSE price, so you donā€™t need to see the open, high and low values.

Using only line chart, you measure from one line to the other line and that is the divergence. If that measures XXX pips (I use 40 to 80 for eu/gu, depending if they are trending or ranging on the 1 hour view, but use the 1 minute to decide my entries) then you sell the pair that is above and buy the one that is below.

After reading a bit further on the thread and making thoughts about the system, I came across following conclusion:

Basically you are not making any hedge (as I thought before)(trading EURUSD/GBPUSD is the same as just trading EUR/GBP, like Jedster said earlier. And trading EUR/GBP is just a normal position).
So, you have normal risk here.

The direction of the position(s) you are taking, doesnā€™t really matter. Thats because you can overlay the two charts however you like (the two pairs just have to cross somewhere). For example at a divergence of the pairs, you can overlay them, so the EURUSD is above the GBPUSD but it is possible as well, that you overlay them with having it the other way round! The pairs donā€™t have a SAME specific point they BOTH refer to. It is like comparing to different things. Even if they correlate in some way, it is still not sure which one is above the other.

For example:
I opened a long EURGBP (1m TF) position on a nice divergence in the evening, but I fell asleep after waiting to long for it to close. In the morning however, a few hours later, I want to see how it worked out and I overlay the charts again! By overlaying them, there comes a convergence (cross), but over night the EURUSD has lost a lot and the GBPUSD has gained quite a lot, so this is another convergence that the one I was actually looking for. So basically it is not a divergence (related to my trade) and there is no possibilty for me any more, to get the ā€œoriginal picture of the chartā€ again and it is getting more and more difficult to close that trade in profit.

Or look at it this way: There is a divergence you are trading at the moment. You are having your position(s) opened and now waiting for the pairs to cross again. But what happens is, that the pair at the bottom falls heavily and the pair at the top rises heavily. But now the charts have adjusted to the new ā€œvaluesā€ and have shifted. Then there comes the cross, you have waited for, but you are still in loss because of the shift.

Now I know, there is the method of fixing this by scale fix. But then, there is no guarantee that there will be any cross any more. And this means, that it is the same probability like trading the normal EURGBP pair with hoping that it will come back any time soon but only that this methods provides us with a very nice method to exit the trades.

This system relates on an subjective view of the charts laying on top of each other but not on an objective point/view!

The problem is that everytime the chart makes a shift, we have a new point, that we are relating the both pairs to.
You can see that as well in the fact, that the (average) number of pips the divergence is making, depends to the timeframe you are using. But if this would be an arbitrage system or something alike, there should be always the same number of pips to take, no matter which TF you are using.
So this system just looks like we are having no risk or small risk by hedging but actually we arenā€™t hedging in any way!

It is quite difficult to put this on paper, but I hope you all know what I meant to say :slight_smile:
Please correct me, if I am wrong in any way. I am just a beginner to this subject.

what i do in that situation is open up oandas fxtrade and align the charts with the original spread (20 pips)
but on a higher time frame, say 30 min. that gives you the correct correlation,
if it blows out on that, move up to the 8hour/1day!

or you could just wait until you are 20 pips in profit then close

Please continue reading until the endā€¦ we already talk about this, and is a very frequent topicā€¦

Medisoft,

Just wondering why you wait for 40-80 on the eu/gu pairs?

Thatā€™s a number I found to be safe enough. But thatā€™s for me. Test yourself to see what is best for you :slight_smile:

Now about 100 pips divergence EU/GU from yesterday

Because the portfolio takes equal lots of EUR and GBP it isnā€™t exactly equivalent to a position in the cross.

Yes, there is still risk that the pairs never come back together, same as any strategy has risk. This is why Medi uses the stop that heā€™s discussed.

Iā€™m currently looking into a more dynamical approach normalizing using a long run average of the cross. I believe this effectively reproduces the vitrite, zoom and overlay method. Also thinking about modifying the lot sizes because equal lots biases the movement in the larger currency (GBP).

However I spend too much time crunching numbers and not enough time tradingā€¦ :stuck_out_tongue:

I am in -80 pips EU/GU. I hope that they converge next week. Is anyone in?

Iā€™m in!!! (Message to short filler)

Intro

I just wanted to thank everyone for their patience with me as I try to get myself sorted out with this method as well as
Forex Trading in general.

As I am sure some of you successful traders can appreciate, when your new to the world of Forex and have really put the time in learning about it (i.e completing the babypips.com school) introduction at least, your brain is literally in ā€œinformation overloadā€, in my case anyway, I have been confusing myself rather easily and that is currently where I am at with a few things.

What I have done below is written out my current understanding of this method and how to properly execute it from start to finish. I would appreciate it if someone could correct me anywhere there is a mistake and perhaps graciously suggest the remedy to my error(s).

I would like to have the ā€œnuts and boltsā€ of this method down soon so I can begin to demo trade it for a couple of months before even considering putting real money into this method.

I should note, the only pairs I am looking at are the EU/USD and GDP/USD, I know you guys are tearing it up with other pairs
but until I figure this out, I am going to stick to these pairs, so my focus of my post will be exclusively on that basket.

Proper Execution of the Stat/Arb Method

This is how I would start to trade this method based on my limited understanding (please note I plan on reading this thread in itā€™s entirety, once my brain recovers over the abuse of the last couple of weeks regarding information overload).

$5,000 Starting Capital
1.0% risk per trade = $50 per trade
400 Pip Stop Loss
$0.125 pip value for the GU/EU basket.

I. The Set Up

  • Wake up and check charts
  • By checking charts I should state that I have two MT4s running (Oanda) and have used vitrite to be able to few both currencies pairs on the same chart.
  • Everything is lined up perfectly and sized the same on both MT4s.

Now as of this writing the GDP/USD is above the EUR/USD on my chart, the line graph is higher. However, the value of GDP is 1.6288 and the value of EUR/USD is 1.3029.

However, I only used the 1.6288 number (last close) on the 1 min chart and use the cross tool and then move it down to where it hits the EUR/USD Last Close and its still measuring the GDP/USD chart ā€œpipsā€ (the 1.6xx numbers)and it hits the EURO Line at 1.6277 (In GDP/USD PIPS) so clearly this isnā€™t as a valid entry as the EUR/USD would need to be around the 1.6257 area to make this a valid entry.

At which time, hypothetically if that was the numbers we had here instead of the current numbers, I would sell GDP and buy EUR, sit back and watch them come back together eventually.

If they widen to 40 pips apart, I enter a 1% trade of $5000 order and if they widen to 60 pips apart I would again enter a 1% trade of $5000 order and then if they come back together I close all positions. If the divergence didnā€™t go past 20 pips, as soon as it closes we would close our first positionā€¦

You guys refer to the 1hr chart just to confirm that they are temporarily uncorrelated?

Am I remotely on the right track?

Thatā€™s the reason I enter at 80 pips if they are trending, because they tend to diverge a lot. The max divergence I have registered from my forward testing is 260 pips, so 100 is still above the half of the max. I should expect to diverge up to 180 and then retrace :slight_smile:

I think you are making good work.

Just remember that Oanda is tied to FIFO rule, so if you open 2 baskets you will be forced to close the first basket before the second.

I suggest you to look for a broker that does not impose that limitation to you.

And as I said before, ignore the real value of the pairs, use simply the pip value you got measuring with the cross tool.

Medisoft:

Sorry for being really dense here about measuring pips, but could you expand on what exactly your measuring?

My charts just show the pips on the right for the ā€œbig pairā€ i.e GDP/USD at 1.6XX i only count the pips pertaining to that pair between the two lines, so if the GDP the GDP line and the EUR/USD line, but I use the GDP/USD pips, is that right? if not, I am not sure what pips you mean?

I.e on my chart if GDP is 1.650 and then I measure down to where the EUR/USD is and happens to be at 1.630 on the GDP Chart, this would be an entry right?

FIFO Rule:

Just curious why this is a disadvantage? (Remember I am a noob :20:) My understanding is you hang on to all the positions until the lines hit, at that time I would liquidate all my positions of that trade as fast as I could? Not sure why it makes a difference that I would have to liquidate my first 20 pip entry first? because they would be all touching and I would start liquidating, please explain :slight_smile:

Or sorry, do you mean that if say I had made a first trade with GU/EU then decided to take a trade with other pairs, I would have close the GU/EU trades first, which would be problematic if they hadnā€™t touched before the 2nd trade was ready to be closed, so if the 2nd pairs like **/JPY and **/JPY I couldnā€™t close them because my GU/EU trades were still open?

Thanks for clearing this up for me

At the right you are looking for PRICE of the pair, not PIPS of the pair.

The cross tool can be used to measure pips. Simply click on one chart, and without releasing the mouse button, drag the mouse up to the other line of the secondary chart, and you will see that it shows 3 numbers separated by a slash /

number of bars / number of pips or pippetes / dont remember what else

If you are using a 5 digit account, like 1.66323 the number of pips is in pippettes, you need to divide them by 10, if you are in 4 digit account like 1.6632 they will be simple pips.

On a 5 digit you will see 53, but that does not means 53 pips, but 5.3 pips, did you catch this?

I.e on my chart if GDP is 1.650 and then I measure down to where the EUR/USD is and happens to be at 1.630 on the GDP Chart, this would be an entry right?

Yes, if you are looking for a 20 pip divergence.

FIFO Rule:

Just curious why this is a disadvantage? (Remember I am a noob :20:) My understanding is you hang on to all the positions until the lines hit, at that time I would liquidate all my positions of that trade as fast as I could? Not sure why it makes a difference that I would have to liquidate my first 20 pip entry first? because they would be all touching and I would start liquidating, please explain :slight_smile:

I donā€™t wait until the lines hit, but until I reach my desired profit for each basket. If you are thinking to close all basket at the same time, then ignore the FIFO rule, it is only a problem if you close each basket at different times.

Or sorry, do you mean that if say I had made a first trade with GU/EU then decided to take a trade with other pairs, I would have close the GU/EU trades first, which would be problematic if they hadnā€™t touched before the 2nd trade was ready to be closed, so if the 2nd pairs like **/JPY and **/JPY I couldnā€™t close them because my GU/EU trades were still open?

Thanks for clearing this up for me

I think FIFO only affects on each pair, but not sure, they donā€™t apply the rule to me because Iā€™m not US citizen.

Thinking out loud.

Hi all. Sorry for my English.

There is an idea to use a demo account as an indicator. For example:

Earlier in the week to enter the market on a demo account. Wait for example 40 pips and at this moment to enter the market on a real account in the opposite direction and wait for 20 pips profit, and then again on a new and so on.

Your opinions gentlemen?

Medisoft:

Thanks very much for your help, I think I finally am figuring it out.

I click on the higher pair and see three numbers like you describe, 0/0/1.62203, then I hold that and drag down using the cross tool and you can see the number changing as you drag down and you hit the EUR/USD and then the numbers shows 0/234/1.61***, I just need to divide 234 by 10 in this instance and would get 23.4 pips?

Thanks again.

Hi! This already was discussed in the thread :slight_smile: please read it all. I didnā€™t remember what was the result of that specific topic, but remember it.