Statistical Arb/Pairs trading strategy!

[medisoft]

Thanks for your input Talon. I have been racking my brain to try and figure out why the profit seems to disappear. I have 2 charts from same broker overlayed on 1m time frame. I should have made more profit when I go back and calulate the pip moves on each individual chart. So what I am seeing with the charts overlayed is not the real correlation picture. Maybe adjusting one of the charts up or down or throwing up an indicator such as ATR on each chart and see if that gives some signal to help view the real picture? I will continue to dig deeper.

Do you checked out the spread (commission) of each pair? Maybe the spread increased a lot Like it does on friday.

[christinaa]

Do you zoom all the way out on the 1hour time frame?

[timehopper]

The only thing I disagree with there is that it is rare. I think it is just as likely, roughly 50 50 chance it will cross and go that way as touch and go the other way. If you look at the charts you will see the two pairs crossing back and forth fairly often. Although with EurGbp being in a downtrend recently it may be more likely to find EU line below the GU line than the other way around so not quite 50:50. But I think that trend would affect higher time frames and not be so much of a factor on shorter time frames where price seems more random.

But still it’s good to have a well defined exit point. That’s a problem with some trading strategies, they may have well defined entry conditions but so many of them don’t seem to have well defined exit conditions.

You are off in this matter. if they pairs have a +.80 correlation by definition they CANNOT have a 50/50 chance of crossing in the opposite direction and continuing along that path for very long . they would have to move in the same direction 80% of the time. meaning they can ONLY TRAVEL IN DIFFERENT DIRECTION 20% OF THE TIME THUS MAKING THAT A RARE EVENT. if you were correct the correlation would be much lower. if you check your math you will see this. It is far more like they will touch again and travel in the same direction together regardless of where on the grid this happens to occur. this is why ALMOST always one side has to be a loser.

[timehopper]

Do you zoom all the way out on the 1hour time frame?

yes, regardless of the time frame being used I zoom all the way out. I believe in getting as much data as possible. I like pairs that are almost always touching in the timeframe I choose to use for the trade. the reason I have been silent on matching the charts up is because it seems to me most people on this thread are using metatrader. I am not, so I don’t have to set pixels per pip or something or other. I simply open two charts on the same pair to make sure they are perfectly lined then change one window to another currency pair.

[timehopper]

Optical illusion…

@Talon D and FXEZ . so I looked at the charts as you suggested and could not figure out why you thought they crossed and continued in opposite direction on a “50/50” chance. then it dawned on me that you have your charts set up in a fashion that is not normal. The problem is you have changed your rising and falling candles to the same color let’s say red on one chart (e/u) then on the other (g/u) you have made them let’s say blue. my guess is you have done this so when you overlay them you can see them better. that’s your problem, you can’t see if each bar is in fact rising or falling bar ,creating the illusion that they cross and continue on because there is only one color per chart. Thus giving the impression that all the bars are moving in one direction on one chart and another on the other. but if you were in the trade you unrealized profit/loss would reflect otherwise. for example look at christinaa’s chart above… to the untrained eye it looks as if they cross and then continue in opposite directions. This is in fact not the case. if you were to look closely you would see that even though the pairs touch and cross each other then “seem” to continue on; almost all the peaks and dips line up. this is because of the correlation. they almost always peak and dip at the same time regardless of the position on the grid. if you were to look ever closer (literally) by zooming in and changing the rising and falling candles back to the original colors ( red/blue) on the same chart and did the same for the other pair then lined them up you will find that when one candle is blue 80% of the time (or what ever the correlation is for that time frame) the other pair will be blue as well. if it is not then it’s the 20% of the time when they are not in lockstep and is about to present an entry opportunity.

It only looks weird because one pair has to outperform the other (bigger/ faster peak or dip) in order to create the entry and exit opportunities otherwise this system would not work at all.

hope that helps.

[guandi]

Hello Timehopper,

what platform aare you using since you are not using MT4?

Best regards.

-guandi

yes, regardless of the time frame being used I zoom all the way out. I believe in getting as much data as possible. I like pairs that are almost always touching in the timeframe I choose to use for the trade. the reason I have been silent on matching the charts up is because it seems to me most people on this thread are using metatrader. I am not, so I don’t have to set pixels per pip or something or other. I simply open two charts on the same pair to make sure they are perfectly lined then change one window to another currency pair.

[TalonD]

You are off in this matter. if they pairs have a +.80 correlation by definition they CANNOT have a 50/50 chance of crossing in the opposite direction and continuing along that path for very long . they would have to move in the same direction 80% of the time. meaning they can ONLY TRAVEL IN DIFFERENT DIRECTION 20% OF THE TIME THUS MAKING THAT A RARE EVENT. if you were correct the correlation would be much lower. if you check your math you will see this. It is far more like they will touch again and travel in the same direction together regardless of where on the grid this happens to occur. this is why ALMOST always one side has to be a loser.

They are just as likely to cross in one direction as the other. If there is a 20% chance they will diverge then 50% of the time that they diverge EU will go up and GU down and 50% of the time that they diverge EU will go down and GU will go up. If you look at a chart you don’t see the two lines exactly overlayed so that they appear as one line very often.

edit: I just saw the post about the candle colors. Yes I have mine set to a single color and it’s set to the line style instead of candles. But still if the candle colors match 80% of the time while they are moving apart then that’s on a smaller time scale than the divergence you are looking at.

[timehopper]

They are just as likely to cross in one direction as the other. If there is a 20% chance they will diverge then 50% of the time that they diverge EU will go up and GU down and 50% of the time that they diverge EU will go down and GU will go up.

(your previous post…Also the touch point is not maximum profit. If they touch and continue moving in that direction then you have even more profit. So for example suppose you go short on EU and long on GU. EU moves down and GU moves up and they touch so you are in profit. But suppose you don’t close your trades and EU continues to move down further and GU continues to move up further and they continue to diverge that way then you have even more profit

Let me be clear…
I was not responding as to which pair will cross which or in what direction. I nor this system couldn’t care less since it’s just a matter of time.
My post was in response to your previous posting, about not getting out when they touch. the situation you described can only occur in a minority of the time. I.E. sticking with the +.80 correlation example for ease of numbers. the aforementioned situation can ONLY happen 20% of the time, since the pairs can only be out of correlation ( travel in different directions ) by that much. they other 80% of the time they travel in the same direction regardless of the position on the grid. When money is involved 1 out of 5 is rare to me. i would rather bet on the 4 out of 5 side. but that’s just how I trade. to each his own…

[timehopper]

Hello Timehopper,

what platform aare you using since you are not using MT4?

Best regards.

-guandi

Guandi,

Sorry, not to seem rude but not MT is about all I can say on the matter. But for the time I’m more than happy to help with any other questions you may have.

edit. Some of the others posted some things about getting the charts right. check the middle of this thread

[christinaa]

There may be a solution to the pip per pixel ratio if I understand the issue correctly.

If you right click on each of the two chars you are overlaying and go to Properties > Common tab then click Scale fix.

You then need to change the figures so that the [B]difference [/B] between the fixed maximum and fixed minimum on both charts is the same. This will make sure both overlaid charts are showing at the same scale which should result in pips per pixel being the same across both charts!

Hope this makes sense!

Danny
When I click Scale fix whether or not I change the numbers, the chart then has a vertical scroll on the right
Did you get this?

[dannyg86]

Danny
When I click Scale fix whether or not I change the numbers, the chart then has a vertical scroll on the right
Did you get this?

Yes I do!

For everyone as a helpful tip when you lock the two scales and have lined up both charts on top of each other, what I do is then put in 2 horizontal lines on each chart i.e 1 blue and 1 red on each chart. Then make sure that each matching colour line overlays exactly with same colour line on the underlying chart. If you later notice that you can see more than 2 lines then you can clearly see that the charts have shifted for what ever reason.

This should help stop a lot of issues that people are getting when their currencies meet and they are seeing their trades are still making a loss.

The only 2 ways people are making a loss when their currencies cross over is because they couldn’t cover the spread or their charts have shifted since entering the trade.

[guandi]

Hello Timehopper,

sure, many thanks for answering.

-zhao

[oakman]

I was thinking checking scale to fix might help and set up my charts that way yesterday. I am going to try and trade it that way. The horizontal line is a good idea.:35:

[oakman]

Looks like instead of 20 pips apart to trigger trades I need to wait until 45-55 pips since I have checked scale to fix on the charts. I have 1m charts overlayed and zoomed all the way out.

[TalonD]

Mentioned it somewhere back there… not that it’s useful but it’s interesting. Fifteen thousand plus data points going back to June 2011 of EU and GU.

The second chart with the purple line shows the difference in their prices over that time period.

Good luck and green pips!

[timehopper]

all trades (2 sets) opened yesterday were closed this morning for a profit well over 40 pips. Used the
aud, cad,usd and nzd currencies… not bad for putting on the trades and just waiting for around 24 hours. I would like to hear some success stories from others using Stat/Arb. I think if some of us started posting our success as well as failures it might be a good thing for all. Although I think it will be more success than failures… :53:

[atlee73]

If you didn’t make money yesterday/this morning using this system you are doing it wrong, it should have been a big day for everyone yesterday into today.

[patforex7]

I have made profits, almost on all the trades. I pick any trade which provides me about 15pips as a gross profit. So far all of them have been successful without a single failure. Total in about 30 trades. All done either on 1M or 5 Min.

Chart alignment is very important to see visual divergence. So make sure you do it right, as discussed. earlier in the forum at length.

Or find a charting application which provides Chart Overlay feature like FXCM does. I found FXCM is pretty cool in overlaying charts. Not sure if anyone has tried it. If yes, feel free to share your experiences too. Specially the bad experience to be aware of.

I am sure there should be some advanced charting apps that may provide Chart Overlay feature. If we find one, the manual method of adjusting MT4, the ugly method, can easily be eliminated.

Chart overlaying can provide you real divergence without worrying about human errors, that you can make while doing it manually.

[Jedster]

I have made profits, almost on all the trades. I pick any trade which provides me about 15pips as a gross profit. So far all of them have been successful without a single failure. Total in about 30 trades. All done either on 1M or 5 Min.

What has been your equity drawdown during these trades? I assume at times pretty big as targeting 15 pips and having the pair widen (to say 50-100) would result in a fairly large draw down.

Also, out of curiosity, what technique are you using to decide your position size?

[patforex7]

What has been your equity drawdown during these trades? I assume at times pretty big as targeting 15 pips and having the pair widen (to say 50-100) would result in a fairly large draw down.

Also, out of curiosity, what technique are you using to decide your position size?

Forgot to mention they are done in Demo thus far. During this short journey of trades, I have seen only few trades reached to about 100 pips apart / DD but eventually they returned back and I made my profits.

I have decided to use one small trade for 20pips divergence and another trade at 40pips divergence, if they diverge beyond that I wait them to return back. Though my maximum DD that I have decided is 150pips which has not occurred so far in 1m and 5m time frame.