Statistical Arb/Pairs trading strategy!

Thanks for making it! its great. but if i am doing trades on multiple pairs its the same number, is that ok?

I was going through some old accounts and found these, and I was surprised to see how much in profit I was, do you think the longer you leave the trades even in a loss they will eventually come back as the days go on because EU GU have 94% correlation on the daily chart. As you can see the other 2 accounts were traded after the first and are still in a loss, I will leave these accounts going and update in a few weeks.


Yes, no problem. The ea knows what pair is which other pair.

I have just entered 2 trades
again long eurjpy short gbpjpy
and long chfjpy short gbpjpy
I will update when they close

I have the same experience. The only detail yo need to check is to avoid fall in a margin call.

Guys I use vitrite but not sure it is enough, I had to do something else, such as fix the scale, autoscroll, these things I simply dnā€™t understand, can someone shows his Parameters?

how do you guys deal with the ā€˜free marginā€™ sometimes my platform closes a trade if my trades have used up 1/3 of my capital, and closes the trade even though i still have alot more capital left
im asking because this system originally doesnt have a stop loss

Ohā€¦ I suggest you to do the baby pips school of pipsologyā€¦

You MUST understand leverage & margin to avoid the margin calls!

Also you must know the amount of leverage you are using and the levels for margin call and stop out that your account has. It is CRITICAL for this method to have enough room to move without a margin call.

Hmmmm. Thinking about something here. If crosses are what we want then why correlation? You get crosses all day long with g/j g/chf. I think we need a correlation of about 0. Am I making any sense?

Aud/jpy e/u look at the charts 4hr for example 1 goes up the other goes down. Then the opposite. Do you see it? Maybe Iā€™m missing something.

You have partial credit hehehe. Just suppose the pair extends its divergenceā€¦ In a correlated pair the divergence is small, because they tend to move to the same direction at the same time, on non correlated pairs they could diverge for months, increasing your drawdown.

They donā€™t really diverge when they correlate at approx. zero. They just switch places.

You go long on 1 pair and short the other.

traderjosh, I like the way you think. Perhaps your right that they donā€™t need to correlate - since weā€™re overlaying charts of instruments that go up and down maybe getting two wild pairs with similar pip ranges will work better as they potentially will cross more often in a shorter time period. The few examples I looked at so far show promise. Earns looking into farther. Thanks.

I think it is not going to work on the long term, but test it for at least 2 months :slight_smile:

I will be the first to say that when you think you have found the holy Grail, you will quickly be disappointed. I did a little back testing but not too much, so this idea is certainly open to criticism. I think you are right it needs to be forward tested for at least a month, maybe two or three. Please keep posting your thoughts it makes me and others better traders.

Today I have been using e/u vs a/j and g/j vs g/chf with very good results so there is at least some hope. Lol

Iā€™m thinking in the right way to trade on both sides with Keltonā€™s method. To place a buy eu/gu and also a sell (if the broker allows full hedge)

At the moment I see more risk than benefit.

What I already realized is that If my position can touch the 100 pip stop, that means, a reverse position would be touching a 100 pip profitā€¦ and thinking about having a risk:reward of 1:1 I increased my take profit to 100 pips, equal to my stop loss of 100 pips. That could make trades to last longer, but will improve the risk:reward that is right now 5:1 to 1:1, maybe I can do 1:1.1

usd and jpy are a little correlated.

about the others, I they are totally uncorrelated! hahahaha, please forward test it and tell us the results, maybe that is the real holy grail :slight_smile:

Wellā€¦who knows. Lol