Signals for Simplicity

Hi there,

This is a quick question about the correlation of EURGBP and GBPCHF and the trading 14 pairs straight method.

As we know EURGBP is a BUY pair and GBPCHF is a SELL pair.

After doing some investigation these two pairs seem to have a pretty good negative correlation. This means of course that when EURGBP is going up in price you will find that GBPCHF is going down or when EURGBP is going down GBPCHF is going up.

Because EURGBP is a BUY pair and GBPCHF is a SELL pair and since they have a pretty strong negative correlation these two pairs tend to stick together either in the top half of the basket (i.e. profit) or in the bottom half of the basket (i.e. loss).

When trading 14 pairs straight it would appear that these two pairs would be at odds with each other.

For example let us assume the top and bottom anchor pairs change so that an opportunity arises for a 14 pair straight "BUY" trade. In effect let us say that a BUY pair takes over the top position in the basket and then a couple of hours later a SELL pair takes over the bottom position in the basket.

My understanding is that this is a good signal to initiate a BUY of all pairs. If I do this then the expectation would be to see both EURGBP and GBPCHF to go up in price over time. However given the strong negative correlation between EURGBP and GBPCHF if EURGBP goes up in price there is a strong possibility that GBPCHF will go down which isn't so good since we would actually be long on GBPCHF.

Am I missing something here with regard to the trading of 14 pairs straight?

How does the strong negative correlation between EURGBP and GBPCHF affect trading of these pairs in the context of the Basket Trading System in general.

Regards,

Laurence.

Hi Laurence.

This issue was discussed under the term "optimization" in this thread and Julius's original. Try searching the two threads for this term and see if that helps. I agree there do seem to be some issues around the trading of highly negatively correlated pairs (EU & UC especially). Some posters seem to think it is a bad idea while others think it has a contributory factor to the overall (14 pair) trade.

Personally I always 'optimize' EU and UC ... i.e. trade UC in the opposite direction to the other 13 pairs. As I am very much of the opinion that trading them in the same direction serves no purpose to the overall trade results. Including them together in the same direction has the same net result (profit) as if you had not included them at all.... so why pay the spread?....IMHO.

Hope this helps.

Thanks for the heads-up cdavid - I shall try and find the posts you mentioned.

Regards,

Laurence.

I read the previous posts regarding "optimization" as Cdavid suggested.

Having read the posts regarding "optimization" there are it seems some caveats when it comes to Trading 14 pairs straight. I had thought that when trading 14 pairs straight one would either trade all 14 pairs LONG or all 14 pairs SHORT; this it appears is not the case.

Given the strong negative correlation between the BUY pair EURGBP and the SELL pair GBPCHF mentioned in my previous post should these two pairs be be "optimized" when trading 14 pairs straight?

For example:

If the basket looks like it is favoring the BUY pairs then GBPCHF should be a SELL rather than a BUY.
.
If the basket looks like it is favoring the SELL pairs then EURGBP should be BUY rather than a SELL.

For me the correlation of the basket pairs is an important element in deciding what direction to trade each pair. Certainly trading 14 pairs straight does seem to require a degree of finesse that wasn't immediately apparent to me.

Regards,

Laurence.

For those that are interested I have discovered the reason why the Combinaitons Generator was running SOOO slowly.

I was recreating the string in the text box for each combination that was generated........

Appending to the text box works a thousand times faster.

I have included the new deployment files if anyone wants the much faster version. Also I tidied it up a lot with more error checking, an output to file feature and more user feedback.

Flying through the generations now so it should be a couple more days before I complete the list of new baskets.

Though I can say right now... if the 14 pairs basket is anything to go by ... that there will be 10's of thousands of possible balenced baskets.

Even when reducing the list such that it only contains baskets that have both spread AND ATR absoulte differences below 30% it still leaves thousands. And thats just for the 14's basket. I am expecting it to be higher for the bigger baskets.

I will also be grouping lists. It is possible for a handful of baskets to contain the same pairs but in different orders, so that the spread / ATR difference is different for each of them. So I will include these identicle lists with a grouping value ie .... group1 contains 4 baskets, each with the same currency pairs in them, each balenced, each with spread / ATR differences below 30%... but in a differenct order so that the Buy/Sell spread/ATR is different for each. I will leave it up to you to decide which of the group is the basket you like.