Correlation Trading

Hi! This is my first post on BabyPips. I’ve been trading Forex on an MT4 demo account for many years but never been able to find a method which could realistically be relied upon to provide anything approaching a profitable system over time. It’s claimed that only a small number of active traders are successful, often claimed around 5%, with their profits being unwittingly funded by the other 95%. So why would anyone jump into trading a live account until they can be confident of their ability to succeed? And yet they obviously do because the 95% of unsuccessful traders are real. Perhaps, like it or not, this is how the system operates – without the reliable support from the overwhelming majority of losing traders, there wouldn’t be a workable trading market - a sobering thought.
Having explained my reluctance to fund the profitable traders I’d like to mention my recent focus on ‘Correlation’ trading. This type of trading, compares one currency pair against another, in a logical way, in an attempt to predict the future movement of the identified Forex pairs to identify potential trades. This method uses no indicators, the basics are easily learned, and there is information readily available on brokerage sites to be able to find correlated Forex pairs with the most potential. There’s plenty of info of this subject so I deliberately haven’t gone into the specifics of ‘Correlation’ trading. The main problem I struggle with is deciding which time frame to use to compare potential correlation trades. I see everything from Weekly, Daily down to 5 minutes suggested, so if anybody can give their thoughts on this point I think it will bring more clarity to a system which has a simple logic about it and considerable potential as a profitable trading method.
NB. I am not a qualified trader but I would suggest that trading ‘Negative Correlations’ is more profitable but with higher risk whilst trading ‘Positive Correlations’ has smaller profit potential but with minimal, if any, risk.
If all this makes no sense, have a look around ‘Google’ and ‘UTube’. The subject has a very quick learning curve and I think you too, will recognize the potential.

@pipsqueak234 first of all welcome! To cut right to your question you are describing pairs trading and the most optimal correlation for pairs trading is 0.35 to 0.85. If the correlation is too low then you have random movement between the two and no real trade, and if the correlation is too high then you also don’t have much of a trade as they will track each other too closely.

But pairs trading in Forex is kind of redundant because currency is already paired. For example, if you noticed that AUDUSD and NZDUSD had a 0,5 correlation so you wanted to do a long AUDUSD and a short NZDUSD trade, you are really just trading the AUDNZD basket. I just wanted to get that out of the way, since the vibe I got from this post is that you were pairs trading currency pairs.

As far as time frame, that is completely dependent upon your strategy. Usually in pairs trading you are trading extremes and hoping for a bit of mean reversion, which if that is your strategy you may be holding for days, weeks or months waiting for the reversion to occur, but there isn’t any reason why you couldn’t also scalp a pair on the 5 minute charts if that is your strategy.

Only you can know what the right time frame is.

1 Like

I would suggest that selecting your Entry TF has at least as much to do with your own lifestyle as it does with trading. At the end of the day, a chart is a chart, Price is Price, you’re just looking at it through different windows when you select different TFs. You’re manipulating your view of data to suit you, but the underlying data is still a Price which you’re trying to predict.

Personally, when I was starting out, when my kids were small, and when I was finding my feet, I liked purely EOD trading, looking at the Daily chart for Entries and entering via overnight orders. That is still a large part of my trading. I monitor trades on the TF on which I chose my Entry, so those are low maintenance (although need bigger Stops). If I want to do some lower TF day trading then I will look at a lower TF, 5 minute or Hourly.

In terms of the correlations part, I have avoided that bit as I just avoid correlated Pairs - don’t want to back the same move twice in case I’m wrong!!

But I hope that the first bit helps.

ST

2 Likes

Hi Krugman, and thank you for your welcome.
Yes, I completely agree that to select ideally correlated trades, you are looking towards the extremes of the forex pairs relationship. This relationship is, as you mention, expressed as the bottom and top of a scale, with the behaviour of forex pairs in the middle being regarded as too random, so pairs below .25 or above .85 are the candidates to look for. Incidentally, to avoid any confusion, many of the broker sites that provide correlation data express their evaluations in a range from (minus) -100 being max. negative correlation to +100 for max. positive correlation. So forex pairs in the extreme range, +/- 80 or greater are a good starting point.
You are quite right that I am pairs trading (I refer to as correlation trading) currency pairs e.g. EUR/USD (Short) against GBP/USD (Long), but that’s deliberate. The USD element is cancelled out, as you suggest, and the same effect is obtained by just trading the EUR/GBP. I know this would avoid having to pay the spread twice, but with the profit potential, I’m happy to accept that and monitor the position on a chart overlaid with the two currency pairs so that I can see exactly what’s going on.
Regarding the question of what time frames to use I think you have confirmed my own feelings that there’s no right or wrong answer when you say ‘Only you can know what the right time frame is’ depending on personal preference for short or longer term trading.

Simon Templar’s response (Thank you Simon) also agrees that it’s all down to personal choice. Your observation Simon of not wanting to back the same move twice is understandable but this isn’t what I’m up to, because I’m buying one and selling the other unless I’ve misunderstood.

For a little more clarification, the sort of selection criteria I’m using is:

  1. Downloading correlation data into a spreadsheet with 1Day,4HR, 1HR and 5MIN data, sorted into 4HR correlation order. I can then look for the highest scoring candidates with reasonable consistency across all timeframes.
  2. I use two prepared profiles in MT4, containing historically recognised positive correlation pairs in one and negative correlations in the other.
  3. There are also some broker sites with correlation data in matrix form.
  4. Once a prime choice is made, then look at heatmaps for the original pairs being considered to make sure they’re not heading off in an outlandishly wrong direction.

These methods are used really as confirmation of each other and if they all concur then I feel I’m on the right track.
I think it’s been confirmed that there’s no right answer or magic to selecting a trading timeframe with better performance than any other timeframe. I’ll just keep demo trading to build a track record that will eventually overcome my risk aversion so that I take on a live account and make a fortune. (What are those pink things with curly tails flying past?)

I’m going to try to copy a couple of trading statements over around 2 week periods and another of two trades (i.e. 2 pairs of trades – 4 trades in all) which are presently 27 pips profit. These will show what I’m up to.



1 Like

We might have misunderstood one another, but that does sound as though you have your money tied up twice in the same move. Which to me would double the risk (unless I took half risk with each trade). But anyway, many ways to trade so glad that it’s working for you.