Hello
This is just an idea i've been toying around with for some time, and would like some feedback. I know it is not a new idea, but im just experimenting with non directional trades.
So here goes nothing:
EURUSD and USDCHF have a negative correlation of -97.7 using daily closes for n = 50.
If i remember correctly, -100 is perfect negative correlation, so, it would be safe to assume that when EURUSD is bull, USDCHF is bearish, and vice versa.
Overlaying daily line price graph of EURUSD over USDCHF (easy to do with FXCM's platform ) will easily show that EURUSD & USDCHF move in opposite directions.
Now, if we were to take a directional trade, it would be reasonable (and risky) to assume that if EURUSD & USDCHF are converging, going long with one pair while going short on the other would be profitable. In the event that we chose the wrong pair to go long and the wrong pair to go short, a stoploss in both pairs would be required.
Again, if the pairs were divergent, it would be reasonable to assume that a convergence would at some point occur, hence going short on one pair, and long on another.
But this hypothesis has one flaw: if the pairs do not converge when away, we lose money. Same applies for the opposite, if convergent we cannot know for sure when they will diverge (although technical analysis on both pairs would help).
So here comes the hedging.
What if we placed 2 short orders. Assume pairs are divergent, and sooner or later, they would meet at their axis. If we go short on both pairs, it would be impossible for both to go long, and both go short. So essentially we are betting that the pairs i will converge, but we are hedging our loss from the pair that goes long, in order to meet the pair which has gone short.
Making sense?
Now here is the tricky part.
EURUSD and USDCHF do not have the same price, do not move the same distance, meaning they have different voltility.
So we would have to hedge the negatively correlated pairs, but assume that we will be making a small profit from the pair that is most volatile, while hedging with the pair that is less volatile. A quantitative formula might help here, but I am clueless on how this could be calculated.
So, by empirical data (ATR indicator), i realized that EURUSD (now at 1.4679) is more volatile than USDCHF (now at 1.0301) and because USD has been long oversold, and because for various technical indications i believe that EURUSD is about to fall, which again implies that USDCHF will rise, i chose to go short EURUSD and at he same time go long USDCHF. Their negatively correlated movement implies that their movements will cancel eachother out, leaving room for a small profit from EURUSD if indeed it moves faster than USDCHF. Of course that small profit could be amplified by using large lot sizes.
And as a fail safe, a generous Stoploss could be used on both pairs, just to be sure.
So, ideas, thoughts and feedback?
This is purely hypothetical, it is not my idea (correlation & hedging with regards to forex have been around for long), but i am wiling to test it and fine tune it to see if a small profit can be made.
This is just an idea i've been toying around with for some time, and would like some feedback. I know it is not a new idea, but im just experimenting with non directional trades.
So here goes nothing:
EURUSD and USDCHF have a negative correlation of -97.7 using daily closes for n = 50.
If i remember correctly, -100 is perfect negative correlation, so, it would be safe to assume that when EURUSD is bull, USDCHF is bearish, and vice versa.
Overlaying daily line price graph of EURUSD over USDCHF (easy to do with FXCM's platform ) will easily show that EURUSD & USDCHF move in opposite directions.
Now, if we were to take a directional trade, it would be reasonable (and risky) to assume that if EURUSD & USDCHF are converging, going long with one pair while going short on the other would be profitable. In the event that we chose the wrong pair to go long and the wrong pair to go short, a stoploss in both pairs would be required.
Again, if the pairs were divergent, it would be reasonable to assume that a convergence would at some point occur, hence going short on one pair, and long on another.
But this hypothesis has one flaw: if the pairs do not converge when away, we lose money. Same applies for the opposite, if convergent we cannot know for sure when they will diverge (although technical analysis on both pairs would help).
So here comes the hedging.
What if we placed 2 short orders. Assume pairs are divergent, and sooner or later, they would meet at their axis. If we go short on both pairs, it would be impossible for both to go long, and both go short. So essentially we are betting that the pairs i will converge, but we are hedging our loss from the pair that goes long, in order to meet the pair which has gone short.
Making sense?
Now here is the tricky part.
EURUSD and USDCHF do not have the same price, do not move the same distance, meaning they have different voltility.
So we would have to hedge the negatively correlated pairs, but assume that we will be making a small profit from the pair that is most volatile, while hedging with the pair that is less volatile. A quantitative formula might help here, but I am clueless on how this could be calculated.
So, by empirical data (ATR indicator), i realized that EURUSD (now at 1.4679) is more volatile than USDCHF (now at 1.0301) and because USD has been long oversold, and because for various technical indications i believe that EURUSD is about to fall, which again implies that USDCHF will rise, i chose to go short EURUSD and at he same time go long USDCHF. Their negatively correlated movement implies that their movements will cancel eachother out, leaving room for a small profit from EURUSD if indeed it moves faster than USDCHF. Of course that small profit could be amplified by using large lot sizes.
And as a fail safe, a generous Stoploss could be used on both pairs, just to be sure.
So, ideas, thoughts and feedback?
This is purely hypothetical, it is not my idea (correlation & hedging with regards to forex have been around for long), but i am wiling to test it and fine tune it to see if a small profit can be made.
Attached Image (click to enlarge)
