Recently an occasional visitor to my blog, whom I only know as "Andy," left a comment inquiring about an old Forex carry trade project I was working on, a strategy that involves going short the Turkish lira versus the U.S. dollar, and hedging that with a long USD position against Canadian dollars at a ratio of 4:3 USD/TRY short to USD/CAD long, which keeps the per-pip value of the respective positions roughly equal. The basic background of the idea can be found here.
I set up a game account to model this approach in and started it off with $1000 and set the account to 50:1 leverage (my broker offers free game accounts for practicing Forex trading - the data is real, the money is not, so you risk nothing while learning; try it by clicking here if you are interested). My initial setup was to use 50% of the account balance for the USD/TRY short, which is 12,500 units in an account with $1000, and the USD/CAD long hedging position was thus 9,375 units (75% of the number of USD/TRY units). This resulted in used margin of $687.50 - a very tight fit in such a small account.
One of my initial concerns with this approach was something I refer to as "divergence." For the most part, USD/TRY short and USD/CAD long follow each other, so as one rises the other falls, and vice versa, and their roughly equal per-pip values keep the account from getting out of whack and hitting a margin call. There are times, however, when the two pairs do not track each other very well, when they instead both slide into their respective negatives. A difficulty of creating one of these setups is that it's hard to know with certainty if you are creating it in the middle of such an event, and the first attempt may very well have fallen victim to this - the unrealized loss in the account quickly swelled to -25% or so (this is why using game accounts to test new ideas is so important!). The first model was launched April 23rd, and to compare that one with the same setup initiated at a different point in time, a second one was set up in a different game account on May 21st, with the intention being to see if there was any difference in terms of unrealized profit and loss trends (swings in the net asset value) between the two accounts if created exactly alike but at different points in time.
A few weeks later I hit upon a successful active trading style and quickly forgot about this project. But, a recent comment left for my by Andy asking how it was going prompted me to open up the game accounts and have a look. I was expecting to find two game accounts that had blown up and failed; instead, I found two accounts that have been chugging along steadily and quietly this entire time. Both the May 21st and April 23rd USD/TRY/CAD models are still going strong, each are accruing around $2.40 in interest every day, which based on the initial deposit amount is a mind blowing 87.6% annual interest rate, and during this week I have observed their unrealized profit/loss percentages swing from -23% during weekend off hours to as low as -4.5% during the busiest part of the trading day. What's also interesting is that despite having been started at different points in time, the unrealized P/L percentages of the game accounts stay within 1% or less of each other.
Nearly five months on, the first account has accrued approximately $360 in interest, or nearly half of the capital at risk ("at risk" in this sense meaning the amount of the original deposit that would be lost if the account received a margin call under this setup). At the ten month mark, and assuming that one draws out all interest generated by the account leaving no additional balance in the account to potentially be lost to a margin call, all cash that could be lost under this strategy would have been recovered, and everything beyond could be considered pure profit (as some would say, you're "playing with the house's money" beyond that point). Seeing the success of this model thus far, I'm tempted to begin building a real one, but I will do so only with pocket change as the models are still what I would consider "young" given the sort of time horizon this strategy requires, which is practically infinite. So, Andy, if you're out there, this is how it has turned out thus far. Thanks for asking!
0 comments:
Post a Comment