I am attempting to implement a strategy that goes both long and short on gold. But going long on gold, I will be long on GLD ETF. However, I cannot find anything which is as efficient as GLD for going long The inverse ETF DGZ (-1x) has got very low volume and the bid-ask spread are quite wide (0.1-0.2%) GLL(-2x) is better , but then if I have to go short for a couple of weeks, the expected returns will not match because it is rebalanced daily. Going short on GLD using the broker (e.g. interactive brokers) will attract margin loan which is eating significantly into the profits that I'm calculating. Given that the USD and Gold are mostly inversely correlated, I wonder if I can go long on USD. I have not analysed if the inverse relationship is also on the short term (on an hour basis)
In my mind, the idea is to capture a direction, which ever way it goes. Then you wait for the position to suddenly reverse, so you close the opposing leg and let the other leg continue. It's a gnarly strategy which basically is the same as calling tops or bottoms in a unidirectional trade.
Unless the person doing this comes from a country with a different tax code for this, you can do the same being flat and entering traders at the point you want to add risk. Makes no sense for those from the US. NONE!
But that is just it, OMM. The people doing this sometimes have zero clue on which way the market would move. So they take the net zero approach, and when they are satisfied the market has momentum behind it in a direction, they close the opposing leg and let the profitable leg continue, eventually closing at a +target.
And how is that different in P/L from coming in flat and then reacting buy getting long or short? It's just not.
Because it gives you the benefit of letting time and the fundamentals skew the direction of the trade. You're in, you do not know where it is going. The FA hits and the market reacts. You BE when it hits and profit a bit on the momo of the news. That is the ONLY way I see it working. I never said it was a smart strategy, and it certainly cannot be automated. I'm just answering your question as to why someone like the OP may try to do it.