It is said that the market spends most of the time reverting. If this is true, then why is there so much focus on trend strategies? By focusing on Trend strategies we miss out on most of the opportunities the market presents us.Because of this, I recently have come to decision that reversion strategies are the way to go.
Reversion strategies are more common than you might think, even if not actively traded. For example, many trending strategies employ reversion theory to pick entries, for example, waiting for the price to pull back to a Pivot or Fibonacci level to enter.
The Forex market will revert eventually, you just don’t know when. You need to be in the market to take an advantage of it, but indicators are generally too slow to tell you when to enter. Chances are you’ll probably end up paying more commission than profits!
I believe for a short-term reversion strategy to work, one needs to average down their position by scaling into the losing position, bringing the net position closer to the current price and hence ready for that price reversion which may come at any time. Testing this has shown dramatic improvements in the metrics. However, if you do any research on averaging down, scaling in, scaling down it generally is negative.
No matter what you call it, you are basically betting more on a losing position, and you keep betting more until you win. This is kind of similar to some Grid strategies and Martingale strategies I have discussed in the past, and it certainly has been the recipe for disaster for many investors in the stock market where a price can crash almost to zero and never return …. Forex though is a bit different isn’t it? Ok yes there are emerging markets and unstable currencies like the Russian ruble, but when was the last time USD crashed to zero?
It is said that Jim Rogers and Warren Buffet often average down their positions. If that is true then it can’t be that bad? However conversely Paul Tudor Jones, one of the best traders of all time, is pictured below with one of his key ideas stuck on the wall behind him – Losers average losers.
Well, why can’t they all be right? Tudor Jones was mostly a trend trader which is not what we are talking about here. We’re talking about taking advantage of short-term reversions in the market, and there is a right way to scale into a losing position and there is a wrong way.
The easiest way to blow up your account averaging down is to scale in too aggressively, like a martingale strategy. This is because the larger the position we already have, the even larger position we would need to average down, meaning you very quickly either run out of money or margin. So, let’s try to intelligently come up with some formula to use to scale in.
The scenario is we entered into a position and the price did not go in our way. However, we expect it to eventually retrace so we want to keep entering and average down our net position. Let’s test two ideas, the both scale in every n pips but
– one tries to maintain it’s net position at a ratio of the current price slide (Percentage of Price Change)
– the other tries to maintain it’s net position at a ratio to the previous net position (Ratio since previous Net position)
So, plotting these into a spreadsheet we get the following.
So three columns. The first dictates the price details, the middle shows the first idea, and the third shows the second idea. We can see that as the price increases by 10 pips, we decide to enter again. The first idea is targeting to keep the net position at 50% of the price slide, which is confirmed by the Dist column. After 10 pips we are 5 pips down, after 100 pips we are 50 pips down etc. The second idea is attempting to ratio itself at 50% of the distance since the last net position… So after 10 pips we are 5 pips down, after 20 pips the price is now 15 pips away from the last net position so we want to be half of that and end up 7.5 pips down only etc…
The results show the first idea is linear. Pretty safe, we keep entering in the same amount every time since our trade entry (every 10 pips) is also linear. This is okay, except the distance our net position is to the current price is increasing. Perhaps a 100 pip retracement might be a bit too much of an ask? The second idea is better, it ends up being about 10 pips away regardless of the size of the slide… though we can see that this idea approaches martingale and we end up doubling our position every time – oops.
But what if we didn’t target a 50% of the price change (for either ideas), and targeted something else? Well I looked at that too…
Note the scale on the charts…. For idea one, if the ratio is less than 50% then good news! Our trade size decreases exponentially! However, if our target is greater than 50%, bad news…. Though, I guess this depends on your strategy. If you are expecting the price to pull back 70% then that’s great for you, but if you only expect the price to pull back 20% then I hope you have deep pockets.
What about idea 2?
Well, that does not look promising. Let’s think about something else.
Since the trade size needed starts to get a bit crazy as the slide deepens, why don’t we space out our re-entries instead? Well, let’s take a look.
The above is the same spreadsheet we started with but after a few entries we start determining our entry price by a factor of the slide. In this case, after the 7th trade we start shifting out entry price by 20% of the price slide at that time. For idea 1 we end up breaking our linear pattern and adopting the exponential pattern we were trying to avoid. For idea 2 the trade sizes actually are exactly the same, but our distances widen, which is the expected behavior.
While we have spaced out our entries and reduced the rate that our positions increase, we have not really avoided the problem, only prolonged it. Honestly, after a 1000 pip slide, are you really going to expect it to retrace 500 pips? It might, but are you willing to bet millions of dollars on it Even if you did just hang in there, there would have been hundreds of potential mini-reversions you could have taken advantage of while your capital was stuck in that trade. Maybe it is a better idea to give up after a slide of a certain size?
Regardless of what the above has shown us, we have to remember that the maths world is not the real world and we should throw these ideas into a strategy, back-test a lot of data and see if they do actually work or not.
However, we cannot predict the unpredictable without infinite back-testing data so regardless of whatever approach you decide to take, if you do start exposing large positions to the market, it is always an idea to have some kind of safety in place just in case the unthinkable does happen and the United States Dollar crashes to zero.