The ‘two steps back one step forward’ trading strategy
In this article I am going to go over a possible trading set-up which I’ve seen recurring on my charts and appears to provide fairly reliable buy and sell signals in the direction of the dominant trend.
I have decided to call this set-up the “Two steps back one step forward,” – or 2-1 for short – in a humorous play on the real phrase “two steps forward, one step back” – meaning to progress with setbacks, but to progress nevertheless.
First I am going to show you some examples of the set-up and then I’m going to provide the results of an initial back-test.
The set-up is simple and consists of a trend, whether up or down (it seems to work equally well in both) followed by a pull-back for a duration of two bars. There then ought to be a higher probability that the next day will be a resumption of the trend. In the case of a bull trend it should be an up-day, and in the case of a bear trend a down-day.
In the initial version of the set-up the trader opens a trade at the open on the day after second day pf the pull-back, and then closes it at the time of the close.
The set-up seems to work on the principle – noted by the eminent technician W.D Gann, that there is a world of difference between two and three. According to Gann three up or down days in a row was a strong sign of reversal. Given reversals are rare it logically means that two is more likely to signal a continuation.
Theoretical consideration aside, the chart below shows examples of the set-up which should clarify what I am trying to describe in words.
In the chart above we see an up-trend and then a series of 2-1 set-ups. As can be seen prices start to rise and then form a shooting star and pull-back in set-up A; the pull-back lasts for two bearish days; then according to the rules for our set-up we should see a resumption of the up-trend on the following day. In the case of A that indeed happens and the ‘third’ day, which is circled, is strongly bullish. If we traded it by opening a long trade at the open and closing it at the close we could have made about 80 points profit.
The market continues higher until we get to the next pull-back and set-up B. Once again the market corrects for a duration of two days, before rising on the third day – as predicted by our strategy.
There are then several more pull-backs as prices continue higher, up to the 1.39s but none last two consecutive days, so they don’t qualify as valid set-ups.
Then after the 1.39 peak there is quite a steep sell-off which lasts two days and leads to set-up C. According to our strategy, even though there have been two steep down-days in a row the next day should be an up-day – and indeed this turns out to be the case.
The trend then becomes less clear and so I have not included the other potential set-ups – as increasingly they start to consist of three down-days, and the strategy begins to fail.
That is the set-up covered. Now its time to back-test how well it works so we can generate some statistics and quantify the edge it could give us as analysts or traders.
The results imply that the set-up provides a statistically significant trading edge.
Tested on the daily chart of EUR/USD between 2008 and 2010 a total of 60 qualifying set-ups were found, with the set-up correctly forecasting the next day’s direction in 40 out of 60 times in the sample. In the other 20 the set-up incorrectly forecast the next day.
Using binomial statistics to test whether the results were significant it was found that the margin or error was only 3.0%, which is higher than the 5.0% required in normal scientific laboratory tests.
Although the win rate was relatively high, the number of points won overall was not as spectacular as might be expected. The study showed that out of a total of 5919 points, the 40 winners made 3675 points whilst the 20 losers accounted for 2244 points. This gives a net gain of 3675 – 2244 = 1431. This does not include trading costs, or the spread, which at 2 points per trade would add up to 120 points.
An initial study into the profitability of the 2-1 set-up on EUR/USD between 2008 and 2010 seems to have shown that the pattern is a good forecaster of price action, with a success rate of 40 out of 60 and a failure rate of 20 – giving a 66% success rate.
The results are statistically significant, with only a 3.0% margin of error.
The result further showed that if traded the strategy would have proved profitable with a net gain of 1431 profits over the 60 trades (not including trading costs).
One major caveat is that no objective method for defining the trend was used, merely the testers own discretion, based on common sense, and a heuristic approach, which tended to throw out set-ups which erred on the side of not being in a trending market.
Overall, however, it appears the 2-1 is an accurate and profitable predictor of price action.