This paper looks at a contrarian trading strategy which takes a classical technical analysis set-up and turns it on its head, trading it in the opposite way.

The Japanese hammer candlestick is a famous reversal candle. It occurs when the market is in a down-trend, opens a new day and falls even lower. A recovery later on means that it closes within the top half of the candle again. Received wisdom says the hammer signals the start of a new bull-trend.

The ATG strategy takes this set-up and trades in the opposite direction expected. The trader bets on the hammer failing. He or she places a sell short order just below the hammer’s low. If  the bounce triggered by the hammer rolls over it will trigger the trader’s order at the lows. As illustrated below:

One advantage of the strategy is that if the trader is proved wrong and the market does successfully rally the lows of the hammer where the order lies usually remain intact ensuring the trader is often left out of potentially losing moves.

Shooting Star variation

The shooting star presents us with the inverse of the hammer. This candle occurs at the top of a rally and looks like an inverted hammer. Instead of bullish, it has bearish conotations. It can also be traded using the contrarian strategy outlined above, except that instead of placing a sell short order below the lows the trader places a buy order just above the highs, expecting prices to go higher.

Psychology

The psychology of the set-up relies on the panic selling/covering which occurs when the market runs contrary to expectations. In this case the appearance of a hammer – in the case of a downtrend – creates expectations of a reversal and a then a rally and  investors may be tempted to go long. However, when the rally runs out of steam and prices roll over their hopes start to fade. Nevertheless many may be tempted to hold on to their longs in the hope of a recovery, particularly as there was a hammer at the lows giving them confidence that a reversal was taking place.

After the hammer’s lows are breached they are catagorically proven wrong and often  liquidate their postions in a panic, which accelerates the already swift down-trend. It is at this point that the strategy triggers its sell-short orders. The panic selling adds momentum to the move and increases the strategy’s profitability.

Variations

Only the basic set-up and entry-point for the strategy has been covered so far. We have not yet discussed stop-placement and profit-targets. There is also the small problem of defining the trend and how you do that in practice as well as the best timeframe to employ.

It is up to the individual investor to optimize the strategy. The final risk/reward parameters will be dictated largely by the security used and the time-frame employed.

I have used data from the EUR/USD currency pair in my research and found the strategy profitable employing a simple 2:1 risk reward using the range of the hammer/shooting star to take profit and place stops. I found that by subtracting the hammer candle’s range (length of the candle from high to low) from the lows of the hammer’s low to give a profit target and adding the range to the high to give the stop-level produced a successful approach.

As for how the trend is defined, I have found that on a daily chart using the market activity of the previous 3-days as a guide can usually suffice. For example, the trend would be defined as up if the close of the prior bar to the set-up was higher than the close of the bar 3 days before.

On intraday charts I have found a longer period gives a better defination of trend – for example with the hourly chart a length of 5-bars works better than 3.

Moving Averages could also be used to define the trend. Other techniques such as hand-drawn trendlines or discretionary indicators can also be employed if the trader wants a more operative style.

Sample Results

Using the range based stop and profit-taking levels outlined above I tested the strategy over the last 10-years on the EUR/USD currency pair and discovered it yielded an impressive win rate using daily data.

Out of a total of 101 trades, 79 were winners and only 22 were losers. Using the stop and take-profit levels suggested above gave a  2:1 risk/reward ratio and despite this inverted risk/reward it still yielded good returns. One criticism could be it did not generate enough set-ups to provide a stand-alone trading solution, although it could compliment other trading strategies as part of a combined approach.

The strategy was also tested with AUD/USD over a 10-year span using the same criteria although it was less successful than with eurodollar, with 56 winning and 30 losing trades. The 2:1 risk reward ratio meant that the strategy was not profitable when applied to the AUD/USD.

Intraday

I tested the strategy intraday on EUR/USD, using an hourly chart and a 5-hour close comparison to define trend. In a period during 2002/03 I found that the strategy produced 87 winners and 36 losers and was marginally profitable.