Before we get start digging into this simple pin bar system, be sure that you understand the limitations of backtesting.
Past performance does not guarantee future results and all testing is theoretical.
…but it does give you a higher probability of being right, when real money is on the line.
That being said, I wanted to share some results of my latest test.
As I always say, don't take what I'm saying for granted, test it for yourself. This is also not the final version, so this is a work in progress.
You may remember that I tested a simple system before. But that was a very rudimentary test. This test will go deeper.
This method is easy enough to test on almost any manual backtesting software, or even scrolling backwards (then forwards) on MT4.
Alright, here we go…
The Entry to the Simple Pin Bar System
I'm looking for a Pin Bar that has a tail of 61% and a 14-period RSI in the corresponding overbought or oversold territory (default 70/30 settings), at the close of the candle.
I'm also looking for a countertrend move before the Pin Bar (examples below). The Pin Bar also has to be the highest/lowest point in the move. The idea is to find an area where price has been stretched in one direction and will hopefully snap back.
Finally, I'm entering the trade on the close of the Pin Bar and setting a 1R take profit. That may be a little hard to picture (I know it is for me), so let's see a couple of examples.
Long Example
For a long, you look for a Pin Bar (marked by my Forex Tester 2 indicator) and the trend, in the bearish direction. There is also an oversold condition on the RSI.
Short Example
A short trade is just the opposite, of course. Here is an example:
The Stop Loss
Stop is set 15 pips above/below the Pin Bar. This might be too tight of a stop, but let's give it a try.
For this test, I'm risking 2% of the account on each trade.
The Take Profit
I'm looking for a 1R profit target on this round of testing. So if I'm risking 30 pips, I'm targeting 30 pips. For simplicity, I'm not taking the spread into account.
There is probably a better TP, but I'm starting simple.
The Results
Alright, now let's get down to the results. I did 8 rounds of testing on the AUDJPY daily chart, from January 2001 to present. Here are the results:
Not bad. But 20% over 14 years?
That probably isn't going to be profitable enough, especially when risking 2% per trade. This means that I have to come up with a way to get more out of each trade.
But the good news is that a 74% average winners is a great place to start. However, there is another thing to consider.
In testing, there were only about 20 trades per test. This is an example from one test.
So 20 trades over 14 years, probably isn't a big enough sample size to get a really good idea of if this is a viable system or not.
You might also notice that there is some difference between the tests.
This is why we test.
Just like with real trading, you will miss some trades. You might be sleeping, on vacation or just not completely focused. Multiple tests helps us simulate these conditions.
This also shows how concrete your trading rules are. If there is a large variance in the results, this shows that your rules might not be easy to follow in real trading.
Possible Improvements to the Simple Pin Bar System
Given these results, I could test four things next:
- Setting two profit targets, one at 1R and one at the next S/R level
- Test on more pairs and on the H4 timeframe
- Looking for a minimum 2R on the first profit target
- Only look for fake pin bars
Conclusion
I really like Pin Bars and I'm going to show you the entire testing process for this system. My goal is to figure out a system that works for me and I will track the method all the way from testing to tracking live trading results.
I haven't seen anyone do this before, so I hope that this will be a first and will ultimately help you understand the guts of a trading system.
Remember that no technical pattern exists in a vacuum. You always have to watch news and previous price action. But basic testing like this establishes a baseline of probability.
