Preliminary backtest results on new USDJPY system

I'm currently backtesting a new system based on my research here. In my last entry, I observed that the USDJPY's daily high or low was most likely to be established in the first four hours of each day.

Suppose we looked on the 4H chart and traded the break of the first 4H candle that appeared for the day, with our stop loss located on the opposite end of the candle?

Example setup:

I basically looked for small 4H candles at the start of each day (a 'small' candle being defined as a candle with a range less than 0.75 * ATR(30)). If we're going to trade the day's move, we want to be entering just BEFORE the move, not after, hence why I only looked to trade the breakout of small candles. If the candle's large, then the day's move has probably already started and you'll be entering late. 

My preliminary backtest consisted of 534 trades from April 2009 to yesterday. Where I set my profit target didn't seem to matter. All TPs showed profit. The best TP though, was 1.25R. This yielded a profit factor of 1.51, which is very good. 

This is the equity curve with a $25k starting balance, at 2% risk per trade. 

R^2 is very good, at 0.93. The max drawdown was 21%. 

I tested most of the data between April 2009 to yesterday. I hope to finish the entire backtest in the next few days. The signs are looking very good, though.