I'm testing a very simple strategy from this research paper: Intraday Foreign Exchange Reversals

The backtest range in the paper is 2003-2005 and the results look promising but when I tested the strategy on 2004 and 2005 data (I don't have data for the whole 2003) I got much worse results than the authors. They included spread and commission which should be larger than fees I'm working with. This is the corresponding line from my asset file:
Code:
EUR/USD,1.12000,0.00005,-0.418,0.0560,0.0001,10.00,2236,0,100000,0,


So my question is - can anyone else here check the paper and my implementation if I overlooked something? Maybe even try it on different data? (I use Dukascopy for testing).

Here's the code:
Code:
var sessionStart = 10;
var sessionEnd = 15;
var period = 4;

function tradeIR()
{
	vars closePrices = series(priceClose(), period);
	if(lhour(ET) >= sessionStart && lhour(ET) <= sessionEnd){
		if(NumOpenShort == 0) {
			if(priceClose() > SMA(closePrices, period)){
				enterShort();
			}
		}
		if(NumOpenLong == 0) {
			if(priceClose() < SMA(closePrices, period)){
				enterLong();
			}
		}
	}
}

function run()
{
	if(INITRUN) {
		slider(1, 0, 4, 12, "m", "Period for SMA calculation");
		slider(2, 0, 2003, 2016, "Start", "Start year");
		slider(3, 0, 2003, 2016, "End", "End year");
	}
	set(EXTRADATA+LOGFILE);
	MonteCarlo = 0;
	//History = ".t1";
	BarPeriod = slider(0);
	period = slider(1);
	StartDate = slider(2);
	EndDate = slider(3);
	
	tradeIR();
}


And my results for EUR/USD, years 2004 and 2005:

2004:
MA length = 4: Trades 524, AR 4%, SR 0.04
MA length = 5: Trades 460, AR 82%, SR 1.09
MA length = 6: Trades 428, AR 15%, SR 0.19
MA length = 7: Trades 376, AR 7%, SR 0.09
MA length = 8: Trades 372, AR 6%, SR 0.07
MA length = 9: Trades 352, AR 62%, SR 0.72
MA length = 10: Trades 337, AR 74%, SR 0.78
MA length = 11: Trades 325, AR 79%, SR 0.81
MA length = 12: Trades 315, AR 47%, SR 0.53

2005:
MA length = 4: Trades 493, AR 26%, SR 0.46
MA length = 5: Trades 435, AR 53%, SR 0.87
MA length = 6: Trades 397, AR 25%, SR 0.48
MA length = 7: Trades 359, AR -9%, SR -0.17
MA length = 8: Trades 345, AR 5%, SR 0.12
MA length = 9: Trades 337, AR -3%, SR -0.05
MA length = 10: Trades 327, 1%, SR 0.04
MA length = 11: Trades 310, AR -11%, SR -0.23
MA length = 12: Trades 314, AR -9%, SR -0.16

Results from the paper:


The strategy is described mainly on pages 3 and 4 of the linked paper.