In the whole paper, there's no mention of whether the market itself was growing or shrinking. This makes the results kind of hard to trust. Assume the market lost money for these dates, then it would follow that most traders on that market lost money as well. The results should at least be controlled for the overall rate of change of the market.
I don't think that's how day trading is supposed to work. Successful day traders are supposed to make money whichever way the market is moving, by taking short positions when the market is going down and long when the market is up. Buy-and-hold investors (quite the opposite of day traders) are the ones who make money when the market as a whole goes up.
