Back in the days when I read slush I could reject manuscripts at a rate of about 20-25 an hour. It doesn't seem surprising to me at all that if you are looking at these papers all day long, it's not going to take very long to learn how to quickly weed out the obvious nonstarters.