You're more likely to notice this if you picked a queen or a jack, of course.
The trick manipulates "change blindness" - our failure to accurately register notable changes in our surroundings. Or, in the case of this trick, any card other than the one you picked.
If you noticed that there was a king, two queens and two jacks, whilst picking any card other than the king, you're more likely to see through this trick.
The trick has also been used in studies to show that most times we subconsciously register which cards were present before.
In research conducted by Luis Martínez of the Institute of Neurosciences in Alicante, Spain, the team found that, after a performance of the trick, when forced to choose between two cards as to which was in the first hand, 80 per cent of the time a participant answered correctly. As New Scientist points out: "If they really hadn’t been aware of the other four cards in the first set, their score should have been only 50 per cent – no better than chance."