Well my take on the saying is as follows (and I posted it so my ideas count for more than others
If I toss two coins in the air and one of them shows a head, what is the probability that the OTHER coin shows a head? Many people think that the THEORY is that the other coin has only two choices, so the chances are 1 out of 2 (or 50:50 for some of you). However, do an experiment, by tossing a coin lots of times, (and ignoring all the cases where both of the coins do NOT show a head), and you will find that the other coin only shows a head 1/3 or the time. I.e., the PRACTICE shows the initial theory to be wrong. The same situation occurs in a game show where the winning contestant has to choose between three doors, one of which has a big prize, the other two booby prizes. As the winner, you choose, say, door no 1. The host now tells you that the prize is NOT behind say, door 3. Of course there is always a door he can say that about, and he knows where the prize is. What should you now do? There are two doors, one of which has a prize, one a booby. Should you change your choice, or stay with the original, or does it not make any difference? Well you are twice as likely to win the prize if you CHANGE your choice than if you stay with the original selection. PRACTICE will prove me right, no amount of (simple) theory will help you get here, unless you have a deep knowledge of Probability.
And I am sure that many of you have had a graduate student come to work for you who knows EVERYTHING in THEORY about your operation, except that he has no PRACTICAL knowledge of what really happens...