First, you observe something about the universe.
Then you try to explain that observation by creating a hypothesis.
Next, you create an experiment, and repeat the conditions that you think will produce the same result.
If you run the experiment, and you get the result you want, your hypothesis has been (so far, anyway) proven correct.
HOWEVER, if you run the experiment, and you get an unexpected result, your hypothesis is incomplete, and you must modify it to account for the new aberrant observation.
I saw Kevin Henson hurt himself once.
Later, I saw it again.
I began to pay closer attention, and noticed that these accidents invariably coincided with him attempting to prove physical prowess by performing some dangerous feat.
…so I ran the experiment…
I said “The next time Kevin shows off, he’ll hurt himself.”, and let the experiment run.
A few dozen samples later, my hypothesis had been proven right.
Kevin Henson only hurt himself when he tried to show others how smart and/or physically capable he was.