Murphy's law comes from the fact that some technician didn't know what he's doing and the engineer Murphy, said "If there is a way to do it wrong, he'll find it". The law was then immortalized from then on.
As to the outcome that Rand Johnson is referring is the Law of Large Numbers. In other words, if you take that there is a particular probability that something can fail, then it will eventually fail when doing a series of experiments, although there is no given time frame when that problem will occur. Take this in reverse, if there is a probability that something will can fail, then there is a probability that something will go right. Hence, you can apply the same logic and expect something right to occur under the same experiments.
The general idea is that the event of failure is usually of low probability and so the expected result is success. Hence, Murphy's Law, "if it can go wrong it will."
Although, do not be fooled into the Gambler's fallacy in this statement. Just because you have flipped a fair coin on heads 500,000 times in a row, does not mean that your next flip has any better probability of getting a tail than before. The probability of flipping a tail is still 50% as before. But eventually if the coin is fair, you will flip a tail.