When I was a junior in college, I wrote a program for the Intellivision that played backgammon. One of the odd issues that came about was whether the program cheated. The company was working on a lot of other game cartridges, and we programmers routinely asked each other to test out the games.
These guys were really smart, but none of them had ever played backgammon. Backgammon is a game that relies on the roll of the dice, and this ensures that even novice players can sometimes beat a more experienced player. So, my colleagues got beat once in a while when they played my program, and it stirred up a certain amount of resentment. Finally one of them went to the owner of the consulting company that employed us, whom I will call G.
G was worried that Mattel would get word that the game cheated, so he decided to look into it. He sat next to me for an entire day, and wrote down every dice roll that the game produced. I was using a linear feedback shift register algorithm, modified to be used for backgammon, and (of course!) I knew that the game did not cheat.
At the end of the day G applied a few statistical tests to the dice rolls, and came to the conclusion that there was no merit to the complaints. The fans were not happy, but, hey, you win some, you lose some.