You and your family are trying to decide where to eat for dinner. There are three new restaurants nearby; two are nearly empty and one is crowded. Do you go to the one that is crowded, assuming that it might be better because it is popular?

If you do, Ivars Peterson thinks he has a surprise for you. He presents this story: suppose that the first person to arrive at the restaurants chooses randomly and suppose that subsequent patrons choose a restaurant with a probability proportional to the number of people already in the restaurant. A little simulation shows that the restaurant with the most patrons may well not be the best restaurant, it may have the most customers simply by chance.

Mr. Peterson doesn't seem to realize that, far from being a surprise, his story is an example of a phenomenon that has been well studied by economists. We call it a negative network externality and it is unlikely to persist in the real world. Is there no word-of-mouth about restaurants and are there no restaurant reviewers? Can't the better restaurants find ways to convince potential customers of their quality? Would people continue to eat at a restaurant solely because they had eaten there before by chance?

Economics: don't try it at home without consulting a professional.