Is that the tomb we are looking for? assumption and reality

Today on WSJ there's an article (The Numbers Guy, which can be read for free) talking about the lost tomb of Jesus. Basically it said there's a tomb found in Jerusalem in 1980 with ten names on their coffins/ossuaries. Those names, by incident or fact, belong to Jesus family. A statistician, Andrey Feuerverger from University of Toronto, argued that there is 1/600 chance those names will come to a family that is not belong to Jesus, based on his assumptions and calculations.

To build a model for something, usually we need assumptions. For example, if an economist wants to figure out how people spend their money, first of all she has to make some assumptions on their type thus behavior. An usual assumption is about their attitude toward risk, such as people are risk averse or risk loving. Conditional on this assumption, she can go on to explore what they gonna behave in this virtual world. The best thing about modeling is, it can assume away those complicated but unimportant factors and extract the true essence that induce the result. Take the spending example, of course we spend some bucks on things we don't really need, but that is certainly not our regular pattern. When economist make a model on spending behavior those incidental purchase will be assumed away, to make sure the model can be more clear to grab the most important behavior, such as money spent on insurance. If we want to include as many factors as possible in the model, it will get more close to reality, but it will also get more complicated as reality.

As a result, a good model is one which makes some reasonable assumptions and neatly derives the result. On the other hand, the risk is also coming along with the assumption it made. In this tomb story, Andrey Feuerverger made his calculation based on many assumptions, and that is where people attack him. Just as the following quotes from this article, it is also what model-maker should be aware of:

"As you pile on more assumptions, you're building a house of cards," says Keith Devlin, a Stanford mathematician and NPR's "Math Guy."