There's a famous problem in Mathematics called "the hundred suitors problem."

Basically, the riddle is this:

There is a maiden who must select one husband (and only one ) out of one hundred suitors. Each suitor is presented in random order, but only once. The maiden must accept the suitor, or reject the suitor and pass on the next. Of course, once she makes a choice, it is final. The question is, what method of choice should she use to maximize her chances of picking the best husband?

The solution:

1. reject the first third suitors, for comparsion
2. then pick the first suitor that is better than all the others in the comparison group

Using this methodology, the chances of picking the best suitor is about 1/3. Not bad odds. Contrast this with picking one at random, or picking the first suiter, the odds are 1/100.

With religion, there is a similar problem in that a person is presented with N religious options, and given a choice between them.

If someone really wants to maximize their probabilities of finding the best religion, they would categorically always reject the first religion presented to them. The likelihood that the first suitor is the best is fairly low. This also helps to couteract the ingrained biases one develops in a geographic region where one particular relious viewpoint is dominant. Of course, an objection might be: why can't a person examine all regions, then make her choice? This is a valid object to the model I presented.

But almost always, a person's religious views are simply a copy of what their parents believe, which are a function of the geographic region and culture. In other words, the basic method that 99% of people use is: they accept the first religion presented to them, and justify it with someting like Pascal's wager.

It was unfortunate that Pascal gave up mathematics to pursue religion, since otherwise he could have increased his chances of survival.