Pascal’s Wager is one of the more popular and silly arguments for why one should believe in the existence of some particular god. In brief, the premises are that believing in a false god is not very costly, while failing to believe in a true god entails a huge cost (at some point, at least), so erring on the side of belief is a rational choice. [Note the prima facie appeal to logic, as Pascal’s Wager was designed to work on educated skeptics.] The standard counters to this idea include the difficulty in willing oneself to believe something for which one has no evidence, which seems to assume a cognitive process that enjoys as little support as the existence of god(s), and the problem of there being myriad gods to choose from, which makes the entire argument rather unhelpful to those who are only seeking to hedge their bet. (For more, see, e.g., http://rationalwiki.org/wiki/Pascal%27s_wager.)
More problematic, in my opinion, are the consequences of any religion using this argument in an open marketplace of ideas, which, these days, will almost always include at least one alternative religion. This is best seen by applying Bayes to the question, under which the strength of the suggestion to believe in god X is proportional to the cost of not believing the claim. If, for example, the penalty for non-belief is that one simply loses access to some form of after-life, then this will be quite ineffective on most atheists, as they already expect to be “paying” this cost. In contrast, if the cost of non-belief is to roast in the fiery pits of hell for a really, really long time, then even an atheist might be tempted to give this religion a try.
As long as only one religion is using Pascal’s Wager on targets and the costs of non-belief remain vague, futuristic, and unfalsifiable, there’s not too much danger. The target need only apply any of the numerous counters and carry on as usual. But when neither of these conditions are met – when multiple religions are each using this approach and employing costs that are immediate and tangible – then we arrive at the serious problem of Pascal’s Bidding War.
If the strength of an argument based on Pascal’s Wager depends on the costs of non-belief, then, in order to “win” this argument over competitors, adherents to a particular religion need to make the costs of non-belief in their claims much higher than the costs associated with non-belief in the claims of all others. And to prevent the target from ignoring these costs on the grounds that they only occur in the future, the religion needs to make these costs immediate and tangible, which brings us to anything from the Spanish Inquisition to modern-day Islamic beheadings. And it can only get worse, as bidding wars – by definition – always escalate. What may have been somewhat effective in the past – e.g., roasting in the fiery pits of hell for eternity – would no longer be an acceptable bet; it is now barely an ante, I’m sorry to say.
edit/update: tl/dr; http://www.jesusandmo.net/2006/08/23/hell/