the logical conclusion is that anything that has a chance of making the world a better place ought to be pursued.
Ahh, but here's where the important notion of probability comes into play. Given that two choices are mutually exclusive, the one with the higher probability should be pursued (this is the central premise of rational choice theory, that decisionmakers tend to choose that which provides the highest gain at the least risk).
A sperm is a low-risk but low-payoff item. For a typical given act of sex, roughly 3-5 million sperm are ejaculated, and only one will fertilize the egg. This assumes that the egg is in a state in which it can attach to the uterine wall and that the act will be successful. Given 100% success rate (which is NEVER the case), a sperm is a 1 in 3 million payoff.
According to Wikipedia, miscarriage can occur up to 25% in normal situations. That does, however, give a 75% success rate. Assuming the payoff is equal, the likelihood of a "clump of cells" making it to the "potential payoff" over the single sperm that was going to fertilize (assuming condom or withdrawal practice) is 2,250,000 times higher [0.75 / (1/3,000,000)]. The sperm then has the 25% chance of miscarriage as well.
That's an entirely different argument from, "it has potential to save the world" -- everything does.
When you start bringing probabilities and such into play, it almost sounds like Utilitarianism. That is, for anybody here who doesn't know it, attempting to make the decision that will have the best outcome for everybody, raising the amount of "good" in the world.
The problem here, and I think it's the same problem all along, is, how do we define a good outcome? I think that, statistically speaking, the chances of somebody born from an 11-year-old mother has a very small chance of bringing more good to the world than harm. However, if you consider another human life being created to count as "good", then having the baby would be beneficial.
Looking at it from utilitarianism seems to favor, I think, having an abortion, because there is, quite likely, at least one less person suffering (whether it's the mother or the child).
I don't know if that's an accurate portrayal of Utilitarianism, though. It seems to me that murder is hard to justify as utilitarian, even if it's murdering a bad person (ie, an execution). But that would being us back to what I was saying earlier about whether or not it would count as murder.
