Earth’s twin will be discovered in 2013, astronomers predict

[D Gary Grady]

9. It's actually not an unreasonable prediction

There's more evidence to support the prediction than one might think. We have now observed enough stars and found enough exoplanets to have a notion of the approximate statistical distributions involved (mass, semimajor axis, etc).

Based on that and the parameters of the observing program, it's possible to compute an estimate the likelihood of observing at least one planet that meets a given set of criteria (mass in a given range and so on). No, the estimate isn't exact, and even if it were exact it wouldn't be a guarantee, so such a prediction obviously isn't a sure thing. But it's also far from being a mere unsupported claim.