As is well known, there are two crucial arguments in the realism debate. According to the no-miracles argument, it would be a miracle if our best scientific theories – namely, those which successfully predict novel phenomena – were not true (or approximately true). So, we should take theories that yield novel predictions as being true or, at least, approximately so. Clearly, considerations of this sort are raised to support realism. On the other hand, according to the pessimist meta-induction, many of our best-confirmed theories have turned out to be false. So, how can we guarantee that current theories are true? Considerations such as these, in turn, are meant to provide support for anti-realism.