An important rule of criminal evidence is that evidence of a defendant’s previous misconduct (evidence of bad character) is prima facie inadmissible. The usual justification for this rule is that, in most cases, such evidence is either irrelevant or likely to have an unduly prejudicial effect on the fact-finder. This article questions this justification. After reviewing the psychological research on character, it examines statistical data on recidivism. The statistics suggest that those with previous convictions are much more likely to offend than are those without a criminal record, which implies that evidence of bad character will usually be sufficiently probative to justify its admission as proof of guilt. The remainder of the article examines various criticisms which can be levelled at this argument, in particular, that the recidivism data are misleading and that the way in which fact-finders reason renders evidence of bad character uninformative.