People v. Collins 68 Cal.2d 319, 66 Cal.Rptr. 497, 438 P.2d (1968) Collins and his wife were accused of robbery. Collins was a black man with a beard and his wife was a blond white woman. At Trial, the prosecutors had difficulty establishing a positive identification, so they resorted to probabilistic evidence. Basically, they brought in a math professor as an expert witness to say that since witnesses claimed that the crime was committed by a 'black man with a beard and a blond white woman' there was an overwhelming probability that the crime was committed by any couple answering to such distinctive characteristics. Only 1 in 12 million couple share these characteristics. The Trial Court found Collins guilty.  He appealed. The California Supreme Court reversed and remanded for a new trial. The California Supreme Court found that guilt cannot be determined by odds, and that the introduction of probabilistic evidence infected the case with fatal error. The testimony itself lacked an adequate foundation in both evidence and statistical theory. The expert appeared to have pulled the statistical evidence out of his butt. The testimony distracted the jury from its proper function of weighing evidence on the issues and made them rely upon an irrelevant mathematical demonstration. Basically, even if you could prove that few couple met the description, this evidence has no relevance as to whether or not Collins and his wife committed the crime. What if the true criminal was wearing a fake beard?  How would that skew the statistics?