In the second half of the 19th century, when words like ‘gene' or ‘genetics' did not yet exist and Mendel's theory had sunk into oblivion for decennia, Francis Galton developed his own, very influential theory of inheritance. Unlike Mendel, Galton was very well acquainted with modern statistical techniques, some of which he had helped to develop himself. So, given the fruitful role of statistical inferences in present-day scientific research, shouldn't we expect Galton to have found the ‘best’ theory of inheritance? Of course, every one knows that Mendel is still considered the founding father of genetics, while Galton's name is now only associated with dubious disciplines such as phrenology and eugenics. So the question is: what was wrong with Galton’s use of statistics? I will argue that Galton’s acquaintance with modern statistics biassed his research so that he could not have even recognized Mendelian patterns of inheritance. It suggested the wrong explananda (the normal distribution of inherited characteristics and regression towards the mean) and it led him to the wrong explanans (his ancestral theory of inheritance).