The shaken baby syndrome was originally proposed in the 1970s without any formal scientific basis. Once data generated by scientific research was available, the hypothesis became controversial. There developed essentially two sides in the debate. One side claimed that the clinical triad of subdural haemorrhage, retinal haemorrhage, and encephalopathy, or its components, is evidence that an infant has been shaken. The other side stated this is not a scientifically valid proposal and that alternative causes, such as low falls and natural diseases, should be considered. The controversy continues, but the contours have shifted. During the last 15 years, research has shown that the triad is not sufficient to infer shaking or abuse and the shaking hypothesis does not meet the standards of evidence‐based medicine. This raises the issue of whether it is fit for either clinical practice or for the courtroom; evidence presented to the courts must be unassailable.