The various trials of soldiers for the excessive use of force or for the ill-treatment of detained civilians during international armed operations have attracted considerable publicity. The explanation frequently given by senior commanders is that misconduct of this type can be explained by seeing the alleged culprits as ‘bad apples’ and, once they are removed, the military system should be able to operate without further incidents of serious wrongdoing. This article explores why the various forms of misconduct take place and, in so doing, it considers the effectiveness of training, the effect on the soldier of any uncertainty of action, command failures and group influences. The action likely to be taken (or not taken) by way of legal or other proceedings against both soldiers and their commanders is assessed. The article concludes by taking the position that whilst the ‘bad apple’ theory might explain why some, but not all, soldiers engage in military misconduct during armed operations it fails to address the more important systemic issue, namely, whether adequate training and effective command responsibility can control such conduct. It argues for the need, on the part of military commanders, to reduce uncertainty of action or conduct to its irreducible minimum (through measures to counteract the causes of misconduct discussed in the article). This may then prove to be the key to reducing the need to rely predominantly on taking legal proceedings against individual soldiers in a search for alleged ‘bad apples’.