But the question remains what could be done to improve the situation. To our mind it is not simply a matter of better training. Not only do jurists who have attended training courses in interpreting forensic evidence do not understand likelihood ratios better than other jurists, also we have demonstrated that among forensic professionals themselves, many basic mistakes are made. One of the factors causing proper understanding to be deficient, especially in jurists, may be the fact that likelihood ratios are so often ‘sold’ in one package along with an advocacy to use it within the Bayesian framework for weighing evidence. But we need to stress that this is our opinion, and not a line of thought that finds direct support in our study findings reported above. Bayesians (e.g. Meester & Sjerps, 2004) hold that the diagnostic value of evidence (the likelihood ratio of the evidence under rivaling hypotheses) should be used to update the belief in the likelihood of the offender being guilty against the likelihood of him being innocent. In Bayes’ theorem an a priori belief (before assessing the evidence) is updated into an a posteriori belief of guilt/innocence using the likelihood ratio of the evidence. Many experts appear unaware of the fact that using likelihood ratios is not necessarily tied to the Bayesian framework as a prescriptive model for dealing with evidence. We are strengthened in our belief by the observation that many forensic scientists use the terminology of ‘Bayesian reporting’ when they are talking or writing about the likelihood ratio approach.