If evidential legal reasoning is probabilistic, and rational certainty about a factual hypothesis is unattainable, then we need rules, called “standards of proof”, to determine the degree of probability required to be able to deem the assumption as proven. Only if we have standards of proof that indicate evidential thresholds will we be able to use other decision-making rules, such as burdens of proof or presumptions. Indeed, the purpose of burden of proof rules is to determine who loses the case if there is insufficient evidence of any of the hypotheses at issue. However, to apply these rules we need to know when the evidence is sufficient. The same is true of rebuttable presumptions, which require that a hypothesis be admitted in court (and used in the reasoning) in the absence of (sufficient) proof to the contrary. Therefore, for example, the presumption of innocence as a procedural rule means that defendants must be found innocent unless there is sufficient proof of their guilt. But when is the proof of a defendant’s guilt sufficient? Once again, implementing the presumption of innocence (like any other rebuttable presumption) presupposes the existence of a rule that establishes the evidentiary threshold.