Law enforcers frequently issue warnings, as opposed to sanctions, when they detect first time offenders. However, virtually all of the law and economics literature dealing with optimal penalty schemes for repeat offenders suggests that issuing warnings is a sub-optimal practice. Another observed phenomenon is the joint use of warnings and sanctions in law enforcement: person A may receive a sanction, whereas person B is only warned for committing the same offense. This situation can be explained through the use of hybrid warning strategies, a concept not yet formalized in the law enforcement literature, where law enforcers issue warnings to x% and sanctions to (100-x)% of first time offenders. This paper uses a two-period optimal deterrence model to provide a rationale as to why it may be optimal to issue warnings. When uninformed individuals are present and their punishment is assumed to be costly, there is a trade-off between such costs and reduced levels of deterrence. Depending on the cost structure associated with the punishment of uninformed individuals; hybrid, pure or no-warning strategies may be optimal. A secondary contribution of this paper is to point out that lack of information concerning laws may lead to optimal escalating punishments for repeat offenders.