This chapterHistory matchingreservoir presents a vast and complex reservoir history-matching case for the Troll field on the Norwegian Continental Shelf. It follows the same history-matching procedure as in the previous chapter. The Troll examples serve as a further confirmation of the learnings and recommendations coming out of the REEK experiments, and the examples illustrate the properties of the history-matching problem when conditioning on production rates. In addition, it uses the adaptive pluri-Gaussian method for parameterizing facies uncertainties and updating these in the history matching. The examples start with a standard history-matching procedure, computing global updates. After that, examples illustrate how the results improve when one properly accounts for uncertainties in the model controls and the facies parameterization. Finally, an adaptive localization scheme reduces the impact of spurious correlations and significantly improves the history-matching results.