In many of CCAMLR’s new and exploratory fisheries, stock assessment techniques are still in development. The failure of the tag-recapture program to provide robust assessments in many areas and doubts about using unstandardised catch rate data have led to a desire to draw inferences about stock status and fishing mortality from length-frequency data alone. In this paper, a number of length-based stock assessment methods are revisited and some limitations in their ability to draw robust inferences about stock status and fishing mortality for long-lived fish species such as Patagonian toothfish (Dissostichus eleginoides) are illustrated. While length-frequency data can be useful in assessments when integrated with other informative estimators for stock abundance and productivity, the review and the simulations highlight that interpretation of raw length-frequency data and single length-based indicators for stock status and fishing mortality is problematic and should be avoided.