Reliabilist accounts of knowledge are widely seen as having the resources for blocking sceptical arguments, since these arguments appear to rely on assumptions about the nature of knowledge that are rendered illegitimate by reliabilist accounts. In Scepticism and Reliable Belief José L. Zalabardo assesses the main arguments against the possibility of knowledge, and challenges their consensus. He articulates and defends a reliabilist theory of knowledge that belongs firmly in the truth-tracking tradition. Zalabardo''s main analytic tool in the account of knowledge he provides is the theory of probability: he analyses both truth tracking and evidence in these terms, and argues that this account of knowledge has the resources for blocking the main standard lines of sceptical reasoning--including the regress argument, arguments based on sceptical hypotheses, and the problem of the criterion. But although Zalabardo''s theory can be used to refute the standard lines of sceptical reasoning, there is a sceptical argument against which his account offers no defence, as it does not rely on any assumptions that he renders illegitimate. According to this argument, we might have considerable success in the enterprise of forming true beliefs: if this is so, we have knowledge of the world. However, we cannot know that we are successful, even if we are. Beliefs to this effect cannot be knowledge on Zalabardo''s reliabilist account, since these beliefs do not track the truth and we cannot obtain adequate evidence in their support. Zalabardo ends with the suggestion that the problem might have a metaphysical solution: although the sceptical argument may make no illegitimate epistemological assumptions, it does rest on a questionable account of the nature of cognition.