Critics have raised concerns that investment arbitration tribunals treat like cases differently – raising problems of consistency – and different cases the same – creating problems of correctness. Using network analyses of case citations, this chapter examines an observable selection of what a tribunal considers to be ‘relevant’ precedent. The author finds that tribunals are more concerned with consistency than correctness, with many tribunals citing precedents based on textually dissimilar treaties. However, this is contrary to what states consider the priority in ISDS reform debates – in which correctness is higher than consistency in the hierarchy of policy preferences. The chapter concludes that states can resolve that mismatch by hard-coding their policy preferences into institutional design. It is argued that, as part of the ISDS reform, states should make the ordering between correctness and consistency considerations explicit when reshaping adjudicatory authority in future ISDS institutions.