Now that the immediate fallout from the Supreme Court’s blockbuster Amex decision has cooled, this Article aims to give a first draft of its place in antitrust history and to offer a roadmap for the next stage of the evolution of platform antitrust analysis. We focus on several issues that have not been fully analyzed in the literature. First, we argue that, rather than mangling the law of market definition, the Court should have explicitly permitted multi-market balancing of effects across the separate markets in which the platform was active. Second, we propose standards to implement such balancing in cases involving multisided platforms in a principled manner, and we illustrate those standards by application to the facts of Amex itself. Third, we highlight three significant omissions in the Court’s analysis that confounded its assessment of the case and threaten to mislead future courts. In particular, we show that the Court failed to take into account that: (1) all three major card networks had parallel anti-steering rules during the period of alleged anticompetitive conduct and harm, which increased the anticompetitive harms by (a) increasing the incentives for fee increases for each network, (b) reducing the incentives for fee decreases by each network, and (c) leading the Amex restraints to harm holders of other credit cards as well as consumers who pay with debit, cash and checks; (2) higher merchant fees caused by these parallel antisteering rules placed consumers into a prisoners’ dilemma game, which led inevitably to increased use of credit cards above the efficient, competitive level, making the volume of card transactions a poor proxy for welfare effects; and; (3) American Express’ cardholder rewards were not necessarily evidence of procompetitive consumer benefits, as some rewards may have represented rents from anticompetitive conduct. Finally, we suggest several options for the courts and Congress to remedy the problems caused by the Court’s faulty approach and place platform antitrust back on track.