In 1932, Edwin Borchard’s Convicting the Innocent offered the first systematic attempt to document and describe the existence of Judge Hand‘s “ghost” of the innocent prisoner wrongly convicted. Three quarters of a century later, Professor Brandon Garrett has published a book of the same title with similar goals – but in the context of the game-changing phenomenon of DNA exonerations. Garrett’s volume summarizes the vast data collected and analyzed in his study of the first 250 DNA exonerations in the United States, and in so doing offers the most empirically rich and conceptually nuanced descriptive account to date of the machinery of wrongful conviction. Moreover, to ongoing debates over the direction of criminal justice reform, it offers a persuasive and sustained critique of the tenaciously resilient notion that the unparalleled procedural protections of the American jury system are effective checks on substantive accuracy as well. And yet, Convicting the Innocent does not offer quite as powerful an explanatory lens as Garrett sometimes claims, and does not advance the ball of criminal justice policy reform as far as it might. Part II of this Review suggests that the project is hampered in fulfillment of its descriptive and prescriptive agendas by constraints intrinsic to the data at Garrett‘s disposal, and by limitations that Garrett‘s own framing and methodology impose. Part II further offers that modest but important qualifiers and additions to Garrett‘s agenda could enhance the prospect that his worthy contribution to criminal-justice-reform conversations will translate into positive and much-needed change.