Not So Obvious After All: Patent Law's Nonobviousness Requirement, KSR, and the Fear of Hindsight Bias

Before the creation of the Federal Circuit in 1982,nonobviousness served as the primary gatekeeper forpatents. When patent holders sued for infringement andlost, more than sixty percent of the time, they lost on thegrounds that their patent was obvious. With the advent ofthe Federal Circuit, nonobviousness became a much lessdifficult hurdle to surmount. From 1982 until 2005, whenpatent holders sued for infringement and lost, obviousnesswas the reason in less than fifteen percent of the cases.While obviousness remained formally a requirement ofpatent protection, there can be little doubt that the FederalCircuit had substantially diminished the doctrine's oncecentral role.A potential turning point arose in 2007, however, withthe Court's decision in KSR v. Teleflex.

In its decision, theCourt, with one hand, rejected some of the key restrictionsthe Federal Circuit had placed on the obviousnessdoctrine, and with the other, broadened the circumstancesunder which obviousness could be found. Taken at facevalue, the Court's decision seemed poised to reinvigoratethe nonobviousness requirement. Yet an analysis ofappellate patent decisions since 2007 reveals that theCourt's decision has led to only a slightly increased rolefor the doctrine. By any measure, the nonobviousnessrequirement remains a pale shadow of its former self.This Article presents this historical background andthen turns to the fear of hindsight bias that seems to havemotivated much of the desire to limit obviousness's reach.As articulated by the Federal Circuit, hindsight biasrepresents the fear that a judge or jury will use aninventor's own invention against her.

Having been told ofthe inventor's solution, the judge or jury will use theinventor's own actions as a roadmap showing how tocombine existing prior art elements to solve the problem athand. With the benefit of hindsight, a judge or jury willtoo readily infer that the inventor's solution was obvious.In a pair of articles, Professor Gregory Mandel has presented results from empirical tests that he believesdemonstrate the potential for substantial hindsight bias inpatent litigation. In this Article, the authors modify andextend Professor Mandel's empirical tests. Our resultsraise questions regarding Professor Mandel's conclusionsand tend to refute the notion that hindsight bias is aserious problem in patent litigation