Obvious to Try

The 'obvious to try' doctrine is one of KSR's rationales for finding a patent claim obvious under 35 U.S.C. § 103. Pre-KSR, the Federal Circuit had held that 'obvious to try' alone was insufficient to establish obviousness — the fact that a skilled artisan could try a combination did not make it obvious. KSR International Co. v. Teleflex Inc. (S.Ct. 2007) rejected this categorical rule and held that 'obvious to try' CAN render a claim obvious when: (1) there is a recognized problem or need to be solved; (2) there are a FINITE NUMBER of identified, predictable solutions to try; (3) a skilled artisan would have had reason to try one of the known options with a REASONABLE EXPECTATION OF SUCCESS. If these conditions are met, pursuing the known options is 'obvious' even without a specific teaching or suggestion to combine them.

In re O'Farrell (Fed. Cir. 1988) established the two-category framework for obvious-to-try analysis that predates KSR and remains influential after it. TWO CATEGORIES WHERE 'OBVIOUS TO TRY' IS VALID: (1) FINITE AND SMALL NUMBER OF SOLUTIONS — the prior art presented 'a finite number of identified, predictable solutions' and a skilled artisan pursuing the obvious solution would have had 'reason to pursue the known options within his or her technical grasp.' When there are only a few reasonable options and each is predictable, it is obvious to try one; (2) GENERAL APPROACH WITH REASONABLE EXPECTATION OF SUCCESS — the prior art described a general approach and there was a 'reasonable expectation of success' in applying that approach to the claimed invention. TWO CATEGORIES WHERE 'OBVIOUS TO TRY' IS NOT VALID: (1) LARGE NUMBER OF UNPREDICTABLE SOLUTIONS — if the prior art gives only 'general guidance' toward an infinite or large unpredictable field, trying one particular solution is not obvious; (2) NO REASONABLE EXPECTATION OF SUCCESS — if a skilled artisan would not reasonably expect success in the particular combination, the attempt is not obvious even if it could be tried.

Pharmaceutical and biotech cases frequently litigate the 'obvious to try' doctrine. In re Kubin (Fed. Cir. 2009): Amgen sought a patent on an isolated gene encoding the NAIL protein (a natural killer cell receptor). The Federal Circuit held the claim obvious to try: the prior art had identified the protein, described partial amino acid sequences, provided conventional cloning techniques, and given skilled artisans reason to isolate the gene with a reasonable expectation of success using known methods (degenerative oligonucleotide probes, expression libraries). The fact that the actual isolation required effort did not prevent the claim from being obvious — the path was predictable even if the work was not trivial. CONTRAST: if a pharmaceutical candidate field has millions of possible compounds and no specific guidance toward a particular class, trying any one compound is not 'obvious to try.' The more specific and predictable the prior art guidance, the more likely an 'obvious to try' argument succeeds. Pharma ANDA litigation (Hatch-Waxman) frequently uses obvious-to-try to challenge formulation and method-of-treatment patents.

Unexpected results are a secondary consideration (objective indicia) that can rebut an obvious-to-try argument under § 103. If a skilled artisan trying a finite set of predictable options would have expected a certain result, but the claimed invention produces a surprisingly different (better) result, the unexpected result can support non-obviousness. Requirements for unexpected results as a rebuttal: (1) COMPARISON TO CLOSEST PRIOR ART — the unexpected result must be compared to the result that would have been predicted or expected based on the prior art; (2) NEXUS TO THE CLAIMED INVENTION — the unexpected result must be tied to the claimed invention as a whole, not to unclaimed features; (3) THE DIFFERENCE MUST BE UNEXPECTED, NOT MERELY BETTER — an improvement that is simply better than expected is not necessarily 'unexpected' in the legal sense; a true reversal of expected properties or a quantitatively superior result beyond the range of reasonable expectation qualifies. Unexpected results do NOT automatically save an otherwise-obvious claim but must be weighed against the strength of the obviousness case.

Obvious to try fails as an obviousness argument when: (1) UNPREDICTABLE FIELD — the prior art points toward a general area of research but there are so many unpredictable variables or possible solutions that no single one is obvious to try (Burroughs Wellcome v. Barr Labs: HIV treatment — specific combination not obvious from broad guidance); (2) NO REASON TO TRY — even with a finite number of options, there must be a reason a skilled artisan would be motivated to try the specific combination in the claim; (3) NO REASONABLE EXPECTATION OF SUCCESS — the skilled artisan, while able to try, would not have reasonably expected success; this is common in biotechnology where expression or folding of proteins is unpredictable; (4) COMBINATORIAL EXPLOSION — if the 'finite' number of combinations is actually very large (e.g., 100 variables × 10 values each = 10^100 combinations), no single combination is obvious despite the finite theoretical count; (5) UNEXPECTED RESULT REBUTTAL — unexpected superior results or unexpected properties can overcome an otherwise valid obvious-to-try argument.