Greg, I would have to go back and find the study, but the issue with fingerprint analysis (and potential flaws) has less to do with the intrinsic ability of examiners to make a blind match (or mismatch) but with confirmation bias. In this study, they sent fingerprint examiners a set of two prints for comparison-- one lifted from a crime scene, one from a fingerprint card. The examiners were almost universally correct in their assessments, just as in the study you posted. There was one problem: these were not truly blind sets of fingerprints but prints the examiners in question had analyzed for police/the prosecution in past cases. And in some alarming percentage-- like 25% of them-- they offered a *different* opinion than the one they gave years before. The study's researchers argued that it was because in their earlier analysis, the examiners were given background information from the cops or prosecutors. Which was the actual problem in the Mayfield case as well. I was told about this by the lats Dr. Cliff Spiegleman a decade or more ago. Cliff had made it almost his life's work to make sure crime labs conducted almost every analysis completely blindly (not just fingeprints) because of confirmation bias issues.
I would also add that there was a refutation-- "no match to Wallace"-- not long after Darby's match. From one of the former heads of a national fingerprint professional association. If you want an amazing presentation on the fingerprint issue-- find James Olmstead's 2003 Wecht Conference presentation on Youtube.
Stu