Perhaps the most damaging mistake is not actually making a real argument. A surprising number of arguments in Appeal Briefs look like this:
Claim 1 recites a widget. Reference A does not teach a widget. Therefore the rejection should be overturned.I don't find that convincing, and the BPAI isn't usually convinced either. Here's what the BPAI has to say about "arguments" like that one:
Appellants’ contention that the references do not teach each limitation of the claim is supported by a summary of the references (Br. 11-14) and quotation of claims 4, 21, 41, and 61 (Br. 15-16). Appellants’ summary of the references and quotation of the claims contains no explanation of how the claim language patentably distinguishes over the references, and thus fails to identify a basis for concluding that any of the Examiner’s findings are erroneous.
(Ex parte Candy.)
For his part, the Appellant does not address these findings. Instead, he merely alleges that "[n]o such generation of suggested edits is taught by Microsoft Excel 2000" (Appeal Br. 11) and that "[t]here is absolutely no teaching whatsoever in Excel 2000 of such hierarchical log of edit regions." (Id. at 12.) These allegations "do not . . . explain why the Examiner's explicit fact finding is in error."Sometimes the argument provides a tiny bit more explanation:
(Ex parte Poynor.)
Claim 1 recites a widget. Reference A teaches a blodget. A blodget is not a widget. Therefore, the rejection could be overturned.But WHY is a blodget is not a widget? The BPAI wants to know:
Appellant argues that the combination of Angell and Orr does not teach or suggest "a stenocaptioning device configured for transcribing audio from an audio event simultaneously with an occurrence of the audio event." First, it is argued that the transcribed text in Angell is "synchronized." Br. 7-10 § A.1. This argument is not helpful because it does not explain why synchronized text cannot meet the limitation.The only time you're going to win with conclusory assertions is when the subject matter of the claim is worlds apart from the subject matter of the reference, so that the assertion "does not teach" is almost self-evident. Say, a claim to a chemical compound and a reference describing a mousetrap.
(Ex parte Pettinato.)
Now, I understand there are valid reasons why prosecutors would rather not make "real" arguments. The first one that comes to mind is prosecution disclaimer (also referred to as argument-based estoppel). If I thought I could get allowances with nothing but conclusory assertions, then I'd be tempted too.
Another reason for avoiding "real" arguments might be a fear of an inequitable conduct charge for mischaracterizing the reference. But based on the case law I've seen, I think this fear is overblown.