I'll discuss three BPAI decisions which illustrate this point.
In Ex parte Ferri, the Applicant "complained" in the Appeal Brief that claims had been amended "in reliance on an interpretation offered by the Examiner, which was then withdrawn in a subsequent Office Action." The Board noted that
[This complaint] has been made by many applicants over the years and was answered many years ago by Judge Rich, who observed that "[t]here is nothing unusual, certainly, about an examiner changing his viewpoint as to the patentability of claims as the prosecution of a case progresses, and, so long as the rules of Patent Office practice are duly complied with, an applicant has no legal ground for complaint because of such change in view. The life of a patent solicitor has always been a hard one." In re Ruschig, 379 F.2d 990, 993 (CCPA 1967) (citation omitted).
In Ex parte Joos, the Examiner had withdrawn a rejection using a particular reference, then reapplied the same reference in a later round. The Appeal Brief characterized this behavior as "unreasonable, costly, and counterproductive." The Board was unsympathetic:
Reapplying references after both the Examiner and the Applicants had earlier thought them distinguished can be appropriate when, for example, the Examiner has a reason to think that a mistake was made. Mistakes may be costly and counterproductive to the Applicants and to the Office, but a good faith effort to correct the mistake is not unreasonable. Appellants’ objection here is not a legal ground for complaint. See Ruschig, 379 F.2d at 993; see also, Blacklight Power, 295 F.3d at 1273-74 (explaining that the Office may withdraw a notice of allowability to revisit an earlier patentability decision before a patent issues).
Ex parte Haley involved a method of gene therapy. In the first Office Action, the Examiner issued an enablement rejection, asserting that "the state of the art for gene therapy was highly unpredictable" at the time of filing. The Applicant got past the enablement rejection, but appealed an obviousness rejection in which the Examiner used the "predictable results" rationale from KSR. In the Appeal Brief, the Applicant argued that this rationale was inappropriate, since the Examiner had admitted in the first Office Action that "the state of the art for gene therapy was highly unpredictable." The BPAI dismissed the argument, noting that "[w]hile ideally the Examiner’s position would remain consistent throughout prosecution, the Examiner is permitted to change his viewpoint regarding the patentability of the claims as prosecution progresses."