Now for the caveats. Appellate success is notoriously difficult to quantify, and any estimate is derived with limited information—changes in the law, uncertainty regarding the other side’s argument, the director’s ability to intervene, and the delay in selecting the Federal Circuit panel that hears the appeal all contribute to that uncertainty. There is no way to predict the exact odds in a particular appeal. But a better understanding of why the typical appellant faces stiff odds and a review of arguments that have succeeded will help us weigh the relative strength of a particular case.

Setting the Baseline Odds

First, we examine the baseline—is the USPTO really affirmed 90 percent of the time on appeal? To answer that in the most straightforward way, a review of Federal Circuit results posting in the first six months of 2016 was conducted. Within this sample, the Federal Circuit decided 108 cases appealed directly from the USPTO. In 56, or roughly half, of the cases, the USPTO decision maker was summarily affirmed without opinion under Federal Circuit Rule 36.¹ These appeals were effectively ended with a single sentence. In another 33 cases, the USPTO was affirmed by written opinion affirming on all necessary merit and procedural points.² In two cases, the USPTO decision was reversed. And in 17 cases, partial relief was granted.³

Thus, the USPTO’s overall record (W–L–T) in the half-year sample was 89–2–17. Counting partial relief as half a win and half a loss, we can convert the Patent Trial and Appeal Board’s (PTAB’s) record to 93.5–10.5. Thus, appellants “won” roughly 10 percent of USPTO direct appeals reviewed. This is in line with the historical disadvantage felt by those appealing from adverse USPTO decisions.

Looking with more granularity, there is no strong evidence that cases of a particular type favor appellants. Though perhaps the USPTO is at its strongest in ex parte actions where the appellant faces the USPTO Solicitor’s Office.

• By forum: the PTAB was 85–2–17, and the Trademark Trial and Appeal Board (TTAB) was 4–0.
• By proceeding on appeal: the PTAB was 45–1–9 in inter partes reviews, 18–1–4 in inter partes reexaminations, 13–0 in patent prosecutions, 6–0–1 in ex parte reexaminations, 2–0–2 in interference proceedings, and 1–0–1 in covered business method appeals.
• By representation: the USPTO was 85–2–17 against represented appellants and 4–0 against pro se appellants.
• By party participation: the USPTO was 67–2–16 in inter partes cases (including 17–0–5 where the USPTO director intervened) and 20–0–1 in ex parte cases.

No matter how it is viewed, the USPTO wins—that is, the USPTO board is affirmed—in an overwhelming number of cases. Starting with a roughly 10 percent success baseline, the question becomes how to gauge the odds in one’s own case. To do that, first consider the manifest reason that the baseline is 10 percent: the standards of review.

The Standards of Review Heavily Favor Affirmance

Fifty-fifty cases before the USPTO boards become long shots at the Federal Circuit primarily because the standards of review favor affirmance.⁴ Academic discussions of the standards of review typically focus on the differing standards between issues of law and issues of fact. To the practitioner interested in appealing an agency’s final action—such as a PTAB or TTAB decision—it may be more helpful to contrast merits issues and procedural issues.

Merits Standards of Review

A challenge to the legal merits is typically preferred because winning on the merits means reversal of the underlying patentability or registrability decision or vacatur and remand for the USPTO to consider the next logical merits issue in the case. In other words, the rejection is undone though the board may have a chance to revisit the ultimate question. Merits challenges typically address whether the agency’s decision on patentability is (1) factually supported and/or (2) in conformance with the law. To be affirmed on questions of fact, the USPTO board need only have relied on “substantial evidence.” Substantial evidence, while more than a “mere scintilla,” is only “such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.”⁵ Questions of law are reviewed de novo by the Federal Circuit, with one caveat: the USPTO generally applies the broadest reasonable interpretation (BRI) to claim language. Thus, while review is de novo, the underlying question often devolves into whether the USPTO’s interpretation is the “broadest reasonable,” which favors affirmance when claims have been found unpatentable.

A successful challenge to USPTO procedures, in contrast, typically results in vacatur and remand for the USPTO to reconsider the same merits issue in light of a revised procedure or additional evidence. For example, in perhaps the most famous procedural victory by an appellant, In re Lee, the USPTO’s rejection based on “common sense” without any supporting reasoning was vacated “for failure to meet the adjudicative standards for review under the Administrative Procedure Act” and remanded for “further proceedings.”⁶ A seeming victory but perhaps a hollow one for Lee, who returned to the Federal Circuit years later with the same application, and the rejection of all claims therein was affirmed without opinion.⁷

Procedural Standards of Review

Procedural challenges typically address whether an agency decision violates (1) a statute, (2) the agency’s own rules, or (3) the Administrative Procedure Act (APA). In each situation, the Federal Circuit typically defers to the USPTO. The USPTO receives Chevron deference if one challenges a procedural rule or decision as inconsistent with a statute. The Federal Circuit will ask only whether the USPTO’s rule is “based on a permissible construction” of the statute.⁸ Similarly, if one challenges the PTAB’s interpretation/application of a rule, the Federal Circuit will likely review the USPTO’s interpretation of its own regulations with “substantial deference,” sometimes called Auer or Seminole Rock deference, looking only to see if the interpretation is “plainly erroneous or inconsistent with the regulation.”⁹ Regardless of the precise nature of the challenge, the agency must not violate the APA’s catch-all prohibition against actions that are “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.”¹⁰

All these standards weigh heavily toward affirmance on procedural matters. But they are not applied to every question. For example, a regulation that merely restates the terms of the statute—as USPTO regulations frequently do—does not warrant deference.¹¹ The Federal Circuit also has historically distinguished “procedural rules” (that warrant Chevron deference) from “substantive rules” (that do not).¹²

Relief Granted Links to the Standard of Review

Scrutinizing the cases of partial relief and complete reversal in table 1 allows a comparison of the relief granted to the standard of review. In nine of these cases, the court reversed the board in whole or in part. The issue in six of the nine reversals was one of law (five of six involved claim construction), and the court determined that correcting the law disposed of the case. The cases also contain 18 issues that were vacated. Of these, five were issues of claim construction remanded for further consideration of art, one was an issue of fact, and 11 involved procedural errors, the most common of which were failure to address arguments (four) and failure to consider evidence (three).

If this sample may be considered representative, it supports the conventional wisdom that claim construction and other issues of law are the best merits issues, and USPTO failure to consider arguments and evidence are strong procedural issues.

So What Are My Odds, Exactly?

A simple three-step approach to answering this ultimate question may be sufficient, if it accounts for the standards of review, the baseline odds, and the historical case law. One such approach is proposed:

1. Vet and select potential issues for the appeal.
2. Put the relief available for each issue in the context of the baseline odds.
3. Compare your issues to case law and adjust accordingly and realistically according to your own judgment.

One hypothetical example may prove instructive. Assume a patentee has lost in the USPTO on two different grounds. The patent’s claim 1 represents all claims rejected under a theory of inherent anticipation. Claim 20 represents all claims rejected as obvious on different art.

Step 1: Vetting the Issues

Vetting the issues—considering the issues you might raise and the relief available should you prevail—is a traditional appellate step discussed in many excellent texts and articles.¹³ Regardless of your approach, always keep the standard of review firmly in mind when deciding which issues to brief.

For our hypothetical case, assume you have vetted the issues and selected three: (1) the examiner improperly shifted the burden on anticipation, (2) claim construction for one phrase in claim 20, and (3) the PTAB’s opinion failed to address evidence of secondary considerations rebutting the obviousness finding.

Step 2: Setting Potential Relief Available against a Baseline

In this example, there are a few merits issues and one procedural issue. So the simple odds matrix, absent more information, might look like this:

Likely Outcomes on Appeal—Baseline

1. 50%—Summary affirmance under Rule 36.

2. 30%—Affirmance on all merit and procedural points in a written decision.

3. 15%—Affirmance-in-part, reversal-in-part, vacatur-in-part, and/or remand.

a. 1–5%—Win on anticipation, lose on obviousness.

b. 1–10%—Win on claim construction, lose on anticipation.

i. 6%—Win on claim construction, obviousness rejection reversed.

ii. 3%—Win on claim construction, obviousness rejection remanded for PTAB’s consideration of art under new construction.

iii. 2%—Win on claim construction, obviousness rejection affirmed.

c. 1–5%—Anticipation rejection affirmed, obviousness rejection vacated and remanded for consideration of evidence.

4. 5%—Reversal on both rejections, procedural error mooted.

Emphasizing the caveats: It should be apparent that there is no way one can render the odds so precisely on an appeal. Frankly, setting odds in appellate law (or any area of law) is a ridiculous game—but one that some clients (and some appellate lawyers) cannot resist playing. Any quoting of odds like these should note that they are overly precise and only reflect a best guess based on experience. With those caveats in mind, we proceed to step 3.

Step 3: Adjusting the Odds in View of Case Law

The final step is comparing your issues to case law and adjusting the odds accordingly and realistically. Accordingly requires one to understand the nature of the case law on the issues identified and to apply your judgment to whether each issue identified is run-of-the-mill or a standout. For our hypothetical case, the odds of vacatur go up if the PTAB explicitly refused to look at any evidence of secondary considerations. The same odds might go down if the examiner discussed the evidence and the PTAB obliquely referred to that discussion. Likewise, claim construction that directly contradicts a definition in the specification has better odds of being revised than one that merely excludes an embodiment.

Realistically, your matrix should add up to 100 percent. The tendency may be to treat each issue separately and independently. A client told that there are four issues, each with a 25 percent chance of winning might think the odds are quite good of winning something. But the most likely outcome in such a case is affirmance. Why? First, because the odds in patent cases are rarely independent of one another. For example, if the PTAB were to be affirmed on written description, it is more likely to be affirmed on a related question of enablement.

Second, even for doctrinally independent issues, like the publication date of a prior art reference addressing one claim element and written description of an unrelated element, the odds are linked subtly by that 3,600-pound gorilla, Federal Circuit Rule 36. Rule 36 is difficult to avoid absent a single strong issue. The Federal Circuit sits in three-judge panels, and at a typical sitting the panel will hear four to six cases. Rarely does the panel issue opinions in every case. Much more typically, one or two of the cases heard will receive a Rule 36 affirmance.

Thus, realistically also means treading carefully before quoting any odds of success approaching 50 percent on a USPTO appeal. If we can take the sample as representative, only 50 percent of all USPTO cases will receive a written decision. Of the cases written on, more than half will be affirmed on all points. Thus, winning an appeal requires beating tough odds.

Conclusion

This article attempts to help you in bringing exact numbers to an endeavor that is uncertain and inexact. That is ridiculously difficult. But if following the suggested method only helps you think about your USPTO appeal from a new angle or with fresh realism, it will have been a worthy exercise.

Endnotes

1. See Fed. Cir. R. 36 (allowing entry of judgment without opinion where the USPTO’s decision “warrants affirmance under the standard of review” or is “without an error of law”).

2. Three such cases were dismissed, at least in part, for lack of standing or mootness.

3. For example, the court might affirm a finding of anticipation for a subset of claims while reversing a holding of obviousness for the remaining claims. The Federal Circuit has multiple flavors of partial relief, each of which combine two or more of vacatur, reversal, affirmance, and remand in a single opinion. See, e.g., In re Man Machine Interface Techs., LLC, 822 F.3d 1282, 1289 (Fed. Cir. 2016) (“affirmed-in-part, reversed-in-part, vacated-in-part, remanded”).

4. For an in-depth discussion of standards of review, see Harry T. Edwards & Linda A. Elliott, Federal Courts: Standards of Review: Appellate Court Review of District Court Decisions and Agency Actions (2007).

5. See Blue Calypso, LLC v. Groupon Inc., 815 F.3d 1331, 1337 (Fed. Cir. 2016) (quoting the classic case Consol. Edison Co. of N.Y. v. NLRB, 305 U.S. 197, 217 (1938)).

6. In re Lee, 277 F.3d 1338, 1340–41 (Fed. Cir. 2002), cited with approval in Cutsforth, Inc. v. MotivePower, Inc., 636 F. App’x 575, 577–78 (Fed. Cir. 2016).

7. In re Lee, 262 F. App’x 275 (Fed. Cir. 2008). For completeness, it should be noted that the rejection in the second Lee case was not identical, although the primary piece of art was the same in both.

8. See, e.g., Cooper Techs. Co. v. Dudas, 536 F.3d 1330, 1332, 1338 (Fed. Cir. 2008).

9. See, e.g., In re Lovin, 652 F.3d 1349, 1353 (Fed. Cir. 2011); see also Auer v. Robbins, 519 U.S. 452, 462 (1997); Bowles v. Seminole Rock & Sand Co., 325 U.S. 410, 414 (1945). But see Perez v. Mortg. Bankers Ass’n, 135 S. Ct. 1199, 1210–11 (2015) (Alito, J., concurring in part).

10. 5 U.S.C. § 706(2)(A).

11. See Gonzales v. Oregon, 546 U.S. 243, 257 (2006).

12. Whether that distinction has any force after Cuozzo Speed Techs., LLC v. Lee, 136 S. Ct. 2131, 2143 (2016) (“That statute [35 U.S.C. § 2(b)(2)(A)] does not clearly contain the Circuit’s claimed limitation[.]”), and City of Arlington, Tex. v. FCC, 133 S. Ct. 1863 (2013) (holding that courts must defer under Chevron to an agency’s interpretation of a statutory ambiguity that concerns the scope of the agency’s statutory authority), has not been resolved.

13. See, e.g., Ursula Bentele & Eve Cary, Appellate Advocacy, Principles and Practice 256–61 (2d ed. 1995).