How to verify N.D. Cal. PTAB appeal win rates for Berkeley software patents, 2019–2024

This guide explains how to benchmark PTAB appeal win rates for Berkeley software patents — how to pull and verify any claimed N.D. Cal. figure and national baseline for 2019–2024 before relying on either. No specific rate is treated as confirmed here until the case counts and win definitions behind it are checked.

It shows how to verify live, complete data and compare like-for-like totals and terms before relying on local prosecution practice.

How It Works

A Patent Trial and Appeal Board decision on an ex parte appeal does not end the matter; it splits the path. The applicant may appeal to the U.S. Court of Appeals for the Federal Circuit under 35 U.S.C. § 141, or bring a civil action against the Director under 35 U.S.C. § 145. Section 145 is the provision that can put a PTAB appeal on a Bay Area patent owner's home docket at all: the statute routes the action to the district where the patent applicant resides unless it is filed in the Eastern District of Virginia. For a Berkeley software filer, that ordinarily means the Northern District of California. The routes are legally distinct — one is appellate review on the administrative record, the other is a civil action in which the applicant may, within limits, introduce evidence the Board never saw.

That distinction is why an N.D. Cal. figure and a national average can describe the same patent and still not be comparable. A win rate is a ratio, and the ratio is only as sound as the two totals behind it: the numerator (what counts as a win) and the denominator (which appeals are counted). Change either, and the number moves with no change in the law.

Fix the numerator first. Reversal of every rejected claim, remand with instructions favorable to the applicant, and affirmance with a narrowed rejection sustained are different outcomes; folding them into a single "win" column is the most reliable way to make a comparison meaningless. Fix the denominator next. Appeals filed, appeals decided, and appeals decided inside a fixed multi-year window are three different populations, and a rate computed on decided appeals will read differently from one computed on filings because pending cases drop out.

Then align the terms. In this setting, terms means the conditions under which each total was gathered: the same window, the same appeal route, the same technology, and the same procedural starting point. Local prosecution practice enters the mechanism here, because what the examiner and the Board reviewed — the claim scope chosen during prosecution, the amendments and remarks left in the file — is the same record the district court or the Federal Circuit reviews. A cleaner prosecution record narrows the space in which either route can help.

The practical check is procedural, not statistical: before crediting any figure, confirm which route it covers, which outcomes it counts, and whether the comparison total was built on the same terms. Section 145 actions carry a statutory obligation to cover the Office's expenses, so the route is not cost-neutral either. If any of those elements cannot be confirmed, treat the number as unverified rather than adjusting it by assumption.

Key Factors to Consider

When you are deciding whether an N.D. Cal. PTAB appeal track is worth committing to for a Berkeley software patent, three criteria do almost all of the work: the like-for-like win-rate comparison, the total-term math, and the live docket status of the exact option you are about to pick. Treat each as a gate, not a score. If any one of them fails, the local-versus-national gap you were counting on is not a real edge.

Start with the comparison itself. A win rate is only meaningful when the case mix matches: same technology category, same appeal stage, same claim type, same time window. Pull the N.D. Cal. figure and the national figure for 2019–2024 and confirm both are drawn from the same population before you subtract one from the other. If the two totals cover different case types or different periods, the difference you compute is an artifact, not a signal. Recompute the division yourself: wins divided by decided cases, with pending cases excluded from the denominator.

Second, the numbers that matter are totals and terms, not headlines. Compare the complete count of decided appeals on each side, the share that ended in an outcome favorable to the applicant, and the elapsed time from filing to decision. A higher rate on a tiny denominator is noise. A lower rate on a large, well-matched denominator is information. Write both figures down side by side before you form a view.

Third, verify the live, complete option before you commit. Docket composition shifts, and a rate published for one window does not bind the panel hearing your case now. Check the current docket for the specific appeal path, confirm the operative rules and any fee in effect at the time you file, and confirm the deadline that applies to your case rather than a remembered one. Where grounding supplies no verified figure, state the mechanism and the check instead of a number.

Local prosecution practice changes outcomes mainly through these three levers: how claims were originally drafted and amended, how the record was built before appeal, and how the appeal itself is framed. That is why a Berkeley software patent can post a different local rate than the national average without any difference in the law. The practical rule: match the populations, do the arithmetic, then verify the live option — in that order, every time.

Common Mistakes

One of the most frequent errors practitioners make is treating a PTAB appeal win rate as a fixed property of the tribunal rather than a moving statistic tied to specific technology centers, art units, and time windows. A Berkeley software patent applicant who pulls a headline reversal rate from a general PTAB dataset and applies it to their own case is comparing unlike populations. The correct check is to isolate decisions from the same technology center, filter by the same claim type (method vs. apparatus vs. computer-readable medium), and confirm the date range matches the period in which the examiner's rejection was issued. Without that filtering step, the win rate is noise.

A second mistake is committing to an appeal track before verifying the live status of the application. An applicant may file a notice of appeal believing the case is ripe for PTAB review, only to discover that the examiner has issued a final rejection that was never properly answered, or that an RCE was filed after the appeal brief deadline. The practical rule is to pull the file wrapper from Patent Center or Private PAIR and confirm three things before any appeal filing: the date of the final rejection, whether an advisory action was issued, and whether any after-final amendment was entered. Skipping this verification step wastes months and can forfeit the appeal window entirely.

A third pitfall involves conflating a PTAB affirmance with a terminal outcome. Some practitioners read an examiner-affirmed decision as the end of the road and abandon the application, when in fact the applicant retains the option to pursue a pre-appeal brief conference, file a continuing application with narrowed claims, or seek rehearing before the Board. The mistake is treating the first adverse decision as dispositive rather than as one step in a sequence. The check is to map every available post-affirmance path before deciding to walk away, and to confirm whether the examiner's reasoning leaves room for a claim amendment that would avoid the cited art.

Finally, applicants sometimes fail to account for the procedural posture that distinguishes N.D. Cal. appeals from those in other districts. A Berkeley software patent that has been through multiple rounds of prosecution may carry a prosecution history that narrows the available arguments on appeal. The mistake is assuming that a strong claim scope survives intact through years of amendments. The check is to read the complete prosecution history, identify every disclaimer or argument made during prosecution, and confirm that the appeal brief does not contradict any prior position taken before the examiner. Inconsistency between the appeal brief and the prosecution record is one of the fastest routes to an affirmance.

Insider Tactics

The tactic most applicants miss is sequencing the appeal around the examiner's own claim construction. Before you file the notice of appeal, map every rejection to the specific claim terms the examiner relied on, then draft your brief so that the construction dispute — not the prior art — leads the argument. PTAB administrative patent judges on software panels tend to resolve appeals at the construction level first; if your construction wins, the art-based rejections often collapse without separate treatment. That ordering is a drafting choice you control, and it costs nothing but brief architecture.

Timing tip: file your appeal brief only after pulling the live docket of the panel assigned to your case. Panel composition at the PTAB can shift, and the precedential appetite of a panel changes with it. Check the PTAB's decisions page for the panel's recent ex parte outputs in the weeks before your brief is due, and calibrate how much of your argument leans on precedential decisions versus persuasive-but-nonprecedential ones. If you cannot verify the panel's recent output, write the brief so it stands on precedential authority alone — that version survives any panel shuffle.

A second non-obvious move is to use the pre-appeal conference strategically rather than ceremonially. Many Berkeley-affiliated applicants with software claims skip the examiner interview before appeal and go straight up. Instead, request the pre-appeal conference and frame your presentation around the single narrowest construction issue. If the conference does not move the examiner, you have now created a clean, documented record showing the examiner considered and rejected your construction — which reads far better to the Federal Circuit under § 141 if you later get there than an appeal filed from an unexplored interview record.

On local practice: before assuming anything about Northern California prosecution norms, check your own file wrapper. Whether local cases tend to arrive at the PTAB with heavily amended claims is an empirical question — pull the prosecution histories for the N.D. Cal. software appeals in your comparison set and count the after-final amendments yourself rather than relying on an unsourced generalization about local culture.

That history matters for tactics — if your application was narrowed repeatedly before appeal, your strongest insider play is to argue the amendments as evidence of patentable distinction, folding the prosecution narrative into the brief rather than treating each rejection as a fresh fight. Verify this against your own file wrapper before committing; the pattern is a tendency, not a rule, and your wrapper is the only record that counts for your case.

Finally, before you commit to the appeal track at all, run the like-for-like check the rest of this guide describes: pull the live PTAB decision data for the relevant period, filter to software-class ex parte appeals, and compare against the national average using identical claim types and rejection grounds. Only after that comparison confirms the track is worth it do the tactics above pay off — sequencing, panel docket checks, and pre-appeal framing amplify a good bet; they do not rescue a bad one.

Comparison

Put the two tracks on one line and the choice usually makes itself. As explained above, the routes are legally distinct: the appeal under 35 U.S.C. § 141 goes to the Federal Circuit on the administrative record, while a § 145 civil action — the route that can land a Berkeley software filer in the Northern District of California — is a district-court case in which new evidence may, within limits, be introduced. The national benchmark is the aggregate PTAB appeal outcome set for the same 2019–2024 window. Do not label a district-court figure as a § 141 appellate figure or vice versa; the two are not comparable.

The winner is whichever track leaves you with a live, complete option at the lowest total term — not whichever one posts the prettier headline rate.

Run the side-by-side in four columns: (1) forum, (2) like-for-like disposition count, (3) win rate computed on that same count, (4) total term to a final, executable outcome. Column 2 is where most comparisons fail. A PTAB “win rate” is only meaningful when the numerator and denominator describe the same event — same claim set, same rejection grounds, same procedural posture. If the local figure counts only appeals that produced a written decision while the national figure folds in cases that settled, dismissed, or were remanded, you are dividing two different things. Recompute both rates from the raw counts before you trust either.

When each option wins. The local track wins when your prosecution record is clean: the rejection rests on a legal question the Federal Circuit reviews de novo, the claim set is stable, and the total term to a final outcome is shorter than the national path. The national benchmark wins when your record is messy — claim amendments late in prosecution, overlapping grounds, or a factual dispute the PTAB already resolved — because the aggregate rate reflects a broader mix of postures and is the more honest comparator for a case that looks like the average one. Neither wins on rate alone; the winner is the track that still gives you a complete option after the term is counted.

Apply the rule before you commit: pull the live docket status for your specific appeal, confirm the case is still pending and not already decided, then recompute both rates from matched counts. If the local count and the national count cannot be matched, treat the comparison as unverified and do not commit. A verified like-for-like total beats a favorable-looking percentage every time.

What to do next

Frequently Asked Questions

Does the 41% win rate apply to all patents in the Northern District of California?

No, the 41% figure applies only to N.D. Cal. Berkeley software cases, not all patents or districts.

Can I use the national PTAB appeal win rate as a benchmark for any patent matter?

No, the national baseline should only be used when the patent pool, appeal type, and time window match the Berkeley software set.

What should I check before applying the local win rate to another matter?

Check that local prosecution practice and the PTAB appeal record align before applying the figure to another matter.

What must be confirmed before relying on the comparison of local versus national rates?

Confirm the N.D. Cal. count, the national count, and the win definition before relying on the comparison.

Is any win rate treated as confirmed in this guide?

No specific rate is treated as confirmed until the case counts and win definitions behind it are checked.

What happens after a PTAB decision on an ex parte appeal?

A PTAB decision on an ex parte appeal does not end the matter; it splits the path, and the applicant may appeal.

Quick answers

Also worth reading: Geely seeks global FRAND rate determination for Nokia cellular patents in China: Geely seeks global FRAND rate · Evolving Trends in Patent Claim Construction From Phillips to PTAB's New Standard: Evolving Trends in Patent Claim · Federal Circuit Confirms Stability of Fintiv PTAB Standard: Federal Circuit Confirms Stability of