« Contra-Validity | Main | Cock-up »

March 5, 2008


The current affirmance rate of the USPTO Board of Patent Appeals and Interferences (BPAI) yields the illusion that examiners deserve a pat on the back for a job well done, but behind every illusion lies reality. Let's peek at the man behind the curtain.

Professor Dennis Crouch goes stochastic, all the while thinking he's got the skinny:

The table below looks at board decisions for each technology center. The relative numbers are interesting - showing a low affirmance rate for rejection of business method cases (TC 3600) and a high affirmance rate for chemistry cases (TC 1700). The BPAI offers its own calculation of these numbers based on all of its decisions that are roughly parallel. [LINK]

(Note, the table does not add-up to 100% because I excluded decisions to remand and deny rehearing).

Thus, according to the BPAI, examiners write legitimate final rejections in 59% of the cases that are appealed. Yet, experience makes these numbers hard to swallow. Do prosecutors have unrealistic expectations of what constitutes allowable subject matter? Is the board completely biased in favor of the examiner? Or, are these numbers merely skewed?

The table is deceiving, as it fails to account for applications that are appealed but never make it to the BPAI. Many notices of appeal result in the withdrawal of the final rejection by the examiner, and never even reach the board.

Bear in mind that patent examination is a game; this is relevant to all PTO transactions.

But first, a quick refresher course on examiner productivity requirements. Examiners are reviewed based on four factors: production quantity (45%), production quality (35%), workflow management (10%), and customer service (10%). Production quantity is by far the most significant measure, not only because it is given the highest weighting factor, but also because it is the only objective assessment. Each examiner is given a production requirement in terms of hours per count. This requirement varies based on art unit (technology complexity) and seniority (GS level). One count is awarded to an examiner for a first office action on the merits, and a second for disposal of the application. Disposal includes abandonment and allowance. Additionally, a count is given when an RCE is filed, and the application is reset to have an additional 2 counts available.

The breakdown:
Restriction requirement = 0 counts
First action rejection = 1 count
First action allowance = 2 counts
Second action non-final rejection = 0 counts
Final rejection = 0 counts
Abandonment = 1 count
Allowance after attorney reply = 1 count
RCE = 1 count
First action rejection after RCE = 1 count
Examiner's answer to an appeal brief = 1 count

This makes obvious why patent examination is a game. Examiners have zero incentive to withdraw a rejection and find new prior art even when an attorney makes valid arguments against a first action rejection. Instead, examiners roll the dice.

Forgetting for a minute about compact prosecution, let's walk a short mile in an examiner's shoes. An examiner writes up a non-final rejection, rejecting all the claims over the prior art. The attorney responds and traverses all of the rejections. Let's assume the attorney makes some valid arguments and the examiner agrees that the rejections should probably be withdrawn. The examiner still has three options: 1) withdraw the rejections and allow the claims; 2) withdraw the rejections and find new prior art to write a second non-final rejection; 3) maintain the rejections and issue a final office action. Option 1 usually won't fly based on the current management mentality of "reject, reject, reject." Option 2 will chew through numerous hours for zero count. Option 3, although also zero count, takes less than an hour.

So, option 3 it is. Worst case scenario: the applicant files a notice of appeal, the examiner is forced to withdraw the final rejection, and then at that point allow the application or find new prior art. In this case, the only loss is the hour or so it took to write up the final rejection. Best case scenario: the applicant abandons the application or files an RCE. Sweet! Free count! And the examiner saved countless hours.

This constant gambling by examiners cause many notices of appeal to never result in an appeal that reaches the board. Additionally, the dice rolling skews the board affirmance rate, masking actual examination quality, but reaps sweet numbers for the PTO to parade to the public.

The game is rigged, and for those eating up the numbers like they were candy, congratulations to the PTO for pulling the examination wool over the eyes of the unknowing. Way to go!

Posted by Mr. Platinum at March 5, 2008 7:37 PM | Prosecution



Excellent analysis. PTO (mis)management is involved in lies, lies, and more lies.

The idea that a 59% affirmance rate represents "quality" is laughable.

Change 37 CFR to require the examiner, in response to an appeal brief, to 1) issue an examiner's answer or 2) allow the application (no re-opening, no new grounds), and then you'll see what the "quality" level of the examiners' work product truly is.

Also, require BPAI, to either 1) affirm, 2) reverse, or 3) affirm-in-part. No remands. No new grounds.

I can dream, can't I?

Posted by: JD at March 6, 2008 6:39 AM

Lucky me. Almost all of my work is in TC 1700, and the Board backs the examiners almost to the point of insanity.

Posted by: bierbelly at March 6, 2008 8:10 AM

Nice post.

Posted by: Michael Martin at March 6, 2008 11:56 AM

Your argument has serious problems.
You say "Examiner's response to an appeal brief: 0 counts".

WRONG. There is a disposal count for an examiner's answer.

That little fact throws a monkey wrench into your analysis. You must not have been at the office long enough to ever write an examiners answer or you would have certainly known about the disposal count. Also, the RCE count is not a separate count, it is merely the disposal count for the first go-round in the case. Once there is an RCE disposal, two more counts (a first action count and a disposal count) are available to the examiner for continued prosecution.

Posted by: surfer at March 7, 2008 6:54 AM


I appreciate your correction. I did make a mistake by stating "Examiner's response to an appeal brief: 0 counts". Thanks for the catch, but let me note that although an examiner does receive a disposal count for an examiner's answer, they no longer receive a disposal count for the abandonment or allowance that follows. Also, I did mention that an additional two counts are available after an RCE is filed.

More importantly, this does not change my analysis. An examiner's answer is very time consuming for the examiner, and whether or not they receive a count for this action is irrelevant. Obviously, an examiner would rather receive a disposal count from abandonment, allowance, or RCE, rather than from spending many hours writing a response to an appeal brief. So, the games continue.

I believe you somewhat missed the point, which is that the high affirmance rates given above are not a good indication of the quality of work coming out of the PTO, since many rejections are made final, only to later be withdrawn after the applicant files an appeal.

Posted by: Jordan Kuhn at March 7, 2008 8:57 AM


Not to quibble, but "Examiner's response to an appeal brief = 1 count" should really be "Examiner's answer to an appeal brief = 1 count."

If the examiner's response to an appeal brief is to re-open prosecution, there is no count.

But your analysis is correct despite this little "monkey wrench."

If the examiner were required to issue an answer in response to a brief (no re-opening and no new grounds), that 59% that PTO (mis)management is so proud of would be MUCH lower.

Unfortunately, many in the PTO are using the old, "appeal, re-open, appeal, re-open, appeal, re-open" scheme to avoid agency, and/or court, review of their actions, which is prohibited by the APA.

There are several cases that have received some notoriety lately due to this treatment by the PTO. I hope the applicants and practitioners involved will take it to a court and get a decision that the PTO can't avoid issuing the application, or even avoid a review of its illegal tactics, simply by keeping applicant in an endless loop of re-opening prosecution.

Posted by: JD at March 7, 2008 11:24 AM

Not all of the withdrawn appeals are withdrawn by Examiners. When I was an Examiner, 6 of my last 10 appeals were withdrawn after the Examiner's answer was written by the various Appellants and either abandoned or RCEd. I never withdrew an appeal. (Unlike my lawyers or examiners posting, I did not always "win". I had an affirmance rate around 70%, though if the abandonments after appeal over the years get included, it would be higher, close to 90%).

Consequently, while the PTO undoubtedly has all of these statistics, until they release them, it is hard to tell what is going on.

Posted by: Ex examiner at March 7, 2008 7:22 PM

An analysis that includes the effect of appeal conferences, but not pre-appeal review, is in my notice-and-comment letter on the appeal rule.

Executive summary: examiners are affirmed about 20% of the time - that is, examiners must concede (or are found to have committed) error (procedural or substantive) about 80% of the time.

The summary table is at page 65, the underlying statistics obtained from the PTO are at pages 51-65.

Posted by: David Boundy at March 10, 2008 9:51 AM

No reopening prosecution? Haha, sounds good. But then the PTO might want to ban all claim amendments just to be fair. Everyone either gets it right the first time or else...

Posted by: sammich at March 23, 2008 8:20 PM

What is also sad is how the USPTO lets foreign patents just waltz through with little effort or research. I know this is the case because a foreign patent was granted, and it was a rewrite of a patent I did 20 years AGO! Needless to say I am having hell with an examiner on a new application that I doubt he has even read and if he did he did not understand it! A review of all of his cases I see reject reject final reject appeal granted.
The guy told me on the phone granting patents hurts corporate business and trade.
The USPTO was not that bad years ago but now with publication of pending patents and the reject American inventors while allowing anything from overseas to pass without nothing more than a modest review it is clear that thanks to the USPTO America is now a horrible place to be an inventor as all you end up doing is educating corporations and copy cats while the inventor is bogged down in red tape. The steal American technology act and google eliminating the patent pending along with the reject and meet the 5 hour quota mmake progress a lost cause.
I hope Gov Richardson fixes the USPTO and puts it back in order fast.

Posted by: John at January 4, 2009 2:12 PM