« Taxing Patent Reform | Main | Pop »

November 20, 2007

The Big Charade

The USPTO released last week its fiscal year 2007 "Performance and Accountability Report", praising itself for "achieving another record-breaking year in performance." The numbers: a record number of patents "examined" = 362,227; quality compliance = 96.5%  (what the hell is that?: dotting the 'i's and crossing the 't's, a triumph of nitpick over substance); BPAI upholding examiners 69% of the time (an intra-agency highly symbolic number), up from 51% in 2005, indicating how much better the PTO has stacked the deck against applicants; and grants down to 51% from 72% in 2000, as if, nationwide, patent attorneys' heads turned to mush as they suddenly spewed garbage patent applications left and right. Yeah, right. All puppet performance, no accountability.

Beyond the numbers is the big lie regarding quality - as Peter Zura succinctly socked it to us:

The PTO seems to be pleased that applications are being rejected. Just as Thomas Blackstone said that "it is better that ten guilty men escape than one innocent suffer," the PTO's current mood appears to be that "it is better to reject ten legitimate inventions than to issue one bad patent."

Jordan Kuhn, former USPTO examiner, now soaring as a Patent Hawk, provides insider insight on the examination regime. [A current examiner remarked of Jordan's step up: "Glad to hear that some examiners manage to escape the PTO compound."]

The USPTO has adopted an “all applications are rejectable” mentality.  Examiners are provided with countless hours of training in the intricacies of 101, 102, 103, and 112 rejections, both during the training academy and throughout their careers, while even a brief discussion of allowability is considered a PTO faux pas.  I received maybe an hour of training regarding allowability.  This hour, prefaced with: this isn’t something you will need to worry about for some time, but….  Many examiners are prohibited, by supervisors, from allowing any applications for at least their first year.  Since new examiners aren’t officially, albeit often unofficially, given a production requirement, they are told to search and search and search until something pops up.  In theory this provides a thorough examination, and might be necessary considering the inexperience of new examiners.  In reality this leads to countless bogus rejections that do not make the least amount of sense and lead needlessly to wasted time and money of the applicant.

It is a sad state of affairs when patentability is determined more by which examiner/supervisor combo is assigned an application than by actual novelty of the invention.  Some applications have zero chance of becoming a patent, regardless of content, before ever being read by an examiner.  Other applications, which have about as much novelty as bad management at the PTO, slip right through to patent, due to misleading claim construction and crafty claim drafting.

Supervisors that have this ingrained “all applications are rejectable” mentality often force examiners to reject, against the better judgment of the examiner who actually spent hours researching the technology, all applications, especially if the claims are the least bit broad.  Have a brilliant novel concept, years ahead of its time?  Want to write some broad claims that are rightfully yours for the taking?  Well, that is an appeal waiting to happen.  Somehow supervisors believe it is worth clogging the appeals board with obviously novel yet broadly written claims in order to transfer responsibility away from their name.

“Quality compliance was 96.5 percent” - another conveniently made-up number. When examiners are told during the last two weeks of the quarter that quality review is no longer reviewing allowances for the quarter, so if you need to bump your production up….   How is this an accurate measure of quality?

When quality is based on review of allowances and not of rejections, so that the thousands of ridiculous examiner rejections are never given a second look, how is this an accurate measure of quality?

When quality review specialists review applications from a wide range of technologies, how can they possibly provide a better examination than an examiner who specializes in the particular art of the invention?


A Patent Prospector reader writes:

1) Has the PTO ever tried to measure the percent of rejections that are improperly imposed by examiners? Or is bad quality just mean a poor job in rejecting applications?  It should be both, because when the PTO issues shoddy examiner rejections, that imposes costs on the applicant and discourages invention.  To my knowledge, the PTO has not investigated this question.

a) It is quite possible to measure, of course.  One good metric of improper rejections would be: How often has the examiner withdrawn a rejection after pure argument by applicant without any claim amendment?  If a rejection is made, an argument comes in response, and the examiner withdraws the rejection and either allows it or shifts to a different rejection, then the first rejection was improper and shouldn't have been made - it should be a "mistake" counted against quality. 

b) Another metric might be:  How often has the examiner shifted from one rejection to another without intervening claim amendment?  For example, I just received allowance of a patent application after six different non-final rejections, despite my having never amended the claims (until all allowable subject matter was indicated, at which point I amended merely to put the allowed claims into independent format).  Does that count against the examiner in her quality review?  No.  Unfortunately.  (I know it kills her productivity, but she blames me for that for arguing too much, not herself for making bad rejections.)

2) The PTO touts the fact that its internal rate of affirmance by the Board has skyrocketed, from about one third of appeals to about two thirds of appeals.  But the comparison uses a year before the PTO set up its pre-appeal -brief conference procedure.  Many of the bad appeals are recently being removed from the system through initial screening.  That is a good thing, I agree - I like the pre-appeal-brief process better than going through the whole appeal - and PTO deserved kudos for it.  But don't try to tell me that the affirmance rate has increased. 

a) How about issuing some meaningful statistic on appeals?  For example, I'd like to know the following percentage:  Let's look at all the times a notice of appeal was filed and see what fraction exists where the PTO stuck with the rejection (i.e., no new ground of rejection by the examiner or the Board) and succeeded in securing affirmance by the Board - and by the way, let's not count as quality "wins" (i) cases where the Board affirms on different grounds or enters a new ground of rejection, (ii) cases where the Board reverses some of the rejections but not others, especially when some claims are allowed, or (iii) cases (likely only a few) where the Board was reversed by later proceedings in the Fed Circuit or D.D.C.  Such a stat would accurately report the percentage of cases where the PTO was justified in forcing an applicant to appeal.  My guess is that the correct percentage is no higher than it once was, and it may be even lower given the pressure to reject applications lately.

In all, the PTO's report is pure gamesmanship using meaningless statistics of the worst sort.

Posted by Patent Hawk at November 20, 2007 2:59 PM | The Patent Office


"One good metric of improper rejections would be: How often has the examiner withdrawn a rejection after pure argument by applicant without any claim amendment?"

Be careful what you wish for. Imagine if they started measuring this, wouldn't it add another dimension to the examination game? You could no longer just argue improper rejections... you'd have to trivially amend or face the added constraint of knowing the examiner will be facing an "error" if he accepts your arguments, so he is less likely to accept them. [In essence, the threat of quality review as you set forth would both hinder him from correcting mistakes previously made and hinder him from making purposeful mistakes in the first place, unless you played the game, but the net effect of that on examination isn't clearly positive to me.]

There must be some game theory that covers this situation...

Posted by: NIPRA anonymous at November 21, 2007 11:27 AM

Newbie examiners are being taught: Quality = Reject, Reject, Reject.

That is not their job. Their job is to make objective findings of fact and make legally sound patentability determinations based on substantial evidence.

The truly sad thing is that once you damage these newbie examiners with (mis)training straight from PTO (mis)management (REJECT, REJECT, REJECT), then they are essentially useless. Not even Room 101 can turn such Winston Smith's into useful examiners. They turn into these "My job is to build a record" type examiners. Utterly useless.

To my friends at NIPRA, you should be aware that many examiners are taught, and believe whole heartedly, that "attorney argument," by itself, is not sufficient for withdrawing a rejection and allowing the application. In other words, some of the idiots over there already actually believe that that is an "error."

Posted by: JD at November 21, 2007 1:46 PM

Anyone heard of the pencil test? It is a semi-joke around the PTO amongst examiners, but some examiners use it. If a claim is longer than your pencil than it is allowable. The idea is that your pencil starts out new at the beginning of each quarter, but by the end is just a nub. Another fine example of consistent quality.

Posted by: Jordan Kuhn at November 21, 2007 2:27 PM

"Anyone heard of the pencil test? It is a semi-joke around the PTO amongst examiners, but some examiners use it."

Yep. And when I was an examiner, one of the primaries had what we called an "allowability gauge" that looked a little like a caliper. If you could slide the application file in and out without contact (because the application was thin), it was allowable (a sort of positive reinforcement/incentivizing for applicants, not too unlike the type being promulgated currently by USPTO management).

Posted by: NIPRA anonymous at November 21, 2007 2:50 PM

P.S. And just so everyone knows, the allowability gauge was a joke (though it did exist).

Posted by: NIPRA anonymous at November 21, 2007 3:00 PM