Crime Statistics

This is an analogy, not an accusation

I am not saying anyone at the USPTO sat down and decided to suppress complaints. If anyone is to blame here, it's Congress, which set up this system and the burdensome fee that falls on the victim.

I have no evidence of any active suppression; I am not alleging it; and if you read this dispatch as an accusation, you have read it wrong. What I am saying is that the system produces a familiar-looking result, and the analogy is the fastest way I know to explain how.

It sure feels like a crime was committed, but that's just my opinion. What I see as plagiarism, mischaracterized prior art, a sloppy examination and an apparently hasty, unproofread allowance. All of this results in harm to me, but is it a crime? Legally, I don't know. I do know there is no administrative relief aimed at the examination itself — no one at the Office reviews whether the examiner was misled or careless. The only administrative handle is re-arguing the patent's validity from scratch, at my expense. No way to report this and feel like it was worth reporting. The analogy here illustrates how stats can be skewed by burdening the reporting of the very thing you want to call insignificant.

The analogy

Suppose you run a police department — or a city, or a whole state — and you want better crime statistics. There are two ways to get them.

The hard way: prevent more crime, solve more cases, and earn enough trust that people actually bother to report. It doesn't look good at first — the numbers go up for a while, because you're finally counting what was always there. Only later do they come down, and when they do, they've come down for real.

The cheap way costs nothing. You do not have to touch a single case or improve the system. You only have to make reporting harder — or leave reporting technically available and let people work out for themselves that it is not worth the trouble. Crime does not go down. Reported crime goes down. And reported crime is the number that gets published.

The PGR is where this lands. The cheap way to improve the outward-facing appearance of competence — cheap for the system, not for the victim — is to make reporting so difficult and painful that few reports ever make it through.

Now put a price on the report

Imagine having to pay a fee to file a police report. I bet the crime stats would drop to practically zero. And that's roughly where the USPTO lands on examination errors — except you can't even file examination errors as a complaint. There is no form for "this examination was careless." You have to buy a validity challenge and hope the carelessness shows through it. (Again, not an accusation, an analogy… but a pretty good one.)

At what point would police reports be financially unbearable for the victim? $60,000 is an unbearable fee for any individual — e.g., the independent inventor. This makes reporting structurally out of reach for the folks (like me) who did the hard work: invented the technology and disclosed it thoroughly. This is part of why innovation in the US is dying a slow, sad, and silent death.

No errors, or no reporting?

The USPTO takes pride in its low error rate. I once heard a senior official tout how few IPRs stemmed from examiner error. That's much like a mayor touting a low crime rate: it skews the whole discussion.

Examination error is a hard number to pin down, because an IPR isn't an error audit — the patents that get one are already self-selected as the ones somebody thought "worth investigating." IPRs and PGRs aren't cheap, nor are they plentiful (less than 1% of issued patents… much less than 1%). They are merits proceedings, and they only ever run on the patents somebody paid to challenge. A sloppy examination can still land on perfectly valid claims — no challenge, no entry in the statistics. A sloppy examination can also land on terrible claims, and it still produces no entry in the statistics until someone like me spends the money to prove it. From the metric's point of view, those two examinations are indistinguishable. What the number actually measures is not how often the Office is wrong. It's how often somebody could afford to say so.

Here's what "affording to say so" looks like. The petition is capped at 18,700 words of dry legal drafting, essentially none of which is about examination errors — the petition has to argue the claims are invalid, not that the examination was careless. For comparison, this site is about 60,000 words, and none of those count for anything.

Then there's the fee: about $60,000. Roughly a dollar for every word on this site. I'd laugh if it weren't the whole problem. (Yes, there's a cheaper cousin — ex parte reexamination — but it's narrower: patents and printed publications only, no § 112, you're a spectator after filing, and it still costs thousands. More on that in a coming dispatch.)

I don't have $60,000 to feed into the PTAB incinerator. So unless somebody at the USPTO happens to read this blog, this error-ridden examination will never appear in the Office's error statistics — not as a complaint it rejected, but as no complaint at all. Its numbers will be exactly as good as they'd be if the examination had been flawless. It wasn't flawless. See for yourself.

Why would anyone want this?

If you want people to believe innovation in the US is thriving, you certainly don't want to admit the flaws. Just like the mayor who wants to tout how great their town is, you don't want people pointing out all the crime. The arrangement effectively buries the sloppy-examination problem, while simultaneously making it look like the US is thriving. It's not. This is one example of why.

Up next: the petition

My time is nearly up. The deadline to file is days away. I drafted the petition, but who can personally afford $60,000 to ask the government (nicely) to fix its own errors?

Related dispatches

• Dispatch No. 2 — Why a PGR? — the ~$60,000 price of the only full-strength way to report.
• Dispatch No. 6 — The Third-Party Preissuance Submission — the cheap reporting channel, and the "Considered" stamp that closed it.
• Dispatch No. 10 — Post-Allowance, Preissuance — What could I do? — the shrinking list of ways to report, dispatch by dispatch.
• Dispatch No. 12 — A Private Conversation, Held in Public — being the one person with the most to say and no standing to say it.
• Dispatch No. 26 — TL;DR; All the system failures — the full list of failures this analogy compresses into one idea.
• Dispatch No. 28 — Let's Fix It — the fixes that would make an objection countable.
• Dispatch No. 30 — Allowance Rates — the published statistic this dispatch asks you to distrust.
• Dispatch No. 31 — Word Count — the burden of the report itself, and the same "let the public log a concern" proposal.
• Dispatch No. 33 — The PGR Petition — the report I wrote anyway, published instead of filed.