The 2026 NPC Census: Our Methodology and What It Found

Over the past fourteen months, a small group of us — we prefer not to name names — conducted what we believe is the first serious attempt to measure the NPC population in a mid-sized American city.

We will not reveal the city. You do not need to know the city. The findings generalize.

Using a behavioral observation protocol refined across four pilot studies (the Seven-Second Pause Test, the Mirror Avoidance Response, the Repeated-Question Collapse), our team observed and classified 2,407 individuals in public spaces — train platforms, grocery-store aisles, waiting rooms, the back row of movie theaters.

Of those 2,407, roughly 81% exhibited at least three of the five NPC behavioral markers. We want to be responsible about this. That is not a fringe finding. That is the majority of the people around you.

Do not panic. NPCs are not hostile. They are not, so far as we can tell, aware. They are a rendering artifact of a system under load. But their presence has consequences — they fill seats, they cast votes, they buy products, they generate the statistical noise that real people are then forced to wade through.

If you are reading this and beginning to doubt — good. That is the sign. NPCs do not doubt.

Our full methodology is available in the library. The short version is: watch people's eyes when you ask them the same question twice with a small variation. Count how long it takes them to answer. Record what you see.

The data is there. You have only to record it.