Reference-context engagement scorer

One photograph, placed in a measured reference population of 91,810 posts. The instrument reports where it falls and how much its own ensemble disagrees.

What this number is

Predicted engagement for this image posted solo, as the thumbnail, by a standardized account, under this population's revealed preferences — with page reach, follower count, posting history, carousel position and source compression held fixed.

Holding those fixed is the point. Without it, a score mostly measures how many followers an account has. The reference context pins every one of them to the same value for every image, so what varies between two scores is the photograph.

What it is not

Not an objective measure of anything.
It reports one population's revealed engagement behaviour. That is a fact about the population, not a fact about the person in the photograph.
Not a photo-selection tool.
Multiple photos of the same person will score close together. This is measured, not a guess: three instruments were built to detect within-person photo responsiveness, and none of them detected it. If you upload five pictures of yourself hoping to find the best one, the instrument cannot tell you — and the spread you see between them is mostly its own noise.

How much to trust one number

The ensemble band is typically wide — a median of 39 percentile points between the lowest and highest member, measured over 200 held-out images. The band is shown at the same weight as the estimate because it is the same size as the estimate: on most images the honest reading is a third of the scale, not a point.