How AI dick rating actually works

One photo in, six numbers out

A scan returns a score out of ten, and almost everyone stops reading there. The score is a summary of six separate judgements: size, shape and proportions, skin texture, head, overall appeal, and erotic impact. Every one of them is scored on its own, and the breakdown is the part worth reading.

The reason it is split that way is that the six axes genuinely do not correlate. Two photos can land on the same total by completely different routes, and the breakdown is the only thing that tells you which route yours took. You can see this for yourself on the leaderboard - tap any entry and the six-axis chart behind it opens.

What each axis is actually measuring

Size is the one everyone assumes carries the whole score, and it does not. It is one axis of six, and in practice it is the axis with the least spread across the board - most submissions cluster, so it separates people less than any of the others.

Shape and proportions is about how the parts relate to each other rather than how big any of them is. This is the axis most sensitive to the angle you shot from, which is the single biggest thing you control.

Skin texture reads surface detail: tone, evenness, condition. It is the axis most punished by bad lighting, because a hard flash flattens texture into glare and the model scores what it can see.

Head is scored separately from shape for the obvious reason that it varies independently of it.

Overall appeal and erotic impact are the two subjective axes, and they are where the model is doing something closer to taste than measurement. They are also the two that move most between a good photo and a bad one of the same subject.

Why your total is not the average

The six axes are not weighted equally, and the total is not their mean. Someone scoring high on the two subjective axes and mid on the four physical ones will beat someone with the reverse spread, which is not what most people expect going in.

This is worth internalising before you read your result as a verdict on anything. The score is a photo being scored, not a person, and the axes most responsible for moving it are the ones most affected by how the photo was taken. If your result surprised you, the fix is usually the camera, not the anatomy - the photo guide covers what actually changes the number.

The model is not the only judge available

The AI is fast, cheap and consistent, and it is completely uninterested in you. That is a feature for a benchmark and a limitation for anything else.

If you want the result written in a specific voice, Custom Persona lets you set the tone, intensity and focus areas the write-up is generated against. If you want a person rather than a model, the judge panel is the human side, and that comparison is its own article.