Why the lens you did not choose changes the number

Yes, focal length is a hidden score variable: most phones have two or three cameras behind one shutter button and switch between them by zoom level without announcing it. Each lens frames and distorts differently, so the score can move between two photos where the only thing that changed was which lens the phone quietly used.

A second variable that looks like the first one

Distance changes proportion through perspective: whatever is nearest the camera is enlarged relative to what sits further back. At face scale the effect is large - Ward and colleagues (2018, JAMA Facial Plastic Surgery) calculated that a photo taken 12 inches away makes the nose appear 30% larger in men and 29% larger in women than an undistorted projection, with essentially no difference at 5 feet. Focal length changes proportion more indirectly: a wider lens takes in more of the scene, so people stand closer to fill the frame, and ultra-wide lenses also stretch shapes toward the edges of the frame in a way the main lens largely avoids. The two effects can look the same in a result: the frame reads differently, the score moves, and nothing about the subject changed. That is what makes focal length the more dangerous of the two - distance is at least something you chose; which lens fired is frequently not.

Where the switch happens without asking

Pinch to zoom past a certain point and most phones hop from the main lens to a telephoto one, or the reverse toward an ultra-wide. Front and back cameras are different lenses entirely, usually with different fields of view and different processing pipelines behind them. Some camera apps switch lenses automatically based on lighting or subject distance, which means the same framing gesture can select a different lens on two separate occasions. None of this is visible in the resulting photo unless you know to check the file's metadata, and most people never do.

Why a protocol has to pin it down

The fix is not complicated once it is named: use the same lens every time, chosen deliberately rather than by whatever zoom level felt natural in the moment. For most phones that means the main camera at 1x, unassisted by pinch-zoom, because it is the one lens guaranteed to be present and consistent across software updates. A submission held constant on distance and angle but inconsistent on lens still is not comparable to itself - the variable just moved somewhere less visible. Note which camera and which zoom setting were used, the same way distance and angle get noted, and the note is what makes the next photo a genuine repeat rather than a coincidence.

What sits underneath this and what does not

The reason different lenses render proportion differently is a lens-geometry question with a name - crop factor and field of view, defined properly, is where that belongs, not here. This post is only about the practical fact that the lens is a variable and it switches without telling you. What the model behind a rating tool is actually doing with whichever frame it receives is a separate question again, and a longer one than a protocol note needs.

Why this is easy to miss and hard to undo

Nothing on a phone's screen tells you which lens just fired. The preview looks continuous as you pinch to zoom, sliding smoothly from wide to tele, but the underlying hardware jumps between two or three fixed lenses at specific zoom thresholds, and those thresholds are not the same across manufacturers or even across a single manufacturer's model years. A person trying to reproduce "the same zoom level as last time" by eye is trying to reproduce a number they never saw and that may not have meant the same thing on both occasions. Once a photo is taken, there is no way to recover which lens produced it from looking at the image alone - the only record is in the file's metadata, and most sharing paths strip that on the way out.

Front camera as its own case

Switching from the back camera to the front one is not a smaller version of this problem, it is a completely different one. Front cameras on phones typically use a wider field of view than the main back lens, because they are designed for close-range framing in a hand-held selfie shot, and the image processing pipeline behind them - sharpening, skin smoothing, colour - is frequently tuned differently from the back camera's pipeline as well. A submission taken on the front camera is not comparable to one taken on the back camera even at the same physical distance, because two variables changed at once: the lens geometry and the processing behind it. If a protocol calls for the back camera, that has to be a fixed rule, not a default that gets abandoned for convenience on days when a mirror or a second pair of hands is not available.

The gear side of it

None of this is an argument for buying a better phone. Camera choice belongs to the gear side of a setup more than to the score side of it, and the lens that matters here is not the best one available on the device, it is the one used consistently. An older phone used the same way every time outperforms, for comparability purposes, a newer one whose lens selection changes from session to session.

A note for anything sent to a person

The same inconsistency shows up if the recipient is a judge rather than a tool. A photo sent for human review still carries whatever proportion the lens produced, and a viewer reacts to that framing the same way a model would score it - the fix is identical either way: pick the lens, use it every time, write down which one it was. Rate Cock's guide to taking the photo assumes the main lens at 1x for exactly this reason, and it is a reasonable default to carry into any tool, not only that one.