Scores
When the tool declines, and what that means
A rejected submission is a statement about the image, not a hidden zero; reading it as a bad result is a category error.
Guides on Scores: How much weight a rating tool result deserves, Every part of a rating result, and what each one is for, What can and cannot be compared, and how
No - "could not score" is not a low score. It means the tool declined to produce a number at all, usually because the image was unclear or outside what it evaluates, so it says nothing about where a score would have landed.
Two different outputs
A score is the tool's judgement about a submission it was able to evaluate. A rejection is the tool declining to produce a judgement at all, because something about the submission fell outside what it will attempt to evaluate - the image was unclear, ambiguous, outside its accepted category, or triggered a refusal condition entirely unrelated to quality. These are not two points on the same scale. One is a result; the other is the absence of one. Machine learning research treats declining as a deliberate design choice, the "reject option": Geifman and El-Yaniv (2017) showed an image classifier could guarantee 2% top-5 error on ImageNet by answering only about 60% of inputs and rejecting the rest.
Nothing about "could not score" tells you where a would-be number would have landed. It might have scored a 6. It might not have been scorable in any coherent sense, because the image did not contain what the rubric needs to evaluate. The rejection carries no information about that number, because no number was ever computed.
Why this gets misread
A results page that fails silently reads, emotionally, exactly like a bad result: no number, no praise, an interruption to what you expected. That similarity in feeling is doing all the work in the misreading. Structurally, a rejection is closer to a form validation error than to a low grade - the tool is saying "I cannot answer this," not "the answer is low."
Treating a rejection as a bad score also produces a bad response to it: resubmitting a worse or more manipulated version to try to force a number out, when the actual fix is usually a clearer photo under conditions the tool can parse, which is the entire premise of treating standardisation as separate from flattery.
What actually causes a rejection
That is a longer subject on its own - what a good tool refuses to score, and why covers the specific triggers. The point here is narrower and comes first: whatever the cause turns out to be, it is not evidence of a bad underlying number, because no underlying number was ever produced.
A rejection from an automated tool is also a different event from a human reviewer declining a submission - a person's refusal usually comes with a stated reason, which is a different kind of information again - and it has nothing to do with a physical measurement, where there is no equivalent "could not measure" state once a tape is actually on the subject. Whether an image triggers a rejection is a property of the model layer itself, decided upstream of anything visible on the results page. Rate Cock treats a rejected entry as absent from its public board rather than displaying it as a floor score, which is the correct handling of the distinction this piece is making.