Arguably, a prediction depends on a ranking but contains less and more data. Less data in that only part of the intrastudent rank is maintained, more in that the ranks are tied to grades.

It would seem that you agree that teachers cannot be expected to produce fair absolute grades, but will produce fairer rankings - which is supported by the articles references.

In terms of the objection being obvious - it is obvious in a political sense that people would quarrel with the prediction being dropped as it both favours them and is more personal. Mathematically its presented as obvious that this is preferable [1].

[1] personaly I object in principal to us estimating grades full stop. I'd have preferred spending the last five months rewriting exams based on covered material and then a mammoth effort at getting kids to sit them safely. However that requires foresight and funding.

It's clearly a massive kludge - those rankings are only estimates, lots of kids would have performed better or worse than estimated, and effectively they are being granted grades based on a whole load of factors other than their own actual performance. I feel very sorry for them.