[ivanbakel]

>... and this was portrayed in the press as proof that it was an attack on state schools because their results were disproportionately affected by the algorithmic downgrading.

While I won't dispute the point on prediction accuracy, the press complaints were actually about the algorithm itself, which benefited students in a smaller corpus[0]. Typically, only private schools have such small class sizes.

[0]: https://www.bbc.co.uk/news/explainers-53807730