Selective schools in the state sector don't set their entry bars especially high for the sixth form, so the grades required shouldn't be a particular problem.
The one I know of has a low entry bar BUT it only offers that low entry bar to those with the highest predictions.
So it ranks all external applicants by predicted grades, then only offers places (with, I agree, a relatively low entry bar) to the highest 20 [or whatever]. Where there s no uniform approach to predictions, that top 20 could be entirely randowm.
I also ask again; I know that your DD's school SAYS that when they give a 9 for a test, they THINK it will be correlated to a 9 at GCSE - but how do they know? Do they have a statistically significant number of pupils such that they can predict the grade boundaries based on the % that the government have released? Or are they going on 'gut feel and approximate linkages to old grades' - ie they might previously have predicted a high A on the old curriculum for such a pupil, so they will say it is a 9, or they would have predicted a low A so it is an 8?
I agree that such an approach can give a 'range' of results: 'This is definitely more 8ish than 5ish', but to say 'Yes, this is definitely a 9, despite changes in curriculum, assessment and grade boundaries' is something i suspect they are MUCH less confident about than you are.