@Railingsohno
It's not so much being slack necessarily as space and time.
At my small school we'd have exams in full exam conditions twice a year, in our large hall that was really only used for assemblies. Teachers invigilated and we had a different timetable over the fortnight of the winter exams and for exam period in the summer.
At my dc's large comprehensive they need the sports hall and main hall just to get all the GCSE students in. They have outside invigilators and can't have the whole school doing exams at once, so only the exam years have changed timetables. Is it any surprise that they choose to do exams in a normal lesson with their normal teacher? It's simply a matter of logistics and finance.
I think NewModel is right in that it will again advantage the small schools who have the space to do proper mocks.
I wonder if they could have added to the algorithm that any grade that was 2 or more out from the CAG was then looked at individually (or maybe the cohort if it was large numbers for one group). They could have used the normal markers to help here, and I think they said it was only about 2%, so we're not talking about huge swathes.
Any mark that was 2 grades or more out, the school would have been asked for evidence for that specific pupil. Now they didn't have to say they were doing that-could have made out it was random checks.
It would have meant, I suspect that most of the two grades out would have been checked and probably gone to closer than the correct grade. They could have even done it in large cohorts that then means they look at all those above as well. It would have put a safety net in, and also on any small cohorts as a few people have mentioned where the average results have shot up by 2+ grades these would also have been looked at and modelled down if they weren't fair, and then people wouldn't be looking at them sideways if they were a fair representation.