Part of the calculations used to decide a university's benchmark
Universities which have to choose between well-qualified applicants say their "benchmarks" for admitting pupils are fundamentally flawed.
Annual statistics show how many students they admit from state schools and from poorer backgrounds.
These tend to highlight universities which were, on the face of it, wide of the mark - missing their benchmarks significantly.
They say the true picture is very different, if the benchmarks are adjusted to reflect the way they actually admit students.
And this year the benchmarks were changed - but related to the 2002/03 year. They involved students who were already in the system.
So universities were publicly "shamed" for not having met targets which did not exist at the time they admitted the students.
The benchmarks - officially, not "targets" or "quotas" - are devised by the Higher Education Statistics Agency (Hesa).
It uses complex formulas, based on policy set by a "performance indicators steering group" including representatives of the funding councils, universities and Department for Education and Skills.
In very simplified form, the benchmark for the statistics on "participation of under-represented groups in higher education (young full-time first degree entrants)" for a given year works like this:
Every student entering higher education is defined as doing a degree in one of 13 broad subject categories. Each subject a given university offers is also allocated to one of these.
Students' previous attainment is put into one of 22 entry qualification groups, according to the number of points they achieved using the "tariff" system devised by the Universities and Colleges Admissions Service (Ucas).
Then for each subject at a university, the number of students in each entry qualification is added up.
The number of "eligible" students nationally - those they might potentially admit - is identified from the subject and entry qualification groups.
There is also a "location adjustment" reflecting the way a proportion of students choose to study at a university near their home.
What was changed this year was the use of the Ucas tariff. Previously the basis had been the A-level points score.
The University of Cambridge is one of those which has condemned the system.
In its case, the benchmark for participation from state schools or colleges rose from 65% to 75%. It achieved 58% - up from 55%.
But a Cambridge spokesperson said the benchmark took no account of how it actually operated as a university, and rated quantity over quality.
The standard Cambridge offer is three A grades at A-level - equivalent to 360 Ucas points. But a student with 360 Ucas points might have got AAA or, say, BBCC, including general studies.
"The new benchmarks assume both students are equally well-qualified for Cambridge entry when, clearly, that's not the case," she said.
"This means that our benchmark has shot up to a figure that we don't believe is either realistic or possible to achieve.
"The previous system wasn't perfect but it was a much better yardstick than total Ucas tariffs."
Nationally, 75% of students who achieved 360 points or more attended state schools - but only 62% of those who gained three As did so.
So Cambridge argues that 62% would be a more realistic benchmark, and that to attain 75% it would have to actively discriminate against many students from independent schools.
It also has a problem with the subject categories.
To study engineering, for instance, it wants students who have done maths and physics A-levels. Many universities ask only for the maths.
So in the Hesa subject category for engineering and technology there were students with enough tariff points to appear to have been eligible for Cambridge, who in fact did not meet its entrance criteria.
A spokesperson for the Higher Education Funding Council for England said the move to a tariff-based system came about because that was the way in which Ucas now published its student statistics.
It would be reviewed, as the whole process always was. If there were concerns the steering group would take account of them.