By Gary Eason
BBC News education correspondent
The answer to the question, "Which is the best school in the country?" is, more than ever, "On what measure?"
There were 14,759 mainstream primary schools in England in the league tables this past year, but results are published for only 13,532 of them.
Most of the others had 10 or fewer pupils sitting the tests - and the Department for Children, Schools and families (DCSF), which issues the figures, says publication would risk an individual pupil's results being identified.
The conventional benchmark is the aggregate score out of 300 of the proportions of pupils attaining at least Level 4 of the national curriculum in their Key Stage 2 tests in English, maths and science.
Of the 13,024 schools which had valid published results this year and last, 6,084 (47%) had a better performance on all three test subjects and 308 (2%) did the same.
But 6,632 (51%) did worse.
A total of 268 schools had the maximum possible 300 - compared with 329 last year and 255 the year before.
Using, as a tie-break, the average point score (APS) - which reflects attainment right across the ability spectrum - the top school was Hampstead Norreys Church of England Primary School in West Berkshire.
It had the highest APS of any: 32.9.
Eight schools this year had the record of 300 in each of the four years' worth of results published in the tables.
The lowest Level 4 aggregate score was 0 (zero) - in somewhat unusual circumstances.
It was recorded at one of the two dozen or so academies which cater for primary age children, the Steiner Academy in Hereford.
The academy promotes the Rudolf Steiner method of learning tailored to the individual's need and their personal level of progression, and teaches its own curriculum.
Being state funded, it was required to register the children for the Sats.
But parents had a different idea, according to the principal, Trevor Mepham.
"The parents were involved in a campaign against the children sitting the tests," he said.
"We registered the children and the parents in large numbers decided that they would not have their children sit the tests."
The Department for Children, Schools and Families said that in science, all pupils were deemed to be "working below the level of the test" so were not required to sit it.
In English and maths, five were in the same position - and the other 19 were absent on the test dates.
The turn of events might be regarded as an embarrassment for ministers who promote Academies as the answer to under-performance in England's schools.
But the Conservatives also are keen to see more Academies - and to have more parents' groups setting up schools and controlling their destinies.
English and maths
The next worst performing school was St Mary's CofE (VA) Primary School in Worcestershire, with a combined score of 50 out of 300.
There, only 7% managed the expected level in both English and maths - which is becoming the new benchmark.
This is the last year that children will be required to sit the science tests, while political mileage is being made out of children's scores not just in English or maths but both English and maths.
Attaining Level 4 in these subjects is deemed to be crucial to thriving in secondary school.
In the top aggregate performers all the children do so - because obviously if you have 100% getting Level 4 all three subjects you have them doing so in any two.
But beyond that, only 14 schools managed 100% in both maths and English (but not in science).
There are different ways of trying to establish how much progress children have made while at a school, rather than their absolute attainment.
A new Progress Measure has been introduced this year - and English and maths feature again.
The government wants to see children making at least two national curriculum levels of progress during each "key stage".
So in primary schools, from Key Stage 1 to Key Stage 2, they should move two levels - typically, at that age, from Level 2 to Level 4.
To focus schools' minds the DCSF is, as usual, using the league tables and this year has incorporated two new figures: one showing the proportion of pupils in each school making at least two levels of progress in English, and one for maths.
All pupils attaining Level 5, which is the highest that is assessed, are treated as having made expected progress.
On the basis of the published results, there were 882 schools with 100% on the English measure, 825 with 100% in maths - and 237 with 100% achieving both.
But many schools have no results - because the analysis falters on the realities of school life: children come and go, and some arrive from outside the country.
It is not always possible to know what their prior attainment was, so it is not possible to calculate how far they have travelled academically.
Where the proportion of pupils used in these calculations drops below 50%, no measure is reported.
But even where it is reported, this means that one school's 100% may not be the same as another's: it might be that 100% of all the pupils made expected progress - or it might be that 100% of only 51% of them did so.
There is still the contextual value added (CVA) improvement measure, worked out for each pupil by comparing their Key Stage 2 performance with the middle performance of other pupils with similar prior attainment at Key Stage 1.
The arithmetic mean of these individual scores gives a score for the whole school. This is converted to a number based around 100.
Resulting rankings can be misleading because although the scores appear very precise they are bracketed within "confidence intervals", largely determined by the number of pupils in a school - a range of scores, any of which might the "correct" one.
Two schools with apparently different scores but overlapping confidence intervals might have the same CVA - indeed, the one with an apparently lower CVA might have the higher true score.
And as with the progress measure, if the proportion of pupils included in the calculation drops below 50% no CVA is reported.
So CVA scores need treating with more than a pinch of salt - but schools with scores at either end of the range are clearly of different quality.
The school recording the greatest pupil improvement as measured by its CVA was Blue Bell Hill Primary and Nursery School in Nottingham, on 105.
The one adding the least value was Gallions Primary School in Newham, London, on 94.7.
At local authority level, the best area this year was Richmond upon Thames where the average Level 4 aggregate was 273 and the worst was Hackney in London with an average of 226. The national average was 247.
These two areas have been in these positions for seven years running.