I have now collated three full SPFL seasons of assessments of key decisions affecting the top two teams in Scotland.
114 matches, 10,260 hours of football.
Again, I would like to thank the Yorkshire Whistler who provides such in-depth, patient, well thought out and argued rationale for the decisions reviewed. His season ended well with Sheffield Wednesday staying in the English Championship and looking forward to the derbies again next season. Mon the Blades!
What are the patterns and trends in this time?
Error Rates
Firstly, are the refereeing performances improving given the first full implementation of VAR?
The number of incidents reviewed has increased from 107 to 110 and then 124 last season. This is not surprising as the use of VAR has highlighted more potential infringements.
The good news is that the error rate has decreased. Using their “split second” judgement only in 21-22 the part-time Scottish refs were estimated to have erred an astonishing 32 percent of the time. Remember, unlike SFA staff (an “independent panel”) marking their own homework, this is the view of a truly independent and expert professional.
Whomever the decisions are benefitting, that is an extraordinarily high error rate and supports the “they aren’t very good” argument. This does not preclude other reasons for errors, and quite the opposite, it rather opens the door.
What also appears to be the case is that adoption of VAR is helping to reduce the number of errors. The error rate last season of 17 percent is almost half that with unaided refereeing from 2021-22.
A case in point is the derby matches.
Without VAR, The Rangers would have had a goal incorrectly awarded in the first derby by Kemar Roofe following a foul on Gustaf Lagerbielke.
In the third derby, The Rangers would have wrongly had a goal allowed following a foul under John Beaton’s nose on Tomoki Iwata in the immediate build-up.
In the final derby, John Lundstram would not have been red-carded for his shin-high challenge on Alistair Johnston.
To be fair, all countries implementing VAR have had seasons of teething problems. Scotland will be no different. The refereeing is far from perfect, how can it be given the personnel and SFA mechanisms, but there are signs of improvement.
I should add for clarification that the selection of incidents for review is subject to some constraints. I strive to include all calls that exercise both sides of the Glasgow divide based on the press coverage and social media reaction. The limitation is usually the availability of video evidence. It is not uncommon for both the BBC Sportscene and the SPFL Youtube channel to miss out on key incidents from their highlight packages.
I have previously asked a couple of online accounts belonging to supporters of The Rangers to provide any evidence of incidents they feel need raising, but no joy. That invitation remains open to any good-faith fan.
In the main, the lack of video evidence means incidents affecting both clubs are sometimes missed.
Patterns of Assistance?
Taking three seasons' worth of reviewed incidents into account, are there patterns of assistance?
Over the three seasons, The Rangers have been estimated to benefit by 5.24 expected points. Celtic has been estimated to have been penalised to the tune of -8.15 expected points.
That is a cumulative pattern of benefit to The Rangers of 13.39 expected points.
This is significant in terms of a “tight” title race. An average of 4.46 xPts difference in The Rangers favour when the actual winning title differences have been four, seven and eight points respectively could have had a significant impact.
As it played out, Celtic should probably have won each league earlier given an equitable distribution of refereeing errors. So at least Sky TV and the SPFL are happy that the title tension was prolonged!
“Conspiracy”
Simultaneously, these are not differences that indicate a cohesive and predetermined effort to cheat!
Whenever one points out data discrepancies in refereeing outcomes, those predetermined to resist that message scream “conspiracy theorist”.
There may be consistent patterns of assistance for one club without there being an explicit conspiracy.
But, when we look at other refereeing outcomes through the data, they also point to a consistent pattern of assistance for one club. The various strands of analysis are fully viewable on this site.
However, I doubt this is labyrinthine nor complicated.
Indeed, it is rather all in plain sight.
The SFA is a secretive, uncommunicative, amateurish organisation with a history of hostility to Celtic (all a matter of public record from the displaying of the Irish tricolour at Celtic Park; through Jim Farry; to the collusion with Rangers over the EBT scandal and UEFA licenses; the Hugh Dallas scandal; Dougie McDonald and last season, David Dickinson).
SFA referees are part-time and are so highly rated they cannot get an appointment at the upcoming Euro 2024 championships nor the Paris Olympic football event. The SFA refereeing service provision is behind the SFA wall from start to end covering everything from recruitment to assignment, performance evaluation and assessment. It is a literal (man in) black box. It isn't a system set up to produce the best people.
Given the SFA’s refusal to even discuss refereeing allegiances, this further puts doubt in people’s minds and where doubt exists, so trust is eroded.
We are told by the supine local media that Scottish referees, above all others in society, are beyond suspicion, And yet one high-profile ref had well-documented gambling issues (not a risk flag, oh no) whilst another current employee seems to have been pilfering the public purse for private expenses. That same media fails to hold power to account.
I rather suspect given the well-documented culture at the SFA, and the default defensive nature of the organisation, that employees simply revert to expected behaviours. Furthermore, they perform in a way that would best guarantee advancement and a chance to suckle at the teat of the UEFA competition expenses train. Don’t most employees do similar?
A more sociologically minded academic study would best be undertaken to understand why patterns of assistance in refereeing across multiple data points, exist.
What is clear is that VAR is helping here and avoiding some of the more egregious decisions to manifest. Hate it all you like, it does seem to be leading to the right decision more often, or at least reducing the extraordinarily high error rate.