I appeared on Graham Spiers’ Press Box podcast this past week to discuss “Do Scottish refs favour Rangers”. If podcasts had open goals…………whits the goaly daen’?
Anyway, I was given the space to run through at a very high level some results from various pieces of analysis by myself and others, which, when taken together, point to evidence of the potential pattern of assistance for one club. Since the pod, Graham has given a supporter of The Rangers space to put up what has been styled a “rebuttal” and then free reign to run amok in the comments.
The “rebuttal” article is free to air while the pod remains behind a paywall. I have asked Graham twice now to either put both free-to-air or both behind a paywall. He is “thinking about it”—a weighty decision balancing the rights of the people who have already paid versus balance and fairness. Graham is just the man for such a moral conundrum!
Why have I not responded, and why am I not engaging in the debate raging over on the Patreon? I will cover that in a moment, but first, a comment on the podcast.
I do have a regret. On it, I allowed myself to get drawn into a debate on “why” all this was happening. Unlike the investigative data, I wasn’t particularly prepped for that, and I felt it diluted the message somewhat.
It allowed the likes of Matthew Lindsay to deploy lazy tropes and to try to “Old Firm” the debate. This is about the Scottish Football Association, NOT The Rangers, although it is touching some of their fans feel the need to defend the governing body so vociferously. I also made at least one mistake – for example, Willie Collum went six months without refereeing The Rangers, not 18, as I stated. I’m happy to correct that, and that’s all on me. Mind you, Lindsay claimed The Ranger’s current jeopardy-free run was only home games, so we all make mistakes.
For me, the far more serious aspect of this is the analysis. Both Graham and the respondent he published fall into the same trap. This was well expressed by my dear friend @jucojames earlier in the week. He pointed out that the “kerfuffle” (it isn’t a debate): “is a good case study on the importance of distinguishing between “stats/data” and having a robust analytical framework grounded within a good faith attempt at objectivity.”
My chronically balanced trans-Atlantic buddy nailed it.
I have tried to explain the method and rationale, be consistent with timeframes and scope, and explain the decision made. I will document this.
All of those things are up for good faith debate, and I deserve to be challenged on them. There will be imperfections– as James and I often nerdily joke, “all models are crap; some are just slightly more useful than others”. When a bad faith actor responds with lots of unstructured “stats/data”, shifting scope, timelines and selectively segmenting that isn’t an analytical framework, it’s just throwing numbers at a wall and hoping some land.
This is about holding power (in this case, the SFA) to account. It isn’t about The Rangers. All football fans in Scotland should be interested in whether there are issues with officiating as it probably impacts their team, too.
I am a Celtic fan, and I am biased as regards how I see on-field decisions.
I will commit to being transparent about the data used, the scope, the timeframes, the safeguards, the assumptions, the method, the rationale and the reasoning behind the conclusions. Any subsequent debate has to be in good faith AND utilise good practise.
My objective in all of this is to establish whether there is a problem with Scottish refereeing. I will publish all this in a time and place of my choosing.
Also, I will stick to the data analysis in the future and let the sociologists and anthropologists get into the whys and wherefores of unpacking any potential causes of a pattern of assistance.
Allaboutceltic says
Numbers, like Google, if you play around with them long enough, you will eventually come to a conclusion which suits your narrative, purpose & argument. To solve the problem, everyone must acknowledge & agree to the input & output and the factual data will bring you to the true conclusion. From what I understand, the point of this debate was to find a reason & conclusion as to why *theRangers went over 2 years without conceding a penalty. That is a “fact” and no matter how you dress the numbers, you will always come to the same conclusion.
When you have these type of discussions, and you’re trying to come to a solution which is “fact”, then it’s an almost impossible task if the problem solving team hid a natural bias. This is not to say they shouldn’t be present, but before these discussions start, there should be a solid foundation to which everyone agrees on the route to take and if they get that part right, then the data will lead them to the truth…and that’s a “fact”.
HAIL HAIL
KEEP THE FAITH??
PS,
In my opinion, Celtic by Numbers is a website operated with the highest integrity and provides the data and evidence to which they draw their conclusions and reports. I know of no other such site within the UK who does this on a regular basis and for that, they should be congratulated. Keep up the fantastic work??
celticbynumbers@btinternet.com says
Thanks for the kind words
Actually i had minimal interest in the penalty’s against run. it is one data point and long runs without are not uncommon albeit this particular run IS remarkable
i am more interested in discerning trends in the range of refereeing decisions over a longer period and this will be presented in time
Brucebhoy says
I posted on kerrydalestreet.co.uk… I started reading this clown’s nonsense. The first thing I noticed was that far from keeping his own data and records, he was taking his numbers from Transfermarkt. Which is fine. If you understand that they record regular season and post split data separately. So he said we got 7 penalties last season when we got 8.
He also got the number of penalties since Sevco came into the top tier wrong. Conveniently, he understated the number of penalties Sevco got and overstated the number Celtic got. That was the first two things I looked at and saw 3 big errors. I stopped but his data is clearly not to be trusted, never mind his analysis of it.
Doug says
I made the mistake of taking his data at good faith, then the mistake of assuming he could hold a balanced view, then the mistake of carrying on reacting to his illogical arguments. I’m starting to think his account is there to try and deflate any argument on a rational look at bias in Scotland.
celticbynumbers@btinternet.com says
I was advised similarly Doug and sorry you had to get down and dirty there!
Auldheid says
Res12 which was also about the SFA and reforming it was also portrayed as an OF issue in order to question the motivation behind the facts presented.
It is a standard tactic when SFA are being questioned.
Keep on being factual.
Damian says
That’s not fair play from Spiers. Disappointed in him.
As a Celtic fan who’s disagreed with you on elements of this line of argument, I’m terribly sorry to hear you’ve been facing this kind of bad faith nonsense.
Please keep at it.
On your central focus, I completely agree with you, by the way. Whether or not there is a pattern of active, meaningful bias is actually neither here nor there. The opaque and antiquated nature of the SFA leaves it wide open to people thinking that way.
Charlie Kelly says
Something I always like to look at is the issue of “Sample size”
I believe we are now in to your third season of the “honest mistakes” series. So by seasons end we’ll have gone past 100 games. I think after 100 games we can start to draw some conclusions because a lot of statistical blips will start to even out and all you’ll be left with are Bias/Corruption/Incompetence… whatever you want to call it.
If we take something simple like a coin-toss as an example.
The methodology is simple – We flip a coin
The potential outcomes are limited – There’s only 2
There are no other variables – We flip the same coin. The same person flips it each time…
If we do that say 10,000 times then it’s going to come out as 5,000/5,000 give or take a few. we all know this.
However, within that 10,000 there will be runs where it is 10 heads in a row or the last 25 flips have only produced 3 tails etc… There will be lots of weird runs and statistical improbabilities contained within 10,000 coin-flips but when all is said & done it will balance out at 50/50 heads/tails over such a large sample size.
So with that in mind, My question for Alan to consider is – Is the sample size big enough so far in order to be able to draw conclusions. Or is rangers run of not conceding penalties just a statistical aberration, like getting ‘Tails’ 7 times in a row (less than 1% chance) ?
Justso says
With respect your coin flipping example is comparing apples and pears. Coin flipping is random in so much as the person doing the flipping cannot accurately affect the outcome. They cannot apply a force to the coin that will guarantee an outcome. So the “flipper” has no measurable effect on the outcome. In contrast footballing decisions such as penalties are completely within the gift of the officials both on the field and in the VAR area. So a run of 7 tails cannot be compared with any run of no pens in any way. The pens are a human decision in which Alan is discerning a pattern. Heads and tails are random events.
Justso.
celticbynumbers@btinternet.com says
The coin flip analogy like all analogies, ultimately fails – it was flippin’flippant!
The key is to see over multiple data points over multiple seasons if there is a pattern of assistance
celticbynumbers@btinternet.com says
Hi Charlie
it is a fair challenge re sample size
What i have and will publish in due course are a range of data points
If there is the same pattern across multiple models and data points then perhaps we can be more confident in suggesting there is a pattern of assistance that may need looking into
Loudpiano says
Surely the answer is simple. Spiers provides a link to your original article explaining the model?
celticbynumbers@btinternet.com says
He has put both behind paywall now after allowing his site to descend into a sewar by all accounts. Hopefully a learning.