Oh yes he did oh no he didn’t. The statistical discussion about football statistics

By Tony Attwood

If you have been with Untold Arsenal for a while, you will know that in my writing I rather like the notion of proof, evidence, logic and data, instead of simple assertion and rumour.

Transfer stories are of course the home of rumour, and the stupidity of them is that on average each summer only 3% of the tales are shown to be correct before the window “slams” shut, presumably with the tinkle of broken glass.

We expect nonsense from that quarter, because there is so much of it around, but when it comes to any other story, some supporting evidence is helpful.

Such evidence never proves anything 100%, but evidence, along with some basic logic and a look at the coherence of the tale being offered, usually gives us a clue as to whether the piece is mindless ravings of a desperate columnist who has spent too long in the company of the beer, or a well thought through and evidenced informative commentary.

So for example when the Daily Mail proclaimed in headline recently that sorting out the weak Arsenal backbone was an absolute priority for Arteta, one could look at the number of goals conceded, and then compare it with other seasons in which Arsenal did rather well.  Their logic suggested that we should find a much worse defence now, than we did in the past.   But in fact we didn’t.  So the piece is mindless bunkum.  Best use it to light the BBQ.

We saw this with the story from the Sun: that Arsenal spent £24m on a one year David Luiz contract.  There was no evidence, and to believe the story we had to believe that instead of paying the normal agency fee of 5% of the transfer cost, Arsenal willingly paid 65% of the transfer cost to the agent – for reasons that most certainly would not become clear at this point.

The fact that none of the “outlets” that ran the story even noticed that a key element of the story was missing, and so did not investigate why Arsenal had acted in a way that no club had ever (as far as I could tell) acted before, suggested someone had made it up, and everyone else had merely copied it.

The hope has to be that we might see an increase in the use of proper evidence in football discussions, but unfortunately the reverse seems to be happening – as revealed by The Athletic’s headline “Stats aren’t the definitive viewpoint in football”

Of course in a way that is true.  Statistics tell us a lot, but not everything.  The statistic that bald men commit 35% of the murders in England (which I just made up) is a case in point.  For example if 20% of men aged 40 to 60 are bald, but 70% of male murderers in that age group are bald, then we are on the road to thinking bald men in that age group are more likely to murder than men with hair on their heads.

Although that still doesn’t mean that the baldness causes men to murder.  Which brings us to the issue of supportive evidence.

Turning away from murder, Untold came up with the statistics that the six teams that had spent the most in the transfer market last summer were all doing worse in terms of the league table this season than last.

That doesn’t prove that buying players does not work as a means of improving the team, but it gives an indication that maybe it doesn’t work – at least not immediately.  What might be interesting would be to look at, for example, three summers’ spending, and then note where the club is in the fourth season.

Here’s another one.  Does changing the manager help?  The evidence is not consistent, but certainly the evidence I have seen suggests that one is more likely to see a decline after changing managers than an improvement.  Not always of course, but mostly.

Now putting together a whole range of data like this one can begin to build up a picture of how to improve a team’s performance in the league, and from what I have done along these lines (and I’ll admit my data is far from complete) I’d say the conclusions would be:

  • Gentle changes to a team work better than wholesale replacements of players.
  • If a quality player is going to make a big difference to a team it will be in the second season.
  • Changing the manager, more often than not results in a decline in performance before an improvement.

My figures are far from complete to prove these points so I’m not going to present them, but just for the sake of the debate, imagine that these points are true.  What is noticeable is that all three points are counter to the demands of virtually all of the media and most supporters.

What the media promote are these rules

1: It is necessary to buy new players now.

Usually this is done with no regard to the profit or loss the club is making, whether to player is available and wants to move, whether the funds are available, and how the play can fit into the 25 and home grown rules, and FFP regulations.  It is just a case of “show you are serious” and “just do it”.

2: The player will make an immediate difference.

Some do, of course.  But not all.  Bergkamp, Henry and Pires are three I immediately think of who took a year to settle in.  Fortunately Arsenal had a manager at the time who knew quality when he saw it.

3: Managerial changes will result in improvement because no one could be as bad as the current manager.

This “no one could be as bad as” comes from the way that pundits and supporters present their arguments saying that playing x is self-evidently useless, and as the manager can’t see that, the manager is an idiot.

The fact is however most clubs have signed managers who have not got things right – at least straight away.

And this is the simple reason why there is so much discontent within football.  The three demands of many fans run counter to the reality.   Consider the results of Ferguson at Man U

Ferguson was appointed manager of Manchester United in November 1986.  Here is the record of Man U before and after.  The year he took over is marked in red below

Season Lge pos FA Cup Lge Cup Europe
1981–82 3rd R3 R2
1982–83 3rd Winners Runners-up UEFA Cup – R1
1983–84 4th R3 R4 Cup Winners’ Cup – SF
1984–85 4th Winners R3 UEFA Cup – QF
1985–86 4th R5 R4
1986–87 11th R4 R3
1987–88 2nd R5 R5
1988–89 11th R6 R3
1989–90 13th Winners R3
1990–91 6th R5 Runners-up Cup Winners’ Cup – Winners
1991-92 2nd R4 Winners Cup Winners’ Cup R2
1992–93 1st R5 R3 UEFA Cup – R1

In Ferguson’s first year the club sank to 11th, and by 1990 the club was ready to sack him.  It was just an FA Cup win that kept him in his job.  But it took him seven years to win the league – seasons which included two finishes in 11th and one in 13th.  True he won the FA Cup and the Cup Winners Cup – both significant trophies, but that was two in six years.  No regular top four finishes, or record number of cup wins.

But as we know, in terms of eventual success he did well.   And yet Emery was hounded out of Arsenal, with a record far better than Ferguson.

And we still hear the old cry, “you can prove anything with statistics.”  It is of course nonsense.  You can prove anything with fake stats and the wrong stats, but apply statistics properly and they most certainly will tell you what is going on.

This is indeed why we know something is seriously wrong with Premier League football.  We now know how much referees are influenced by the home crowd because we now have the statistics.  We know how unable commentators are to see this – or how disinterested they are.

The problem we have is that those who commentators on football, really don’t understand the statistics that are on offer.  They are far too interested in the personalities.



One Reply to “Oh yes he did oh no he didn’t. The statistical discussion about football statistics”

  1. Tony

    “That doesn’t prove that buying players does not work as a means of improving the team, but it gives an indication that maybe it doesn’t work – at least not immediately. What might be interesting would be to look at, for example, three summers’ spending, and then note where the club is in the fourth season.”

    Which is basically, though not exactly what I kept saying whenever you said, and to paraphrase you here, ‘Spending big doesn’t work’.

    That statement above is a change from your usual stance on this topic over recent times, because you have stated on many occasions that spending doesn’t work…….

    This from the article linked bellow

    “Here, the story that we have presented before is repeating itself. There is no direct link between league position and expenditure and thus this table has a multiplicity of issues within it. Manchester City massively outspending everyone in order to try and stay at the top and failing. Arsenal in third, despite the endless whining and whinging by AST and its associated supporters’ groups who demand more and more spending.

    Now in Europe, the top spenders. Again the Manchester two at the top. But look at this figure for Brighton and Wolverhampton either side of Real Madrid. Aston Villa we may note being above Bayern Munich”.

    You said:


    ……which is clearly untrue, as I pointed out in a couple of rather long comments back in May, that actually showed how spending big, not only works, but is essential.

    I made 2 comments in the following article:


    The first showing just how dominant high long term net spenders are.

    The second showing just how misleading trying to correlate one seasons high spending with trophies, or one seasons low spending with a lack of trophies can be, in this case using Chelsea as an example.

    And that is the thing with statistics. The smaller the data sample, or the shorter the term over which the data is sourced, the less reliable the conclusions are.

    You have to look at the premier League over it’s entire duration to get a true picture of just how crucial long term big spending is, which brings me to Arsenal and what I believe they need to do.

    I believe if Arsenal can maintain an average net spend similar to that of the last 5 years for the next 5 years they will have at least a chance of not only breaking back in to the top 4, but possibly even challenging for the title.

    But that still may not be enough, if for example, Man city, Man Utd, Liverpool and lets say Newcastle, spend the same or more.

    If they don’t that will give us a chance.

    If we drop to an average annual net spend of say zero to 10 Million over the next 5 years, irrespective of how good our youngsters are, we will stay where we are and could even drop, depending on what happens at spurs and Everton.

    You see, having these brilliant youngsters is all well and good, but there are 2 major problems.

    1 – Will they all, or at least enough of them, turn into the WORLD CLASS players we will need to compete with these bank rolled clubs?

    2 – Will they stay with us if they do, when these bank rolled clubs will be able to pay them more money, and pretty much guarantee them Champions league football and trophies?

    These are the cold hard facts as I see them. To have any chance of recapturing past glories:

    -We need these youngsters to be everything we hope they’ll be.

    -We need to keep our current World Class players.

    -We need to attract World Class players on a regular basis to keep current players on their toes and freshen the squad.

    Can we do it? Only time will tell.

Comments are closed.