Tag Archives: Premier League

Southampton v Hull City. Premier League: Saturday 29th April 2017

2 May


For my money there is no better place for getting a view of a football clubs season than from the barbers chair. The barbers own observations will have been triangulated by a huge number of other perspectives and from over the season and their arguments will well honed through countless conversations on the topic. I’d add that perhaps one of the best barbers in this regard is to be found just over the Itchen bridge in Woolston at Sean’s barbers.

The barbers view is particularly valuable as Saints fans struggle to decide whether this season can be considered as success, or a disappointment. Sean’s view was that the club needs to spend more to compete with the big clubs and that in the absence of this the current position is where the club should expect to be. Puel has, Sean also feels managed players well, particularly Maya Yoshida, who has gone from an error-prone bit-part player to an important team member and occasional captain. Crucially though Puel has provided opportunity to young talent to blossom.

Sean’s points are well considered and I’m in agreement. In the context of the last few seasons perhaps to be in 9th place is a disappointment, but in the wider context of the last decade and given the resources the club has then it’s actually quite good – and that’s not even considering the League Cup final.

One impact of sitting in 9th place with just a few games to go is that the Saints went into the game with Hull with relatively little to play for, neither competing for Europe, nor threatened by relegation. Hull on the other hand were fighting for survival, being just one place and two points above the final relegation spot.

It was therefore unsurprising that Hull seemed the keener side out of the blocks, striking the outside of the post from a free kick early on. They were also as one Saints fan later put it “first in the tackle” throughout, but over the course of the game opportunities for either side were few and far between – the best Saints chance falling to Gabbiadini who blazed wide after being put clean through on goal.

All in all it was a rather poor spectacle with little to excite the Saints fans who were the most subdued I’ve ever seen them at St. Mary’s. They were though finally half-roused when, after a spell of pressure, Maya Yoshida won a penalty in the 89th minute.

There is a scene at the end of the Hitchikers Guide to the galaxy where Marvin the android, on the verge of death, sees God’s final message. The message reads “We apologise for the inconvenience” at which the perennially downbeat Marvin appears to take some consolation “I think… I think I feel good about that.”

If scored the Saints fans would dissipate through the streets surrounding St. Marys with a skip in their stride, the 89 minutes of painfully-dull football they’d witnessed swiftly forgotten.

It was however, the Hull fans who embarked on their long journey home happy as Dusan Tadic saw his well-hit penalty – low into the right corner – spectacularly saved by the Hull ‘keeper Eldin Jakupovic.

This secured a valuable point for the visitors which may, before long, prove crucial. For the Saints fans it simply adds further murkiness to the debate on whether the season can be called a success. Perhaps it’s worth a trip to the barbers.


The Premier League and Championship: Patterns of on-field inequality

2 Jan

PL and Champ CV2

The coefficient of variation is the standard deviation as a percentage of the mean average. This is effectively a measure of how spread out the number of wins is. A low percentage means that the number of wins for each club is more clumped around the average while a high percentage means that the number of wins recorded that season by each club is more spread out from the overall average – therefore implying that the division is much more uneven.

The technical bits out the way the graph shows an interesting pattern. Viewed over a time period of 1983-84 until the present the trend in the Championship appears to mirror that of the Premier League, but with a slight delay; For instance in 2010-11 to 2011-12 the Premier League saw a sharp rise in the coefficient from 33.7 to 43.8. A similarly sharp rise was then seen in the Championship between 2012-13 to 2013-14 29.6% to 34.5%

A gap appeared to open up between the two, when in 1993-93 the Premier Leagues coefficient rose to 38.1% from 23.9% of the previous season while at the same time the Championships declined from 31.4% to 23.1%

This gap was then closed in the late 90s thanks to the Championships coefficient increasing, notably between 1995-96 and 1998-99 when it rose from 20.2% to 35.6%

In the late nineties and early noughties, between 1997-98 and 2000-01 the coefficients of the two divisions were roughly in line however, from this point on the two have diverged with the Premier Leagues coefficient continuing on an upwards trajectory, while the Championships has taken a downward trajectory, where it even reached a low 19.6% in the 2012-13 season. Despite these different trajectories however, the trends seem to still be mirrored with the lines following a similar pattern, albeit on their differing trajectories.

The start of the Premier League era was associated with the top-division becoming more unequal in terms of wins and this trend fed into the Championship a few seasons later in a kind of trickle-down effect. However, over the last decade, the broad trend has been for the Championship to become more even on the field with teams recording a number of wins closer to the overall average for the division.

In terms of explanations I’m stumped. In my last post I looked at the long term trends for the top-flight going way back and my explanation there is that the general trend towards less on-field equality is driven by the reduction in the role of luck which has been an underlying feature from the development of goal-nets to the implementation of goal-line technology.

If anyone has any other hypothesis, or explanations then please feel free to add your thoughts! The data I have used is in the table below.


Has the Premier League made football more unequal?

12 Dec

One of the complaints most commonly levelled at football is that since the advent of the Premier-league there has been a growing disparity between the haves and have not’s. The league has become too uneven: It’s a debate which almost seems to mirror anxiety about growing inequality in wider-society – something clubs with millionare players and near minimum-wage caterers are also part of.

Leaving some of these other issues behind the first challenge is how to measure on-pitch inequality. Football – to the analysts delight -provides a wealth of measures on almost any subject, and the same is true here. We could look solely at honours and titles, points gained, goals scored, or conceded. We could also look at the divide between the top 10 per cent and the rest, or the top 25 per cent, or 50 per cent and so on.

The relative closeness of teams, in a sporting sense, can be measured quite simply by looking at the number of wins each club has gained over the course of a season. This can then be computed into an average number of wins, from which we can calculate the standard deviation, a measure of how spread out clubs are from the average. The assumption is that clubs in more even leagues will tend to have a number of wins closer to the overall average – in a very even league, for instance, clubs should have around the same number of wins

Going one step further, we can work out the coefficient of variation. This is the standard deviation as the proportion of the average which allows us to better compare leagues across time (this is because a season with a high number of wins overall is likely to increase the standard deviation and what we are interested in is the relative share of wins). In effect a higher percentage means that the standard deviation is high relative to the mean and therefore the number of wins recorded by clubs in that league is more spread out and the league uneven.

CV graph

Doing this results in an interesting chart. The beginning of the Premier-League era does result in a sharp increase in the coefficient from 25.9% in 1991-92 to 38.1% in 1993-94 so it would appear that yes, the Premier League has resulted in a league which is less competitive overall, however, this needs to be looked at against the historical trend. If we look back to the 1930s we can see that despite fluctuations that there has been a general underlying trend for the league to become less equal over the whole period.

With the premier league not taking place until 1992 What then, has driven this trend?

Looking back to the 1930s even with what was on paper the best team. Inclement weather conditions, poor pitches, long journey times, poor officiating and injuries could all conspire to affect the final result. With no substitutions one-knock could reduce a team, if not in actual numbers, then in effect to 10 men. Fortune, or rather misfortune, could therefore play a much bigger role in deciding the outcome of games.

It was in 1965-66 the league first allowed one substitute, for cases of injury. The decade also saw the abolition of the maximum-wage for footballers in 1961. This was accompanied by some of the biggest increases in the coefficient, from 23.2% in 1961-62 to 41.0% by the end of the decade in 1969-70

Undoubtedly the premier-league marked a big change, or rather the confluence of several big changes. Following the Taylor report clubs began, to invest heavily in infrastructure – upgrading grounds which had been left to decline for decades. Such projects have a time-frame, and numerous complications; financing, land-availability, planning-permissions and so on. This meant that over the 90s and 00s some clubs lagged behind, stuck in outdated facilities which hampered their ability to generate revenue.

European-level changes to improve the mobility of labour also had a big impact. Clubs had access to a much wider pool of talent with the wealthiest clubs able to purchase not just the top players in England, Wales and Scotland, but the greatest players in the world. Allied with 1995’s Bosman ruling this sparked a huge change.

Annotated CV graph

All this though simply marked the acceleration of a trend begun much earlier.

The ultimate question though, is have we seen the end of this trend, has the Premier League reached an apex of inequality?

It is hard to infer the future from just looking at the past however, part of me thinks that we may have. The reason for this is that we have reached a point where the clubs like Manchester United, who raced to the top in the 90s have slowly been reeled in by the chasing pack as clubs lower down the food-chain have completed infrastructure improvements and have gained experience in running a more commercial operation.

It is also some years since clubs have had access to, and the cash for, the world’s best players – The impact of this has already been felt and now, thanks to the Premier League’s TV deals even modest clubs have a huge amount of spending power.

Are Goalkeepers getting taller?

30 Aug

This post began with a question posed by Andrew of the excellent Hopping Around Hampshire blog: Are football players getting taller? I promised to look into this. 

For a while pondered my approach. First, I’d need a data source. I tried some old league directories, but these didn’t have player height information. I then searched my old football box for my copy of the Official PFA Footballer’s Factfile 1998-99, but was hit by a niggling thought that It had been the one book I’d got rid of. I then tried the internet, buy though I managed to find data on current players at http://www.premierleague.com it only had heights going back a few years so back to the old football box. This time I managed to find a Merlin Premier League 97 album (minus stickers) which had what I was looking for, although not for every single player – just the ones deemed worthy of a sticker.

To begin with I decided to compare goalkeepers as height is regarded as a key attribute for Goalkeepers, providing them with an added advantage. I’ve also heard it said that at some point there was a decided shift away from shorter ‘keepers – like Leicester City’s former ‘keeper Kevin Poole 1.78m. As the Merlin Premier League 97 data was incomplete I decided to compare only the preferred ‘keepers in each team with their Premier League counterparts today.

The headline finding is that, on average, Goalkeepers today are 2cm taller than their 1997 counterparts at 1.91m compared to 1.89m. Today’s tallest Premier League Goalkeeper Southampton’s recent signing Fraser Forster also breaks the 2m barrier, standing at 2.01m. In 1997 the tallest was West Ham’s veteran ‘keeper Ludek Miklosko who stood at 1.96m along with Coventry’s Steve Ogrizovic.In total there are three goalkeepers today taller than Miklosko and Ogrizovic, as well as Forster these are Asmir Begovic at 1.98m and Thibaut Courtois at 1.99m. When it comes to the shortest goalkeepers in 1997 Mark Crossley and Neil Sullivan both stood at 1.83m, the same as today’s shortest goalkeepers Hull’s Alan McGregor and Leicester City’s Kasper Schmeichel.

Goalkeeper Infographic

The list:


Ludek Miklosko 1.96
Steve Ogrizovic  1.96
David Seaman 1.93
Russel Hoult 1.93
David James 1.93
Peter Schmeichel 1.93
Dave Beasant 1.93
Tim Flowers 1.88
Dmitri Kharine 1.88
Nigel Martyn 1.88
Shaka Hislop 1.88
Tony Coton 1.88
Mark Bosnich 1.86
Neville Southall 1.86
Kasey Keller 1.86
Kevin Pressman 1.86
Ian Walker 1.86
Gary Walsh 1.85
Mark Crossley 1.83
Neil Sullivan 1.83


Fraser Forster 2.01
Thibaut Courtois 1.99
Asmir Begovic 1.98
Joe Hart 1.96
Wojctech Szczesny 1.96
Ben Foster 1.93
Brad Guzan 1.93
Simon Mingolet 1.93
David de Gea 1.92
Adrian 1.90
Lukasz Fabianski 1.90
Tom Heaton 1.88
Hugo Lloris 1.88
Tim Krul 1.88
Vito Mannone 1.88
Robert Green 1.87
Tim Howard 1.87
Julian Speroni 1.86
Alan McGregor 1.83
Kasper Schmichel 1.83

European League Attendances 2013/14

5 Aug

Attendance infographic 4

Hollowing out: The shifting geography of the top-flight

7 Dec

How much has the distribution of teams in the top-flight changed in thirty years? I decided to take this season and compare it to 1983/83 marking each team as a red point on the map. The most striking feature is the hollowing-out effect which has taken place around the mid-part of the country. It could, of course, just be all about football, but to me there’s far more to the pattern.  The Early 1980s were a time when some major changes to the very fabric of the country were just beginning to gather pace.  Over the intervening thirty years the Midlands have been hit hard by the de-industrialisation process, particularly the car industry in the West Midlands along with the mining industry in Nottinghamshire and the steel industry in Sheffield. By contrast London’s globally-connected finance-driven economy has boomed.



Ligue 1; Europe’s most equal football league?

26 Nov

At the moment I’m doing a course on stats and it’s got me working out things like standard deviation by hand. If that seems pointless it is because in the age of Excel it is, but I’ve got to memorise it and practice it so what better way than doing this than to get some attendance data and trying to answer the question – which league is, in terms of attendances, the most equal. Just to start with I’ve stuck to the big-5 i.e the Premier League, Bundesliga, Premier Liga, La Liga and Ligue 1 and used attendance data for the 2012/13 season

First up is a rather simple measure of difference, the range. This is the difference between the smallest and largest observation – in this case the difference between the smallest and largest attendances in each league.

Range Big5

From the graph we can see a big difference between the three most unequal; The Bundesliga, La Liga and the Premier League and the more equal Serie A and  Ligue 1. In the Bundesliga top club Borussia Dortmund an average attendance at 80 520 compared to SpVgg. Gruether Furth on 16 864, a difference of 63 656. In Ligue 1 by contrast top club Paris St Germain had an average of 43 239 compared to AC Ajaccio who had an average attendance figure of 6 801; a difference of 36 438.

The range is though quite limited and can be swayed by extreme values. Another measure is the interquartile range – that is the difference between the upper and lower quartiles (the lower quartile is the point at which 25% of the observations are equal or less than and the upper quartile is the point at which 75%  of observations are equal or less than)  . This has the effect of chopping off the extreme values.


This produces a slightly different graph. Notably Spain drops back. This is undoubtedly due to the effect of ignoring Real and Barca who are outliers in terms of attendances. With them out of the picture Spanish clubs are actually a lot closer together in terms of attendances. Serie A though moves in the other direction. This is because Serie A is an oligopoly with several big clubs Inter, AC Milan, Juve, Napoli, Roma and Lazio all drawing bigger crowds whilst at the other end  a larger number of clubs draw much lower crowds. Again too it is Ligue 1 which emerges as the most equal with an interquartile range of 9 854.

Another measure still is standard deviation (SD). Easy to work out using a spreadsheet, less so by hand! This shows how clumped around the mean (average) the figures are. A high SD means more spread (and therefore more unequal), whilst a low figure means the numbers are more bunched together (and therefore more equal). The difference from the previous two methods is that SD takes into account the whole distribution.


What we get is a graph which looks quite similar to the first graph, showing the range. Once again Ligue 1 emerges as the most equal. There is one problem with SD however, and that is it can be quite hard to compare when the means are so different; For example the Bundesliga has a standard deviation of 17 039, which is similar to the 17 265 of La Liga, but the mean  attendance in the Bundesliga is 46 624, compared to 28 327 for La Liga, so comparatively there is less variation in the Bundesliga.

One way around this is to use a coefficient of variation. This converts SD into a percentage of the mean. If we do this we get the following result.


Strangely the Bundesliga and Premier league now look more equal, the coefficient of variation showing that comparatively there is less variation between attendances than the highly unequal La Liga. Ligue 1 appears more unequal as using the other measures the difference between attendances have appeared lower because the overall numbers are lower for all clubs in the league. Looking at this in proportion shows that attendances are similarly unequal.

Perhaps it is interesting to look at results. The Soccer by Numbers blog looks at the standard deviation of wins for the big five and the Eredivisie for the 2010/11 season and finds that La Liga, the Eredivisie and the Premier League are the most unequal with Ligue 1, Serie A and the Bundesliga being much more equal.

Another stat is to look at the number of different league winners in the ten years between 2003/04 and 2012/13

League Winners

Even taking into account the dominance of Lyon over the earlier part of the period, Ligue 1 seems the most open with six different winners. In fact since 2007/08 a different club has won the league championship each season.

So is Ligue 1 the most equal league of the big five? Well, it depends how you look at it, but on balance it seems that Ligue 1 is the most equal with the smallest absolute differences in attendances between the clubs. The highest number of different league winners also suggests a more open league, certainly when compared with the highly unequal La Liga.

%d bloggers like this: