By John Considine

Earlier this week I watched Seattle Seahawks play the Washington Football Team.  On their last drive of the game, Washington led by 8 and had an "easy" opportunity for a field goal on fourth down.  A field goal would have probably sealed the game.  Instead they opted to seek a touchdown.  They failed.  Seattle went the length of the field on their drive, scored a touchdown, and would have tied the game if their two-point attempt succeeded.  What influenced the Washington decision and did analytics play a role?

Ever since David Romer's paper "Do Firms Maximize?", a paper that examines the fourth down decision of NFL coaches, there seems to be an acceptance that the expected points approach is the appropriate way to examine virtually all such situations.  However, there may be situations where it is difficult to have a lot of confidence in the expected points approach.  The situation in the Seattle - Washington game might be an example.  Many commentators suggested that a key factor in the decision was the injury to the Washington kicker earlier in the game.  How much data was available on the potential success of any replacement kicker (or an injured kicker)?  How many attempts did they have in NFL games?  How many from this distance?  How many in the closing minutes of the fourth quarter?  How many with the team 8 ahead?  The number of observations on which to base any decision is likely to be very small.

That is not to say that a small amount of data on a kicker of a football does not add value.  Goalkeepers in the English Premier League like to keep data on a small number of penalty kicks by opposing players.  On the left hand side of the picture is the water bottle of one goalkeeper with data on West Ham United penalty kickers.  There is only a small number of kicks for each kicker but there does seem to be an element of predictability.  It is not exactly "big data" but it might be helpful.  We might call it "small analytics".

Even where there is access to big data we need to be careful in how it is used.  One takeaway from Romer's paper is that play-callers were conditioned by what was done in the past.  Conventions were established, e.g. kick/punt on fourth down.  In expected points terms, these conventions left some points on the field.  Romer acknowledged that the conventions might be explained by other decision making criteria.  However, in expected point terms these conventions were mistakes.  They were mistakes given the expected points from kicking compared to going for it on fourth down.

It is possible that Romer's paper changed the convention.  Brian Goff clearly suggests that Bill Belichick's play-calling was influenced by his knowledge of the paper (Goff has a chapter in Sports Economics Uncut called "Bill Belichick as economist").  Would the results be replicated with data from 20 years after the original paper?

Most would object to using Romer's data to guide play-calling in 2021.  It is clearly from a different era.  Different rules, different players.  But players are constantly changing.  How does Bill Belichick or Josh McDaniels use the data to guide decisions for the rookie quarterback?  They may be willing to pay the expected points price in some situations in order to develop Mac Jones.  Maybe the Washington Head Coach, (Riverboat) Ron Rivera, was not just betting on the current situation.  Maybe he was gathering data on the ability of his quarterback to convert.  Or he may have been investing in the development of his quarterback by providing him with the experience (passing up the chance to give any alternative kicker an opportunity).  Even if the analytics did not support Rivera's decision, we should not automatically assume he was incorrect.  I support Romer's approach but not to the exclusion of all other possibilities.

By David Butler

Recently the 21/22 salary caps for La Liga became public knowledge. The most interesting insight was probably the gap between Real Madrid (1st) and Barcelona (7th) - Real Madrid now have a salary cap about seven and a half times bigger than that of their great rival. Real can dedicate a whopping €739m to salary payments.

It's my understanding that no more than 70% of revenue generated can be spent on salary, but I think there are ways to work around this by strategically managing debt. As these limits are generally based on turnovers though, these data lay bare the dire financial state at Barcelona. 
​
After only 14/15 matches we are starting to see the almost universal relationship between wage spend and output/points accrued come about (R² =.52). Below plots the wage spend (log) against points for this La Liga season.

Of course, this relationship is well-documented in the sport’s economic literature. In 1998 well-known sports economist Stefan Szymanski showed how in English leagues higher wages lead to better Performance (R² =.78) and how better performance then leads to higher revenue.
​
Will Real Madrid win the league? Based on their wage budget they should. Indeed, they can spend more than €500m in excess of the second highest budget in La Liga – Sevilla – at €200m. 

By Robbie Butler

Next month the 15th Annual Gijon Conference on Sports Economics from the Sports Economics Observatory Foundation (FOED, University of Oviedo) will be held in Gijon (Spain).

Gijon is synonymous with sports economics thanks mainly to the work of Plácido Rodríguez and others who have made the annual event in northern Spain part of the sports economics calendar. Along with Carlos Gomez-Gonzalez and Julio del Corral, Plácido has just published "The Who and the What of the Journal of Sports Economics–20th Anniversary Edition". I wrote about this recently here.

The event in Gijon next month does not focus on a 20th anniversary, but rather a 50th anniversary. As the Organising Committee say Gijon XV:
      "is a tribute to two of the pioneers in the field of economics of the sport, with two facts that they did 50 years ago.          Roger Noll organized in Washington, under the auspices of the Brooking Institution, the first Conference on Sports        Economics in 1971 and Peter Sloane, that year, published one of the most relevant articles in the field of sports            economics: ‘The Economics of Professional Football: The Football Club as a Utility Maximiser". 
What a wonderful occasion this will be, to hear about Roger and Peter, both part of the sports economics community. 

The 50th anniversary is very appropriate as the 1970s marked the start of what was probably the birth of "sports economics". 1956 and 1964 were outposts to this point. However, what has followed from 50 years ago has created the field that we see today. 

The forthcoming book Advances in Sports Economics commences with a brief history of sports economics where Noll, Sloane and others, and their seminal contributions to the field, are highlighted. The book is available on general sale from next month from Agenda Publishing and Columbia University Press.

By David Butler

Yes, ranking and seeding are determined by past performance, but I still find it interesting to look at international rankings ability to predict future outcomes. Doing so serves to show just how unbalanced European International football is (as England’s 10-0 win gave us an insight to).

The chart below plots the FIFA ranking of countries prior to the (UEFA) qualification groups being drawn for World Cup 2022 and the points per match countries went on to accumulate.

A countries rank is a good predictor of how many points on average they will go on to accumulate.
​ 
Surely UEFA are happy – all the big teams make or are  very close to making the finals and a full inventory of international fixtures is delivered with minor qualification shocks. 

This piece first appeared on RTÉ Brainstorm and can be found here.

By Robbie Butler

Last week Dean Smith and Daniel Farke became the 4th and 5th English Premier League (EPL) managers to be relieved of their duties since the start of the 2021/22 season. Xisco Munoz (Watford), Steve Bruce (Newcastle United), and August Manager of the Month Nuno Espírito Santo (Tottenham Hotspur), had already had their contracts of employments terminated before Aston Villa and Norwich City announced the latest sackings.

The attrition rate of football managers is quite something. Between 1992 and 2005 the average managerial tenure within the EPL was just 2.5 years. During the 2015/16 season alone 54 managerial sackings occurred in the top four tiers of English football.

English football is not alone. Three managers have already been dismissed in La Liga this season. Paco Lopez (Levante), Michél (Getafe) and Ronald Koeman (Barcelona) all departed during October. In Germany, Mark van Bommel was dismissed by Wolfsberg last month, having only been in charge since July, while Hellas Verona, Cagliari, Salernitana and Genoa have all sacked their managers in Italy’s Seria A this season.

However, such is the job. Would-be managers will be only too aware that tenure is dependent upon results and could be short-lived.

Just ask Les Reed who lasted a mere 40 days with Charlton Athletic in 2006. Alan Shearer and Ian Dowie fared slightly better, and survived for almost two weeks longer than Reed (53 days each) in 2009 and 2010. More recently Frank de Boer was shown the exit door at Selhurst Park after just 77 days, with Crystal Palace dismissing the former Barcelona and Netherlands international in September 2017. The Eagles had lost their opening four Premier League games and had failed to score a goal.

While the sackings of De Boer, Bruce, Koeman, Espírito Santo and countless others have often come as little surprise, does the appointment of a new manager actually improve the team?

Almost 20 years ago Bruinshoofd and Weel (2003) decided to explore this for Dutch football. The two compared the performance of clubs that had been responsible for at least one of the 125 managerial sackings in Eredivisie between 1988 and 2000 with a control group of 103 performance dips where the manager was not sacked. Performance dips were periods when a club could have sacked their manager, after a string of poor results, but decided not to. The authors found that while new managers improved the performance of teams, the clubs that stuck by their managers (control group) performed even better once recovering their form.
 
For example, from the 15th of December 2019 to 1st of February 2020 Manchester United played 9 league games. The club experienced a dip in form winning just three times, drawing twice and losing the remaining four games. This was a return of just 11 point from a possible 27. There were renewed calls at the time for the dismissal of manager Ole Gunnar Solskjær. The club stood by Solskjær and Manchester United won three of their next four league games before lockdown halted play. The return to football in June 2020 saw the Red Devils remain unbeaten, winning six of their final nine league games. Arsenal have had a similar experience with Mikel Arteta. Three defeats in the first three league games this season have been followed by an unbeaten streak, with six wins in the last nine games.
 
De Paola and Scoppa (2011) report very similar results for Italy. Examining sackings in Seria A from 2003 to 2008 the authors find a positive effect from managerial change. However, just like in the Netherlands, the effect disappears once controls are added with the authors concluding that changing a manager neither improves nor deteriorates performance on the pitch, when compared to sticking with the incumbent. A very recent study from Swedish domestic football supports the findings in Italy and the Netherlands, and concludes that sacking a manager during the season does not have a significant impact on the performance of the team in the short run.
 
To illustrate the point Flint, Plumley and Wilson (2013) used the example of Andre Villas-Boas. The Portuguese manager was sacked by both Chelsea and Tottenham Hotspur. At Chelsea he survived for under a year. He fared slightly better at Tottenham, lasting 16 months. In both cases, Villas-Boas was replaced by a caretaker manager who themselves were sacked after less than 10 months in charge.
 
Of course, some managerial change appears to have an instant effect. Performances improve and the team goes on excelling for years into the future. The key is to understand when the manager, rather than other factors (players, board of directors, level of investment in the squad, etc.) is what is holding the team back.
 
Steven Gerrard has just become the 215th appointment in the EPL era and the 123rd person to hold the title of manager in the EPL. Should Aston Villa now improve, and climb up the league table, Gerrard will get the plaudits. Would recently sacked manager Dean Smith also have done this?. Evidence from other European leagues would suggest ‘yes’. 

This Friday Kaori Narita, University of Liverpool will present "Managerial contributions to firm success: Who are the best managers in Europe?" at the ROSES Online Seminar Series.

By John Considine

There is a superb survey article in the current issue of the Journal of European Competition Law and Practice that deals with recent sporting issues.  By reviewing the period 2018-2021 it allows one to catch up on some of the major developments in the area of competition law.  All of this in the shadow of the proposed European Super League.

It considers a variety of issues.  The largest amount of space is devoted to a case where the International Skating Union (ISU) was approached by a Korean event organiser who wanted to organise two international events.  The ISU did not approve the events and threatened any skaters involved with lifetime bans from ISU events.  Consequently, two Dutch speedskaters brought a case to the EU Commission arguing that EU Competition Law had been infringed.  The speedskaters were successful.  It is easy to see the possible relevance for the European Super League if one substitutes UEFA for ISU and substitutes two Spanish competitors for two Dutch competitors.  The authors, Benoit Keane and Ben Foster, examine the implications for breakaway events - including the European Super League.

The reaction of clubs and national sporting bodies to Covid-19 is one of the more interesting sections.  The authors examine how four different national football associations were taken to task for their alternative approaches to ending (or not ending) their seasons.  Clubs appealed to competition law for a range of reasons.  They complained that one governing body stopped the season and averaged the results.  The complained that one governing body used the league standings when games stopped.  Another complaint concerned cancelled promotion and relegation.  The fourth situation involved a governing body who completed the season after the covid-19 induced delay.  Did any of the clubs win their cases?  Should any of the clubs have won their cases?

There is also an informative section dealing with financial regulation in sport.  Again, there is some interesting distinctions discussed.  One involves a club taking a case against a governing body, i.e. Saracens' case against a salary cap.  This case is worth contrasting with others where players complain against a group of clubs that reduced the salaries of players in response to Covid-19 issues.  The relevant legal authorities rejected Saracens arguments against the salary cap but rejected the way the clubs reduced players salaries.  (A Financial Fair Play case brought by a sports agent also features in this section.)

The article opens with a discussion of the word "Crisitunity" (as coined by Homer Simpson) but it is largely exempt from legal jargon.  On a number of occasions there is a discussion of agreements that are anticompetitive "by object" and "by effect".  Here the authors might have used another illustration from The Simpsons.  In an episode given his name, Sideshow Bob Roberts complains that he is in jail for a crime he did not commit, "Attempted murder, now honestly, did they ever give anyone a Nobel Prize for attempted chemistry?"  In EU Competition Law, the anticompetitive agreement does not have to have its intended effect.

y Robbie Butler

The title of this post will be a familiar expression to followers of Gaelic Games. However, what follows is not about Ireland’s national sport but the English Premier League and was motivated by a tweet I saw posted which said every Premier League goal score by a team this season is worth roughly one point.
​
I had never thought about the relationship that way. I decided to take a closer look and collected data for the last 5 completed seasons (2016/17 to 2020/21). The relationship stands up as the figure below illustrates. The R-squared is a whopping 0.94!

The top-right of the picture is dominated by Manchester City and Liverpool. City are the only team to break the 100 goal mark (twice in fact) but with different outcomes. The lower of the two City crests saw just 81 points from 102 goals (2nd to Liverpool in 2019/20) and is an obvious outlier. Liverpool have a different outlier. Winning the league with 99 points that same season with just 85 goals.

The season before Liverpool and City battled to the final day, with City winning the league by one point (99 to 98) and also outscoring the Reds 95 to 89. In fact, one goal (Leroy Sane) was the difference and denied Liverpool both the league title and an unbeaten season.

Two other observations are worth noting. First, Manchester United’s second place finish in 2017/18 under Jose Mourinho saw them collect 81 points from just 68 goals. Huddersfield were relegated the following season scoring just 22 goals. However, this should have gathered more than the 16 points they accumulated.

Of the 100 observations over the past five seasons, 9 times teams got the exact same number of points as goals scored. In another 15 cases there was just one goal between the number of goals scored and the total number of overall points.
​
In future, maybe teams could target 40 goals rather than 40 points if they want to stay in the division. 

By Robbie Butler

This week the Journal of Sports Economics published "The Who and the What of the Journal of Sports Economics – 20th Anniversary Edition. The paper, co-authored by Carlos Gomez-Gonzalez, Julio del Corral and Plácido Rodríguez is a celebration of the first twenty years of the journal. 

The abstract says "This article describes the content published in the Journal of Sports Economics (JSE) for its 20th anniversary. The analysis focuses on the most relevant topics and captures the characteristics of authors, institutions, and types of collaborations over a 20-year period. In total, the sample includes 663 research articles (published from
February 2000 to December 2019) and 856 authors from 490 institutions. The 20-year period offers an insightful picture of sports economics research. The article discusses additional features and trends and reflects on expected directions for future research in the journal."

Among the range of data and issues addressed, one aspect focuses on the "evolution of the topics analyzed in JSE contributions." The trend is quite stable. Questions examining "behavior" and "competitive balance and demand" are 1 and 2 for the 4 five-year periods between 2000 and 2019. Combined these make up about 30%-40% of all accepted submissions. Competitive balance and demand are clearly the most cited papers, with the top 4 cited papers all listed under this heading. Each has more than 300 citations to date. 

Since 2015, soccer has been the most written about sport. Baseball appropriately dominated the 2000-2004 period but has since dropped to 3rd spot on the list, with American Football in second. 

There is also a really nice map called "Geographical concentration of contributions to the JSE". The paper states: "Map 1 displays the percentage of authors who have contributed to the JSE from institutions in different countries and yields a significant imbalance. While the journal is missing the contribution of authors from several countries, especially in Africa, Central America, and the Middle East, four countries concentrate more than 75% of the contributions: the United States (60%), followed far behind by Germany (7%), the United Kingdom (5%), and Spain (4%). Other countries in the Top 11 are Canada, Belgium, Switzerland, South Korea, Australia, Italy, and France."

The full paper, which is brilliant reading, can be found here.

By David Butler

​Last week, the Irish government published legislation to facilitate the establishment of a gambling regulator. Among other rules, new legislation could be on the way to end free bet offerings , credit or loans to players, and VIP services.  

To date, I have not seen much on rules relating to how information is presented to gamblers. A myriad of behavioural research has shown that how information is displayed, it’s context and where it is positioned can impact decision-making.

For example, bookmakers commonly offer ‘specials’ to customers that are prominently positioned and brightly coloured to attract attention.  A common tactic in football betting markets is to combine two very short priced favourites with one longer priced team. All three must win (the conjunction fallacy comes to mind). The extra value to the gambler is often displayed.

I tracked this for football specials for a week on a popular gambling website. About 20% of these specials were winners. Indeed, there doesn’t seem to be anything very special about them.  

If one was to accept that we cannot eradicate problem gambling, maybe it in the interest to improve gamblers literacy of the tricks and traps used by betting firms in an attempt to reduce losses.

By Robbie Butler

Last month our paper "Changing Competition Design and Spectator Turnout: Evidence from the League of Ireland" was published in the Economic and Social Review. The very last line of the paper says:

"While specific strategies to foster habit formation are challenging to pinpoint, a precise policy recommendation arises from the scheduling effect we identify – there is a need to ensure as many fixtures as possible take place between Friday night and Sunday afternoon."

This is one of the key recommendations of the paper. Attendances are higher for Friday and Saturday matches. Last weekend provided a very nice example of this.

Bohemian (Bohs) played Waterford twice in 36 hours. Both games were at Dalymount Park in Dublin. The first was an FAI Cup Semi-Final on Friday night and had full capacity (3,100 approx.). The second was a league match on Monday night. Attendance is reported as "2,000".

Admittedly the games are not the same competition. The cup match generated significant interest and had 900 away fans. The league game was more important to Waterford (from a relegation perspective) but saw a sizable decline in travelling support. This is to be expected as it was played on a Monday night.

Even accepting the changing circumstances and competitions, a fall in attendance of about 35% in the space of 36 hours is quite something and demonstrates the importance of scheduling.

The full paper can be accessed here.