After a month that included OptaPro’s first analytics event in North America and the fifth OptaPro Forum, you’d think I’d have reached my quota of analytics conferences. However, the MIT Sloan Sports Analytics Conference (SSAC) in Boston has been a firm fixture in my calendar for seven years now and the growth of the event to around 5,000 attendees is a reflection of this industry’s progression.
Like me, many of you reading this will already attend the growing number of industry conferences around the world and know that there are significant challenges around these events remaining unique and insightful. I walk out of panel sessions at Sloan every year hearing murmurs of: “Well, they didn’t say much there did they?” and “I could have watched ESPN for that”. I believe this is short-sighted as it very much depends on your own mind-set and what you want to get out of these sessions. Of course, two opposing GMs from the NBA are not going to sit on a panel and lay out exactly how they are using analytics in their draft strategy. However, Sloan does provide a great opportunity to hear a ‘State of the Union’ from each sport when it comes to the impact analytics is having on their game. If you’ve been to this event as often as I have, you can certainly see how the impact is growing year on year.
Sitting in the panel sessions containing some of the most recognised names in analytics, general managers, team owners, former players and hosted by the great & good of ESPN on-air talent, there appeared to be a (possibly unintentional) theme running through this year’s conference. All sports go through an evolution cycle, such as the impact of professionalism in rugby union since 1995, the introduction of the 3-point line in the NBA in 1979 or the ‘bigger, stronger, faster’ revolution across all sports since the dawn of sports science.
What SSAC provided this year was an insight into analytics’ relationship with the evolution of each sport, and the challenges and opportunities that this has presented to league decision-makers and commissioners.
Looking specifically at NHL, in 2004-05 there were no games and no Stanley Cup Champion for the first time since 1919. Unfortunately, this wasn’t for an exciting reason like all the NHL players deciding they were all going to play in the Russian KHL, but for the lack of a Collective Bargaining Agreement between the league and its player association (CBAs are common in all major US sports and show the power the players have over their leagues & associations that really doesn’t exist in soccer).
However, this non-existent season is seen by many as a turning point in creating today’s exciting, fast-paced modern NHL and data played a part in shaping this evolution. In 1993-94 goals per game were at their lowest level in 20 years (6.48).This was a problem for a league that sees as many as 80 shots per game, and for the next 10 years this trend continued with goals per game never rising above this rate and often dropping into the low 5s. NHL Commissioner Gary Bettman took on this role in 1993 and for the next decade was tasked with solving this problem, but by the time the 04’ lockout came, shots per game in previous seasons had been as low as 27.3 per team and goals per game as low as 5.14.
Armed with this information, Bettman was able to shape and implement rule changes in the post lockout season that drastically changed the style of play to consequently create a faster more fluid and exciting brand of sport. These changes saw an increase in the numbers with shots per game in 2017/18 at 31.8 per team and 5.89 goals per game, a figure steadily rising since the lockout, although not as dramatically as first imagined (for more on the impact of rule changes in the NHL, please read ESPN’s Sean McIndoe’s excellent article chronicling the period described here).
However, with that history lesson complete the NHL now finds itself in an interesting position as it sits at the bottom of the analytics pack when compared with other major US sports. In a panel that included Gary Bettman and Tampa Bay Lightning owner Jeff Vinik, they discussed the challenges of using analytics in the ‘fluid game of hockey’ (along with introducing tracking data to the league). This was fascinating because that is the same challenge that has been levied within soccer for many years.
However, we know that with the amount of investment soccer teams are placing in performance analysis and technical scouting departments along with the type of tactically applicable work that we see at the OptaPro Forum (and beyond) each year, these challenges can be overcome.
“Players are on salary caps, but analytics departments are not. This is where you can differentiate yourself.” – @devinpleuler #MajorLeagueData #SSAC18 #TalkDataToMe
— Sloan Sports Conf. (@SloanSportsConf) February 24, 2018
Baseball is often seen as the pinnacle of analytics and it has been long discussed how the sport is better suited to this style of analysis. The work that has been done over the past few decades to model event & scorecard data has meant that any baseball analytics session at Sloan has been the hottest ticket in town with people waiting to hear the latest advancement beyond WAR and any insight that can be gleaned as to what the new OBP equivalent is that are giving teams the edge. However, on the face of it this progression has slowed considerably with one baseball panel this year essentially becoming a nostalgic look back at the ‘good, old days’ when these metrics were first created.
Watch: Next Frontier in Baseball Analytics
What has really changed is that the StatCast ball and player tracking data is available to all MLB teams but is not publicly available in the same way that box score data has historically. Although an increasing amount of this data is being shown on MLB broadcasts and in the media, it has meant that the number of people able to access and model this new data set has decreased significantly. The consequence of this is that although MLB GMs and analysts represented on panels discussed very broadly the significant impact metrics like launch angle and out probability were having on their assessment of players, they were not keen to divulge much detail. There has perhaps been a shift from the past where public analysis could arguably sit alongside (and in some cases inform) work in the professional game, which paved the way for a more open discussion.
However, with players increasingly discussing ‘launch angle’ when describing their improvements in batting average and OPS, this new wave of metrics is starting to leak out into the consciousness of baseball fans. With power numbers again on an upward curve in MLB, it gives another indication of how much analytics continues to shape the evolution of the game.
“I wasn’t smart enough to see that maybe I should have shot 20 times a game. Now we see the point guard a little differently; it is your first point of attack. To a fault I was almost too much of a facilitator.” – @SteveNash #7Seconds #TalkDataToMe #SSAC18
— Sloan Sports Conf. (@SloanSportsConf) February 24, 2018
With those two sports at different points on their analytics journey I found this year’s SSAC basketball panels the most intriguing. Like hockey, early basketball analytics focused on easily accessible shot data, with the majority of the research or publicly available work focusing on various shooting charts and efficiency assessments. However, over the past few years more work using tracking data has showcased some of the most innovative and exciting analytics in sport. Whether this is analysis evaluating how changes in offensive decision-making can effect team scoring efficiency (Replaying the NBA) by Luke Bornn or a research paper identifying the effectiveness of using a double team defensive strategy (The Advantage of Doubling).
The concepts of ghosting and spatial analysis that are starting to drift into soccer were inspired by basketball analysis and what I find most interesting about this is the particularly strong tactical applications. Teams are now able to use this combination of event and tracking data to identify patterns in the plays and how they map offensive and defensive strategies.
The level of innovation and development in analytics in the NBA is impressive but a conversation around the shift in offensive style followed my takeaway theme of SSAC ‘18.
A few years ago at SSAC I was trying to describe to non-basketball fans the dramatic increase in 3 point shooting by using the phrase, “Of course, 3s are worth more than 2s”, which was obviously a mistake when surrounded by data scientists. Despite my slightly over simplistic description, this was essentially the discussion that was had on the Inventing Modern Basketball panel. They described how the historic 05/06 Phoenix Suns team who ranked first in all major shooting categories including 3-point shooting (2097 attempted) would now place them 19th in the current season. This shift has been significantly impacted by analytics, with Battier describing how data began to influence strategies such offensive rebounding and the need to balance this with getting back on defence with the increase in shots taken quickly in transition. With the success of the Warriors’ all-out shooting method in recent years’, analytics will continue to play a role in finding ways to nullify this strategy and a new style of play will evolve.
Watch: Inventing modern Basketball
Back to soccer
With the current position of those sports mapped out, what is the ‘State of the Union’ in soccer? At both the OptaPro Forum over the past few years and on recent soccer panels at SSAC, the focus has been on how we can use data to describe tactical concepts that can more effectively impact the game. At both events we have seen work combining event and tracking data to conduct spatial analysis such as the research paper (Wide Open Spaces) from Luke Bornn & Javier Fernandez (FC Barcelona) that identified how effective players are at controlling and influencing space. This type of analytics not only has direct coaching implications, but more importantly could be presented in a way that can be easily digested by those coaches. With soccer played on a more global scale than the four major North American sports, identifying shifting trends in style of play in the game is more difficult with cultural or league styles a more relevant analysis. As the discussions around each of the sports developed at SSAC, the question was posed “Does it matter that our sport keeps evolving?”. In all these North American sports the leagues have imposed various rule changes to increase offence in some way and tailor the sport to a certain style, but in soccer the diversity of styles is one of the aspects that makes the sport the global game.
Analytics is clearly influencing team tactics. Whether it’s Tony Pulis’ focus on set plays, Burnley’s efficient ‘solid defence leading to attack’ style or even Manchester City taking possession football to new levels, they all have a root in analytical concepts. As with the situation the NBA & MLB finds itself in, the most advanced analytics in soccer is happening behind the closed doors of club training grounds all over the world but it is happening and it is really changing the game.