Today is the day of the Soccer Analytics Panel at the 2012 MIT Sloan Sports Analytics Conference.  There are some late changes to the panel, but I’m looking forward to a stimulating discussion from the panelists.  Follow along for the liveblog at 1300 EST (1800 GMT).

To refresh your memory, the panelists are:

• Steven Houston (Hamburg SV)
• Scott MacLachlan (Chelsea FC)
• Steven Brown (Everton FC)
• Drew Carey (Seattle Sounders FC)
• Alexi Lalas (ESPN)

Marc Stein (ESPN) returns as the moderator for the Soccer Analytics panel.

1248: I’m here, WiFi is up, and we’re ten minutes away from the Soccer Analytics session.

1251: Steven Houston’s affiliation is given as “Hamburg FC”.  Not quite right…

1254: I should say at the outset that this liveblog is not a word-for-word transcript of the panel. I try to capture as accurate a summary of the panelists’ statements as possible, but any errors in the reporting are my own.

1256: Another apology for not having comments working on the blog…I will be reading Twitter and you can leave comments on @soccermetrics.

1300: All of the panelists are on now.  Introducing the panel…and we’re off.

1300: First question on: “Can we look a few signings and say that they were analytics-driven signings?”

• Houston: Amazing to see impact of analytics awareness among team front-office.  Difficult to get buy-in across the board.  Lot more short-team deals (loans, etc) which tend to be driven by data.  Important to get data in as part of the process, preferably at the beginning.
• MacLachlan: Chelsea, Everton, Fulham, Man City, Tottenham using data actively.  Dortmund, Hamburg on the Continent.  Clubs in Italy using some analytics to recruit in South American market

Stein saying “the MLS” instead of “MLS”

• Carey: Seattle is analytic-driven anyway, with Microsoft and Amazon in the area.  Still haven’t found a magic formula, but neither have anyone else.
• Lalas: Important to have someone champion the idea, and also that those champions become champions.  Have to face economic reality in MLS.
• Carey and Lalas making the point that due to small budgets, imperative to make best of choices, using analytics, also diet/medical practices.  Salary cap constraint driven.

Stein “if there was salary cap in Europe, this would be more of an issue, right?”

• Houston: Because of German ownership structure, analytics does have opportunity…most useful when it comes to scouting, how to effectively evaluate players.  Data used to filter players, then use scouting to evaluate those players.  Not scouts vs numbers issue.
• Houston: We have all of the spatial data, haven’t done a good job of explaining those data to managers, executives.  Need to make use of data scientists

Stein asking Steve Brown about use of data in match preparation…

• Brown: Data analysis in performance analysis in infancy.  All EPL clubs working with same level of data now, difference-maker is extraction of actionable information.
• Talking about how Landon Donovan used data in match preparation.  Trending information, tendencies of opposing defenders, etc.

Asking Lalas about the use of videotape during his career…

• Lalas: I learned early on that holding possession in defense was important to the quality of his play.  Wanted to track that information in-game and have it relayed to him.  Data is nice, but data with video is very powerful.  Data without video isn’t worth as much.
• Houston: using data to find out which video to show.  Making data relevant to end-user is extremely important.
• Brown: coaching staff has to see use in it, that is applicable to them.
• Lalas (to Brown): but that’s your job to develop use from it (the data)
• Carey: Culture of team is very important in receptivity to data. Open to best practices

Stein: “What is it going to take for continental clubs to be receptive to data?”

• Houston: “Old Guard” guys actually receptive when you’re making their job easier!
• Lalas: Especially true if it confirms some tacit knowledge
• Houston: German football more analytical esp with youth teams and national team…has guy who led match prep for England v Germany also works for him
• Borussia Mönchengladbach and Dortmund also incorporate advanced analysis.  Houston gave example of a Japanese player (Kagawa) found using technical scouting, now a big star in Bundesliga

Lalas asked how analytics could have been used when he was a GM.  Hoo boy.

• Used analytics more on the business operations of the MLS teams.  If he thought of adapting similar tools for on-field analysis, perhaps things could have gone better.
• Lalas: I’m more convinced of applicability of data than before…but as a tool.  Lack of awareness, budget driving factors.
• Carey: Seattle very desperate for advanced information on players, especially mid-level ones (bulk of players in MLS).  Find best players at mid-level prices –> successful team in league.  (Lalas: especially international players at that market range)
• Houston: Also need to know which data to look at…some metrics can be deceiving, especially physical data…Other perhaps undervalued ones more important.
• Brown: For too long we have paid attention to data without context
• Houston: Complication with data…really interesting data in soccer not publicly available, not available from box score.  Most clubs don’t even have software developers in house.

Very good discussion of the search for quality metrics from basic data.

• Spatial variables are there…Lalas saying that it is the final frontier because of the nature of the game.

Stein: “How much data sharing is there between clubs?”

• Not much sharing in general…want to keep competitive advantage?
• Brown: most teams do internal coding to collect data relative to team’s game plan, style of play.  (Verifies what I’ve learned.)

Stein: “Are we closer to getting measures of off-the-ball movement?”

• Brown: We have the (x,y) data, so it should be possible
• Houston: Defensive metrics hardest to develop

Would be useful to say that the nature of the game makes it challenging.  Could use someone to describe it from a technical POV, but the risk is losing people with the technical detail!

• MacLachlan: Saying that knowing numbers without intuitive knowledge risks losing credibility.  “Would a 19-year-old out of Uni tell someone like Harry Redknapp that his lineup formation is wrong?”
• Houston: A mentor told him, “99% of your scouting is who you don’t sign”.  Use data to highlight players, tell scouts where to go.

MacLachlan: using data to do due diligence also important. Can’t make a multi-million pound deal without knowing anything about your player.

Stein asking for example of data-driven decisions at halftime

• Brown: data-driven vs video-driven decisions.  Giving example of prep for 2009 FA Cup final (Everton vs Chelsea)

Are goalkeepers harder/easier to analyze?

• Brown: goalkeepers difficult to analyze, especially with conventional stats.  (Why not ball distribution?)
• Most teams doing penalty analysis as well (even England national team?!?)
• Houston: try to look at save percentages over long period of time
• Maybe drop in penalty conversion due to knowledge of player tendencies

Set-piece preparation (and penalty preparation) more amenable to analysis

• Lalas: Knowing how to exploit set-pieces hugely important

Audience Question: Are analytics situation-dependent, or is it possible to develop best formation from analytics?

• Carey: Well all plays are situation-dependent, no?
• Brown: Gives example of Stephen Pienaar
• Houston: We’re behind on +/-, optimal formations
• Stein asks if analysts have time to do that.  Houston responds no — demands on time esp mid-season.

Audience Question: What kind of data useful for predicting development or value of player?

• MacLachlan: data nonexistent, also legal issues involved.  Huge uncertainty ellipse
• Brown: EPPP rule coming in UK (or English FA).  Will essentially require data collection/analysis for player development at Premier League academies
• Get the impression that it’s ten or more years away
• Carey: Have to consider competition level, esp with college players in USA.  Having own academies also helps, control of own data. “I don’t think the college system is good enough for MLS.”  (Very true statement.)

And that’s it. Thanks for following along!