Soccermetrics

Research paper schedule released

The MIT Sloan Sports Analytics Conference has released the list of finalists of their research paper competition.  I am surprised that there are only four papers in the final competition; I would have expected six to eight.  One of the papers is -- surprise, surprise -- on the Pythagorean expectation in baseball, examining how run distribution affects the win percentage of a team.  It's written by a physics professor at Iowa State. 

Sloan conference updates its agenda and location

The Sloan Sports Analytics Conference has released its schedule of events.  Several of the sessions will run concurrently, but not as many as I feared.  Sunil Gulati will appear in a session parallel to the Emerging Analytics session (with the officials from Man City and Chelsea), so there may be some popping in and out of rooms.  I will attend the analytics sessions (baseball, basketball, and emerging), as well as the Research Paper Presentations throughout the day.  Of course, the real action happens in the hallways and during the break periods.

The one session that I am looking forward to, other than the Emerging Analytics session, is this one on the limits of "Moneyball", moderated by Moneyball's author Michael Lewis and paneled by Mark Cuban, Bill Polian, Daryl Morey, and Bill Simmons.  I've written a post already on the limits of soccermetrics and I'm interested in finding out whether the panel touches on the same points I argued.  Analytics are great tools, but they're just tools, and are no substitute for a functioning brain in the sports team's front office.

The other major news from the Conference is that it will no longer be held at the Sloan Business School.  Instead, it will be at the Boston Convention and Exhibition Center, a much large venue which indicates how much interest in the Conference has grown in recent years.  I'm just glad that I hadn't already reserved a hotel room!

Some modifications to the Soccer Pythagorean formula

Two weeks ago I said that I would post the document containing the derivation of the Soccer Pythagorean formula, but I was trying to complete a modification before uploading the paper on the site.  It had to do with the fact that the predicted percentages of wins and draws exceeded 1 in some cases, which is impossible, of course.  It turns out that there is an overlap in the integration intervals that I used to calculate the win and draw probabilities, and I aimed to fix that by recalculating the interval in the expression for win probability.  A side effect is to reduce the over-prediction of wins by the soccer Pythagorean, which I estimate would reduce the estimated point total by 15-20 points.

It turns out that the small change makes the integral extremely complicated to solve.  For the one or two of you who care, here's why.  Skip the next paragraph or two if you don't want to know. 

The problem is that when I perform the integration with that modification to the intervals, I end up with a binomial series of the form

which is an infinite series because γ, more often than not, is not a rational number, so I have to make an approximation.  Then I end up with an integral that looks like this:

and if you perform an integration by parts you end up with another infinite series with an infinite chain of integrations. 

I've almost figured it out, but the bottom line is that (1) a simple modification produces an extremely complicated problem, and (2) the resulting formula is nowhere near as simple as the baseball Pythagorean.  The price, I believe, of having a sport with three possible outcomes that occur quite often.

I want to include some results and look at some of the current European leagues using the formula, but I wanted to hold off until making the modifications.  I can upload the original paper that I submitted to the conference tomorrow if there's sufficient interest.  What do you think?

The 7th World Congress on Science and Football

The World Congress on Science and Football is a quadrennial conference that draws academics and professionals from all over the world to discuss issues related to the intersection of football with the physical, mathematical, and social sciences.  It is unique in that it has discussions on all of the football codes -- from association, to American/Canadian football, Australian and Gaelic football, and rugby union or league. 

The previous Congress was in Antalya, Turkey in 2007.  The next Congress will be in Nagoya, Japan from 26-30 May 2011.  The call for papers will be released in July 2010 on the following topics:

• Sports biomechanics and engineering
• Match and skill analysis
• Exercise physiology and training science
• Sports medicine and rehabilitation
• Social science and management
• Psychology
• Interdisciplinary approaches to football

It appears to be a conference more geared toward the academic researchers, as they have the travel budgets to be able to fly out to Japan for a week.  I know some of them read this site, and others might find it on interest as well.  

The second leg effect in Euro club competitions

Lionel Page and Katie Page, "The second leg home advantage: Evidence from European football cup competitions", Journal of Sports Sciences, 25(14): 1547-1556, 2007. [Scribd]

Following up on my discussions of the two-leg competition in football is a very interesting and thorough paper written by British and Australian researchers.  I came across it while reading a Wikipedia entry on the two-leg format in sports.  The paper looks at the last 50 years of European club competition (Champions Cup/League, UEFA Cup, Cup Winners' Cup) and establishes the existence of an advantage to teams hosting the second leg, even after controlling for extra time and the difference in team quality.  They note that this advantage has declined over time, but didn't give a clear reason for that.

I don't have time to make lots of comments on the paper now, so I'll just post a link so that you can take a look at it and I'll talk about the publication later.

A closer look at the two-leg format in the UEFA Champions League

In my last post on the two-leg format in soccer (MLS Cup Playoffs vs. UEFA Champions League), I determined that in a typical two-leg series, the club hosting the second leg has a 60% chance of winning (3:2 on in betting parlance).  I was wondering if that probability held up during each elimination round in either the preliminary or knockout phase.  My initial expectation was that the stronger teams in the preliminary round do not always host the second leg, so the probability of the second leg host advancing would drop closer to 50%.  In the knockout stage, the group winners host the second leg of the first knockout round and the group results are used to determine seeding in subsequent rounds, but the teams in the knockout phase are more even, so the probability might drop as the competition progressed to the final. 

Welcome Sounders!

There are times when I see a burst of visits to a long-ago written post and I ask myself, "Why could possible be the cause of that?"  Such an event happened today when I saw a lot of hits on my latest CONCACAF coefficients list.   So, thanks to Dave and those of you at Sounder At Heart for sending some traffic this way! 

Thanks for visiting, and look around, tell your friends, and leave a comment or two.  There's a post or two in the pipeline that jumps off a thread posted on Sounder At Heart some time ago, so stay tuned for that.

MORE:  Ahhhh, that's why!

Player valuation, with some help from options pricing theory

Radu Tunaru, Ephrahim Clark, and Howard Viney, "An options pricing framework for valuation of football players", Review of Financial Economics, 14: 281-295, 2005.

It is important for football club management to know how much their players are worth on the open market.  This publication is the first to use concepts from options pricing to develop a framework for determining the value of football players from the perspective of the club that holds his contract or the club interested in purchasing his services.  This valuation system opens the door to a variety of options for managing or trading the contracts of players.

Moving forward

Well, I've had some more time to digest the news about my paper submission.  I'm not happy about the final outcome; I thought (like any other submitter) that other people might be interested in what I would be presenting.  It doesn't always work out that way, for a variety of reasons (paper not good enough, not the right audience, an outsider -- that last item occurs more than you think), but it's part of life at these conferences, especially one that has a paper session for the first time.

At any rate, I'm shaking off my initial disappointment and moving forward.  The soccer Pythagorean study, the publication and its results, will be posted here by the end of the week.  There are some other posts on other topics that I will post in the following days, as well as what I hope will be the start of a discussion on what makes a good statistic for soccer.  There may be some other announcements in the coming months but I can't give much details at this time.

What I can tell you is that I will attend the MIT Sloan Sports Analytics Conference.  Even though I won't be presenting, I believe it would be an excellent opportunity to meet some leading figures in the world of sport who are actively working on analytics, and most importantly get some networking opportunities with those who work in performance analysis in soccer.   There will definitely be multiple posts on my conference experience in March. 

Finally, my sincere thanks to Dave, Aditya, Eric, and Axel for their assistance with data collection for the soccer Pythagorean work.  I wish I had some way to compensate you for your work, but I don't.  At least not yet.

Paper outcome

About an hour ago I received the following email from the conference organizers:

Thank you for submitting your abstract to the 2010 MIT Sloan Sports Conference. Unfortunately we received an overwhelming response and, though your paper was excellent, are unable to invite you to the conference at this time. We anticipate that this will be an annual event and we welcome work from you in the future.

Thanks again for your interest we would love to see you there in March.

-----------

Needless to say, I'm very disappointed and a little discouraged.  I'm not as motivated to attend the conference as I was a week or two ago, but I might be by the end of the week.  There would be some good networking opportunities. 

As for the Pythagorean paper, I'll post it on this site in a couple of days. There were a few revisions that I needed to do to correct the total probability exceeding 1.0, and when I finish those, I'll post them as well.