Tuesday, June 28, 2011

Using the Transfer Price Index to Set Expectations at Aston Villa

My latest post at the Transfer Price Index blog is up, and is a must read for any Aston Villa fan.  The post also dissects some of the differences between the long term (M£XI) and short term (m£XIR) transfer models, and identifies when it is appropriate to use one versus the other.  Again, it will be some time before I cross post the material here due to the desire to keep things over at that blog given the proprietary nature of the data.  Please head on over to the TPI blog and check it out!

Monday, June 20, 2011

Taking Requests for m£XIR Analyses

Response to my post analyzing the effects of starting XI transfer cost on match outcome at the Transfer Price Index has been wonderful.  Thank you to everyone for your feedback!

One of the re-occurring requests has been one of additional manager or team analyses using the m£XIR model.  Some people are looking to pick on their least favorite managers, while others are looking to see how their favorites stack up.  This is certainly why the model was created - to analyze club and manager performance!

I've set up a Google Spreadsheet to capture such requests.  Please take a look at what's been requested so far, and if you don't see your request in the list feel free to leave a comment here or send me a tweet.  I'll certainly add it to the list.

A few caveats are to be made.  First, I will have to see how I can fit such requests into my crowded summer schedule.  Second, posting will also be subject to approval at the Transfer Price Index blog.  I am not the only contributor at their site, and the requests must be balanced against the wider blog material there.  Perhaps a few of the posts end up here instead - who knows?  We'll have to play this as it goes.  In summary, it may take a bit longer to get through the requests than some of you may desire, but I will work through them.

Let's do some crowdsourcing - send me your ideas, and I will blog about them!

The Effect of £XI on Match Outcome

I know I haven't posted much over the last two weeks, but that's not because I have been sitting idle. Rather, I have been creating a post over at the Transfer Price Index blog that creates a model that links the likelihood of match outcome (win, tie, loss) to venue (home, away) and the transfer cost ratios of the starting XI (m£XIR) of the two teams. It's 13 pages long when printed, and serves as the third leg in the three foundational models I have created for the Transfer Price Index (match outcome vs. venue and m£XIR, table position vs. average £XI, and table position vs. average Sq£).

I highly recommend jumping over to the post and having a read - it will be a while before I post it here as the data it is based upon is exclusive to the TPI. I will, however, use the m£XIR model extensively in posts at my own blog, so a good understanding of the foundational document is recommended.

Yet again I must express my extreme gratitude to Paul Tomkins for allowing me to use the data, and to Graeme Riley for his updating of the data set with the 2010/11 season and transfer data. Especially important was Graeme's meticulous reconstruction of the starting XI of every match in the history of the Premier League and the 2011 CTPP costs for each player in the starting XI. He's got the tough, time intensive job compared to the tens of hours I take in analysis and writing I partake in once he's done the heavy lifting. There aren't enough pints in the world for me to repay him and Paul for the data they provide.

Sunday, June 12, 2011

The Difference Between Stats and Analytics In Pictures

This is an example of what a friend and I call "stupid [sports] stats". Image courtesy of xkcd.com.

This is an example of soccer analytics. Image courtesy of me, data courtesy of DogFace.

Always striving for the latter so that we can provide much better insight than the former.

Friday, June 10, 2011

Friday Night Links

It's been two weeks since I last posted my favorite links of the week, so yet again this week's post will be rather long.  I'd also like to update everyone on my plans for the blog this summer.

Summer Plans

Over the last seven months I had maintained a rate of about three posts per week: two posts of original material, and one that contained links to my favorite sites or a book review.  For the summer I will take a less structured approach.  My outlook is that posts will be less frequent, but of thicker content based upon what I am working on.  I already have a number of large projects on my radar, so there will be a good bit of foundational work that is laid this summer in preparation for the final third of the MLS season and the kickoff of the 2011/12 Premier League season.  These projects involve some pretty detailed and lengthy statistical analysis.  To ensure the resulting posts' quality, I must take things slowly.

Here's a summary of what I have planned this summer:

  • A two part series on the impact of starting XI transfer costs on the outcome of English Premier League matches.  The first post should go live next week over at the Transfer Price Index blog, with the second post likely following about a week later.  This series takes what we learned earlier this year about the effects of squad transfer costs over the long-term and brings them back down to the individual match level. Definitely some interesting conclusions there, and you'll never guess who played the most lopsided match in Premier League history from a financial standpoint.
  • An update to my M£XI and MSq£ analyses using the 2011 Transfer Price Index data.  We'll get to see how much or little the regression equations have changed, how the team and manager over/under performance rankings have changed, and which teams over/under performed in 2010/11.
  • A new regression analysis using the 2011 Transfer Price Index data to show what a club's cost-per-point should be based upon their squad transfer cost.  This will allow for the identification of how efficient teams have been at earning points based upon their cost structure.
  • A several part series studying the transfer habits and club results of Arsene Wenger.
  • To-be-determined analyses that will result from the combination of MLS players union salary data with Climbing the Ladder's player lineup database.
As you can see, it will be a busy summer!  Stick with me, and I promise the reading material will be highly insightful if not as frequent as in the past.

My Favorite Links from the Last Two Weeks
  • Adidas has a great commercial highlighting what can best be termed as soccer culture, both here in the States and around the world. 
  • Auguso Neto, courtesy the Run of Play blog, has some interesting social commentary on gladiator sports, soccer, and how our world is becoming more Roman every day.
  • Anyone who reads my blog knows I use the economics of soccer to tell a story.  It's no different if the clubs are in the Premier League or in MLS.  Want to know how important this weekend's derby against Vancouver is for Seattle Sounders fans?  Look to the price of a ticket to that match compared to others throughout the season.
  • Chris Anderson at Soccer By The Numbers summarizes European soccer revenue.
  • Speaking of Chris, can anyone help him out in explaining why Ligue 1 in France and the Erdevisie in The Netherlands seem to be so different than the other European leagues when it comes to goals per match?
  • I can't help but agree with MLS Talk when they critique Sunil Gulat's excuses for not siding with the FA in sending a message to FIFA by abstaining from Sepp Blatter's re-election.
  • EPL Talk asks why there aren't more team's like the Clough's Derby County and Nottingham Forest or Kenny Dalglish's Blackburn Rovers - teams that aren't glamorous and are built from the ground up.  Sadly, Dalglish's Rovers were an aberration in the Premier League era.  The genie can't go back in the bottle, and only Financial Fair Play rules can help level the playing field (although they'll never eliminate the disparity).
  • The Swiss Ramble summarizes the transfer funds available to Arsene Wenger, and has a few thoughts on where the Gunners might spend that money.
  • Howard Hamilton at Soccermetrics shows statistically that Arsenal was justified in winning the First Division in 1989, but a Liverpool championship could also have been justified.
  • On Footy asks "Can You Win in MLS without a Big Man?"
It may be the off season in Europe, but it's a weekend full of soccer here in the US.  Our national team is in action on Saturday in group play for the CONCACAF Gold Cup (our region's national team championship).  The winner of the Gold Cup goes on to the 2013 Confederations Cup and gets the benefit of playing in World Cup 2014 stadiums a year ahed of the tournament..  My Seattle Sounders are also at home this Saturday in that derby match versus our neighbors from the north.  It will be a busy Saturday for me.  I hope your weekend is exciting, and I'll see you back here next week!

Wednesday, June 1, 2011

A Few More MLS Salary Statistics

I thought I'd briefly return to a topic that was big enough to get me to take a brief break from my recent three week blog vacation: 2011 MLS salaries.

Recall the following conclusions from my previous post:
  • Median pay was actually down 13% from 2010 to about $80,000 per player per season.
  • Median pay per match was down 24% when one accounts for the four additional games teams will play this year with the admission of the Portland Timbers and Vancouver Whitecaps to the league.
  • New York has led the way the last two seasons in terms of total team payroll after LA had led that category for the previous three years. LA did close the gap this year with the signing of Juan Pablo Angel as their third DP.
  • No other team in the league spends even half as much as LA or NY.
  • MLS crossed the 500 player barrier this year, which was driven by the addition of Portland and Vancouver as well as an increase in the number of roster spots for each team allowed by the league.
  • Ultimately, the league spends as much on total payroll as Chelsea spent on the transfer fee for a single player (Fernando Torres). Since my last post, an additional comparison could be made - in 2011 MLS only spends slightly more than what Don Garber is seeking from the next expansion franchise (scroll down about 1/3 of the way in the linked article to see what I'm talking about).
This all made me think of three areas in the 2011 salaries worth exploring

  1. The effect of designated player salaries
  2. The effect of player position on median pay
  3. The number of players on each team.

Designated Player Salaries

The designated player (DP) may present one of the more interesting contrasts within the league - they're meant to grow the status of the league in the minds of American sports fans while not breaking the bank. In some ways, MLS is trying to re-create the magic associated with certain NASL teams while not making the same fundamental mistake as the NASL -  having team expenses outpace revenues.

This results in a huge pay disparity between DP's and non-DP's. The graph below shows the sum of DP salaries in constant 2011 MLS dollars, as well as the share of total league payroll that DP's utilized (click on the graph to enlarge).

While DP salaries (and their associated share of total MLS pay) dipped in 2009, they regained their steady climb to a new high in 2010 of 38.4% of league salary expenditure. In the 2011 season there are 17 DP's in the league that earn a combined $28.5M per year and take up 35.6% of the total pay in MLS. This means the remaining 501 players have $51.6M to split amongst themselves (or an average of $103k per player).  If the DP's weren't afforded exemptions from normal salary cap rules, the average player pay would increase 50% to $155k per year. For a league that has cried poverty and has focused on improving the quality of play on the pitch it seems strange that they would allocate so much of the league salary expenditure to so few players. I've always questioned how the league expects to get measurably better in the eyes of the soccer-watching public if they're attempting to buy superstars instead of using the money to build teams.

Median Salary and Number of Players by Position

Another way to look at league salaries is to break out median salary by position. The graph below does just that, plotting the salaries of forwards, midfielders, defenders, and goal keepers in unadjusted dollars as well as the count of each type of player by season (click on graph to enlarge).

The graph clearly explains that midfielders and forwards are the best compensated in the league. However, that gap seems to be closing. While the number of players at each position has gone up from 2010 to 2011, it appears that the greatest gain was in the midfielder position. In fact midfielders only outpaced forwards by about 25 players through 2009, but midfielders have now become the greatest number of players in the league by nearly a 65% margin. Median pay within these two groups fell significantly in 2011, while defender pay remained constant and goal keeper pay was slightly reduced. The superior numbers and steeper pay reductions seen in the midfielder and forward ranks are what drove the magnitude of the overall decrease in median pay between 2010 and 2011.

Number of Players per Team

The increase in the number of players at each position is also showing up on each team's roster. The graph below shows the number of players per team over the last six MLS seasons (click on graph to enlarge).

From 2010 to 2011 teams have added an average of four roster positions. This should bode well come playoff time (along with the front loaded schedule for teams participating in Champions League), as I showed in a previous post that the difference in the number of games played over a season between first round playoff teams is a statistically significant determinant of the winner of the series. How much of a benefit this will be is yet to be seen. Not only do managers need to utilize the additional talent they have at their disposal when playing in US Open Cup and CONCACAF Champions League matches, but the gain seen over the last season only gets the teams back to the roster numbers they had in 2008. That was the first year that the "extra matches" theory took effect in the first round of MLS playoffs.


MLS is certainly increasing its salary expenditures and consequently increasing the number of players on each team. This bodes well for the league that had reached a low in 2009 in terms of players and salaries in constant dollars. I am wondering how much longer the DP model can last in its current form - only two teams are able to spend the money to have anything beyond the token DP or two, and they're spending the money on top talent that is contributing this season. I just don't see MLS players putting up with such a pay disparity over the long term. Finally, I'll be interested to see how much the additional roster spots will help in the playoffs. I have made no attempt to hide my dislike for the whole concept of a playoff in this league, but the specific failure of successful teams to make it out of the first round due to apparent fatigue from participation in so many extra matches is especially frustrating for those of us who want to see long-term success rewarded. Additional roster spots may help with such fatigue, if managers are willing to take the risk of not playing their usual starting XI in knockout tournaments like the US Open Cup and CONCACAF Champions League. I remain skeptical, and will make a comparison between the first round opponents at the end of the regular season and make a few predictions.