Were the Union the league’s unluckiest team in 2013?

American sports fans love their statistics, which could be one reason why soccer isn't as big as it could be in the US. Look at how baseball is presented to sports fans. During the October baseball playoffs, when a baseball relief pitcher entered a game, the television screen reported 7 numbers (with strange acronyms) that helped us understand the pitcher's performance over the last 8 months. And what about when soccer is on TV? Also In October, when Kleberson entered the game in the 79' against Toronto before kicking the goal of the season, all we saw on the TV was a green arrow and his jersey number, 19. There was nothing to define who Kleberson had been all year. The average viewer really can't have any context for his entry other than to know that when #19 does something, that's Kleberson. Such is the curse of being the beautiful game. Beauty is very often difficult to quantify. Our eyes seems to do a better job of defining a player's performance than any a collection of numbers.

But the soccer community and the MLS have been making progress defining the beautiful game with math. Three years ago the MLS introduced the Castrol Index as the official performance index of the MLS. The index, according to, tracks on average 1,800 player actions through the course of the game. "The Castrol Index tracks every move on the field and assesses whether it has a positive or negative impact on a team's ability to score or concede a goal." That sounds pretty innovative to me. But color me skeptical. One number defining a player? Goalkeepers and forwards alike? The final index for the 2013 regular season was released a few weeks ago. Let's see what the index can tell us about the Philadelphia Union season we just witnessed.

The Castrol Index analyzes performance at the player level. The first thing I did was calculate the average score of the team per minute played. The team with the best result per minute played was Sporting Kansas City with a score of 503 (out of 1,000). Ok, I can get my head around KC scoring well. The lowest scoring team was Chivas USA with 432. Again, so far so good. Next I ran a regression of each team's score to see if it was predictive of their total points in the final table. The results are graphed below.


My first observation is that the Rsquared is .763 (Rsquared in this case tells us the amount of the variance of points across the MLS table that can be described by the Castrol Index I summarized). I've read a fair amount of criticism of the Castrol Index but to think that one statistic can predict a team's points with that level of statistical certainty is impressive. To put that in context I'll turn to the NFL. I tried to predict a team's current wins by looking at their net points (points scored minus points allowed). The Rsquared for that is .786, just a little better than the Castrol Index. If that Castrol Index is on par with net points in the NFL as a predictor of success, that seems pretty strong to me. So let's look at the Union.

I've circled two data points on the chart. The red circle represents the team that earned more points than their Castrol Index would suggest they should. The red circle is Real Salt Lake. They earned 52 points this year but the Castrol Index and my regression indicate they should have only earned 43. The blue circle represents the team that had the biggest gap between points they should have earned and the points they actually earned. That blue circle team is our beloved Union. The Union earned a score of 492, the 2nd best score in the league. The model suggests the Union should have earned 58 league points when in fact they earned just 46.

So what does this say about the Union's season? Were they just extraordinarily unlucky to not have more points? Were they somehow unable to capitalize on all the positive actions the individual players were making? Does this reflect in some way on the coaching? If they were just unlucky we should see this trend reverse next season. If this trend continues there may be something systematically wrong with the way the Union play. It's something to keep an eye on for sure. And it's great to see soccer joining the world of innovative sports statistics. Now if we can just get them on TV.

In This FanPost


Trending Discussions