How Data Changes My View of MLS or a Frank Exploration of Luck in Dallas

by Harrison Crow (@harrison_crow)\

As I made my way to Toyota Stadium on Friday night I was concerned about just making it to the game on time. The traffic was horrendous and it was my first time driving around Frisco that collectively dragged my pace of getting to the park and was the reason I was walking up to the gate as fire works were set off and the National Anthem finished.

I stood just outside the south gate waiting for my ticket to arrive as Dominique Badji scored the Rapids first goal of the season and I felt a sense of validation in thinking that this was going to be a game that Colorado could be competitive and challenge for full points leaving Texas as I had implied with my post Friday morning.

It wasn't that I thought Dallas was a bad team as I wrote about them. I think Dallas is a very good team even after that beating, and I'm pretty certain they'll make the playoffs out of a very stacked and competitive Western Conference. The problem is that prior to Friday night they had the second highest PDO in Major League Soccer, a metric that is a measurement of luck based upon finishing and save percentage.

FC Dallas had scored a total of eight goals as a team behind the contributions of Blas Perez with three, Tesho Akindele just behind him with two and Fabian Castillo trailing with just one. Those three are what is going to drive the Dallas bus to success, just as the trio did last season, and though goals will come from other sources these are three that you can point to as "the guys".

The problem is that all three have been scoring goals with a much higher efficiency than what we'd seen previously from them. Now from what we've learned about scoring rates is pretty basic; they've had maniac highs and depressive lows. Even with the number of quality chances they're gotten, as described by our expected goal metric, it's not something that we could reasonably expect to continue. Again, not because they don't have fantastic goal scorers or that those players are of a lesser quality to the rest of the league. There are few players in the world that can score at their current rates.

Their high PDO meant that if the volume of Dallas' shots didn't change, they weren't going to continue to scoring goals.

Likewise, Colorado was riding a similar wave of eventual regression.

While Dallas had a high PDO, Colorado had a very low one (956, tied for second lowest in MLS) that was largely driven by their complete lack of goals across 48 shots. Yes, 48 shots without a goal. They should have, by our own measurements, scored four goals by the time they arrived to play FC Dallas this week and instead were sitting on a goose egg. Few teams can take near 50 shots over any given time frame during the season from the attacking third and come up that empty.

Just for a bit of applied science; over the past five seasons only two players have taken more than 48 shots and not scored a goal: Juninho for the Galaxy in 2014 and Kalif Alhassan for the Timbers in 2011. Based on shot leverage we can tell that Juninho was shooting from long distance and wasn't finding a lot of good chances. Likewise, being that he only scored five goals through 94 matches and is now playing in the NASL, it's possible Alhassan does not possess the finishing skills required from most goal scorers at the MLS level.

I'm not trying to say that it was certain Colorado was going to win a game or even score a goal.  Dallas could have very well done things different and there is luck to account for, too. Don't think for a second that shots like Dillon Serna's happen every week, there is a reason why it was special. The shot could have gone either wide or high and I'm a bit surprised that Walker Zimmerman didn't get a boot on it. Most players across soccer LEAGUES (not just MLS) convert those shots into goals in less than three percent of opportunities.

Colorado's eruption was mind blowing in the sense that I didn't expect them to score four goals, blow off the doors and leave Dallas with a clean sheet and all the points. But it's not as though I didn't think it couldn't happen either. That's the thing about all of this; we aren't trying to get a high definition picture of the future or to take the beauty out of anything, but instead it's to give the accomplishment context and measurement while understanding why it could have happened and if we should continue to expect it to happen.

Colorado isn't a team that I'm convinced is going to be anything great. They're probably at very best a 5th or 6th playoff seed if their defense holds up all the way through, which is another topic entirely. Likewise, Dallas has the attacking pieces to continue to beat their expected goal parameters and a top-3 seed isn't out of question. But if they either can't create more opportunities or continue to finish chances at a high rate, their regression may continue.