MLS 2013: A Technical Evaluation of Strength of Schedule

Screen shot 2013-01-13 at 11.13.01 PMSteve Fenn puts MLS’s 2013 schedule into the data blender

MLS has released the 2013 schedule, which is once again not balanced, and with further expansion coming soon probably never will be.

Every club will play each cross-conference opponent once. All Western clubs will play everyone else in their conference 3 times, and the East will play 2 intraconference opponents twice and face the rest 3 times.

Here’s a graphic layout of the entire schedule, with hosts on the left and visitors on top:

......

…(click to enlarge)…

For (very cool) interactive graph, click here

Strength of Schedule

Given these imbalances, it is only natural to ask which teams gain advantage.

By linking opponent 2012 goal differential to every fixture, we can approximate strength of schedule pretty well. Across professional soccer leagues, goal differential has been proven to be a somewhat reliable predictor of the next season’s success or failure. Each individual match’s difficulty is tallied by dividing the opponent’s 2012 goal differential by 34 then adjusting for the average home field advantage in 2012 MLS (0.49 goals per game).

This strength of schedule figure is an approximation of how many goals each club should be favored by over the entire season based on their schedule alone. Like a betting line, negatives are good. Here is graph of 2012 goal differential vs 2013 strength of schedule:

...(click to enlarge)...

…(click to enlarge)…

For (very cool) interactive graph, click here

The West lines up perfectly with its best-fit line, the East is a bit jumbled.

Each Eastern club’s skew is driven by which two intraconference opponents they only face twice. Those above the line have it most difficult. Houston (orange) are the most hard done, facing Toronto and Philadelphia less than others, while Red Bull New York are the most fortunate, dodging 3rd matches against Sporting Kansas City and Chicago. Keep in mind that over the course of the season these differences in schedule strength are still quite small.

The worst-case scenario, though, would be the conferences becoming imbalanced. East and West were more-or-less equal in 2012, with goal differential of zero in interconference matches. However, when a season is unbalanced, that year’s Supporters Shield could deserve an asterisk, and the determination of the MLS Cup host might be questionable, too.

Monthly Schedule Strength

The month-to-month difficulty of this schedule should also be quite relevant to club expectations at various points of the season. To that end we have a graphical representation of the number of matches each club will play each month and the per-match difficulty of that month:

...(click to enlarge)...

…(click to enlarge)…

For (very cool) interactive graph, click here

The smallest squares are the easiest months, and the green months contain 3 or 4 matches, while the red have 5 or 6. March is clearly the easiest month in the schedules of Los Angeles, Vancouver, and Dallas. Don’t be surprised if these clubs’ records appear too good to be true on April Fools Day. If they struggle out of the gate, they may find it very difficult to catch up with the rest of the field in the West, though.

Additional Notes on the 2013 Schedule

Opportunity lost, again: MLS has yet to take full advantage of the slowest day of the sports year. Fans will have no other options on the day after the MLB All Star Game, July 17, but MLS’ only offering will be Toronto FC at Chivas USA on MLS Live. This is one weeknight that should feature a couple marquee matchups every year.

Steady match distribution: Of the 152 club-months, only 2.6% feature 6 matches, and 13.1% have only 3 matches. The US Open Cup and CONCACAF Champions League will congest some clubs fixtures, but MLS has at least set a mostly even pace.

Staggered season finales: I went over this last season, but it bears repeating. The last week of the season should be synchronized for competitive balance. At least one of the teams playing on Sunday after their competitors are done will unfairly get to rest starters heading into the playoffs when they should be sweating the results.

About these ads

23 responses to this post.

  1. To make this post more accessible, I decided to strip away some of the mathematics speak and caveats, but figure some will find them interesting.

    First the geekery:
    – The R^2 of GD vs SoS for the West alone is 1.00, meaning that GD drives it entirely.
    – The R^2 of GD vs SoS for the East is 0.8219 which is still very high.

    Now the caveats:
    – There will surely be clubs who perform decidedly better or worse in 2013 than their 2012 goal differential would predict. This Strength of Schedule (SoS) is a starting point, not an absolute.
    – It’s no surprise that SoS correlated so well to GD, because that is what I’m basing SoS on.
    – It might be that one home field has a bigger influence on outcomes than another.
    – Travel distance also varies in this schedule. At some point I hope to investigate the influence this factor has historically had on MLS results.

    I’d be happy to address any questions or criticisms here or on twitter @OptaHunt

    Reply

  2. Posted by john mosby on 2013/01/14 at 10:14 AM

    i do want to say though that whenever i am betting on a match, i want to be getting points, not giving them, just as a response to the thing about the betting lines.

    Reply

    • Right, but in a betting line it’s a compliment for a team to have a negative number next to their name. Essentially, SoS is saying how many goals the schedule is favoring to that club over the course of the season.

      Reply

  3. Posted by john mosby on 2013/01/14 at 10:16 AM

    so who has the easiest schedule, portland and DCU?

    Reply

  4. Posted by john mosby on 2013/01/14 at 10:18 AM

    maybe it is fc dallas

    Reply

    • Technically, San Jose has the easiest schedule, but that’s mainly because they don’t have to play against themselves. The real swings given by the schedule come in the distance between a club’s SoS and the best fit line.
      In the West, this distance is 0 for all 9 clubs. In the East, the (dis)advantages are tiny. Here are the extra goals Eastern clubs look to be gifted by the schedule alone over the course of the season:

      New York Red Bulls 0.69
      Sporting Kansas City 0.33
      New England Revolution 0.19
      Philadelphia Union 0.09
      Montreal Impact 0.06
      Columbus Crew 0.04
      Toronto FC 0.04
      D.C. United -0.08
      Chicago Fire -0.41
      Houston Dynamo -0.94

      This is driven entirely by the two opponents these clubs play only twice. Sorry I didn’t post that in the original article. That was an oversight.

      Reply

      • Oops, the formula that generated that list was backward. That was the right order from best to worst schedules, but the positives were negatives and vice versa. Corrected:

        New York Red Bulls -0.69
        Sporting Kansas City -0.33
        New England Revolution -0.19
        Philadelphia Union -0.09
        Montreal Impact -0.06
        Columbus Crew -0.04
        Toronto FC -0.04
        D.C. United 0.08
        Chicago Fire 0.41
        Houston Dynamo 0.94

        That’s what happens when I try to rush a calculations. Sorry for any confusion.

        Any more questions?

        Reply

        • Posted by john mosby on 2013/01/14 at 3:45 PM

          so basically, the eastern conference champs, based on schedule alone, are going to give one goal, over the course of the season, and red bulls are getting about 7/10s of a goal over the same season?

          Reply

          • That’s a pretty good summary. Season-long, the main takeaway should be that the unbalanced schedule itself won’t lead to a larger competitive imbalance unless one conference becomes clearly inferior to the other. At that point a close Supporters Shield race becomes sways disproportionately toward the best team in the bad conference. Similarly, the worse team might end up hosting MLS Cup
            Really, I think the shape of each club’s schedule in the month-to-month chart is the most telling part of this.

            Reply

  5. Hey Steve. Love the monthly breakdown. One clarification, just to make sure I’m clear. . the second chart (with the tightly correlated line) is comparing to a team’s own 2012 goal diff to their 2013 SoS (which is based on other teams’ 2012 goal diff), correct? In that case with any kind of nearly balanced schedule the difference between teams will be dominated by the fact that a team doesn’t have to play themselves. I did a similar thing for Sounder at Heart and it was clear that by far the largest factor in the Galaxy’s vs, say, Chivas schedule is that the Galaxy doesn’t have to face the Galaxy three times and Chivas doesn’t get to face Chivas three times.

    It’s an interesting fact because in retrospect it’s obvious but I don’t think it’s the sort of factor that people are expecting when they think of a strength of schedule comparison. . I suspect most people assume that schedule differences or home/away allocations are the major factors.

    Reply

    • Correct, and if you follow the link to the interactive graph you can see it even more clearly. When clicking a team name in the legend Tableau gives you the option to exclude that club from the graph. The really cool part is that the best fit line adjusts automatically.
      Exclude everyone in the East and it becomes clear that the West forms a perfect line. Exclude the West and you get a clearer illustration of the unearned swings in schedule strength for the Eastern clubs. Hence, the figures in my comment response to John Mosby above.

      The more I think about it, the more I wish I’d gone into more detail on this particular topic. Because when looking at season as a whole many fans’ main concern isn’t really strength of schedule, it’s “how fair is our schedule.”

      There are other possible factors that I didn’t address here. Clubs do travel different distances than other clubs travel to play at their stadium (though it’s mostly close). Also, some home fields are probably more advantageous than others. These would require more research than I had time for here, but I’m strongly considering an in-depth look at the effects of travel distances.

      Reply

      • Posted by simianvector on 2013/01/14 at 2:13 PM

        I think focusing on the monthly schedules helps there. Over the full season, SoS is basically dominated by the teams’ inability to play themselves, but when you break it down by month you can compare harder to easier stretches. RSL’s October, for example, is a nice piece of data to pull out.

        Reply

  6. [...] MLS 2013: A Technical Evaluation of Strength of Schedule – from The Shin Guardian: Given these imbalances, it is only natural to ask which teams gain advantage. [...]

    Reply

  7. [...] months the Timbers and other MLS sides will have the easiest and hardest go of things based on strength of schedule. Imagine how easy the schedule would be if the 2013 Timbers got to play the 2012 Timbers three [...]

    Reply

  8. Reblogged this on Soccer, A Love Story and commented:
    A schedule breakdown with such artistry to the science, I can’t help but love it.

    Reply

  9. [...] The Shin Guardian takes a stat-based look at the 2013 schedule. [...]

    Reply

  10. Elsewhere some readers recommended that I tweak the last graphic by making number of matches drive square size and SoS drive color. Here’s a link to the interactive result:
    http://public.tableausoftware.com/views/MLS2013difficultyandmatchespermonth/Sheet1?:embed=y
    I’m honestly not sure which one I prefer.

    Reply

    • Posted by john mosby on 2013/01/16 at 8:07 AM

      so really a big red box is the last thing you want to see, and green is good, meaning negative SoS, means weaker team, something like that? based on this new graph, id have to say that real salt lake have the most favorable schedule. seattle as well

      Reply

      • Depends on how you define “favorable.” They each have one of the easiest individual months, but in both cases that month’s countered by one of the hardest months elsewhere in the season. Overall, all West schedules are essentially fair, with the only edges coming from teams not playing themselves.

        To me, the monthly visualization is more about illustrating the general shape of each club’s season. Is it more favorable to start light (like LA, SKC, FCD, and VAN), or finish up against some patsies (like RSL or Montreal)? Hard to say, but it does cast a more realistic light on what those clubs should accomplish each month.

        Reply

        • Posted by john mosby on 2013/01/16 at 4:59 PM

          one thing that seems clear, is that it should be easier and better for a team playing three games in thirty days, as opposed to a team that is playing once every five or six days. rsl start the season off with a tough run, but they are one of only four teams with two months where they only play three games, and they do have four months that are in the green.

          i wonder if there are any early futures odds for MLS

          Reply

          • Yes, but in RSL’s case that’s countered by a 6-match August, and 2 months with 5 matches. Seattle has 4 months with 5 matches.
            If Salt Lake can make the playoffs (remember that their offseason changes were significant), they could well benefit from a relatively light schedule over the last 2 months of the regular season, though.

            Reply

  11. [...] MLS schedule the cycles to this analysis for the domestic league take more care. If you can recall previous columns on MLS strength of schedule, the calculations there came in handy and did the footwork for the [...]

    Reply

  12. [...] tougher fixtures, the Timbers achievement is much more impressive. Especially when you consider that schedule strength will more or less even out in the [...]

    Reply

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 247 other followers

%d bloggers like this: