clock menu more-arrow no yes

Filed under:

What the stats say about the Sounders’ prospects in 2020

New, 14 comments

Seattle Sounders’ Sisyphean Shield Search

Sounders vs. Timbers: Photos

MLS is a parity league; sustained excellence over the course of a full season requires both exceptional talent within the relatively tight constrains of league roster construction and a healthy dose of luck. This only goes to show the difficulty of the Seattle Sounders’ playoff run — making enough adjustments to rate well under competitive adversity (consider, for example, that the team had lost Clint Dempsey for the season and Chad Marshall to retirement in its two MLS Cup seasons). One doesn’t project any team to be the absolute best (or worst) with any great level of confidence — fivethirtyeight.com gives LAFC the highest probability of winning the Supporters’ Shield, at roughly 1 in 3.

Still, to make things interesting, as we look Seattle’s 2020 schedule, let’s focus on two very unlikely conclusions to the season: one in which Seattle earns 46 points or fewer (putting the playoffs very much in danger), and one in which the team matches or exceeds the points record at 72 (very likely earning the Shield in the process).

The Model

I’ve constructed a fairly crude model to simulate Seattle’s season using 2019 goals for and against for all teams on the 2020 schedule, split by performance at home and away. The result of each game is derived from a combination of Seattle’s typical goals scored and conceded (a randomly-selected value from a normal distribution built from the 2019 dataset) and the same calculation for each opponent. I then run the season 1,000 times. This “Monte Carlo” simulation step itself is reasonably similar to the projections with which you may be familiar at fivethirtyeight or sportsclubstats. Sportsclubstats, in contrast, weights by in-season record and a generalized home-field advantage (neither of which are particularly complex or insightful adjustments). Fivethirtyeight uses expected goals (good, if a cause for debate), a poisson model for goal-scoring distributions rather than my normal distributions (good), adjusts team “performance indices” with head-to-head results (great), and includes Transfermarkt player valuations in preseason performance index estimates (terrible).

In my simple model, each simulated season produces an end of the year point total. The count (y-axis) of each point total (x-axis) is shown below for two scenarios. Scenario 1 uses Seattle’s unmodified goal distribution from 2019, which produced a goal differential of +3. Scenario 2 uses the Sounders performance in regular season games for which at least 3 of 4 key players were available to start (Raul Ruidiaz, Nicolas Lodeiro, Gustav Svensson, Jordan Morris), which would project to a +14 GD over the course of a full season.

Scenario 1 exhibits a peak at 50 points or just below — more likely than not to be a playoff season, but only just so. A +3 goal differential is a borderline playoff performance, and the 2019 Sounders can be considered somewhat fortunate to have done so well with so little (or comparably unlucky to be forced into so much within-season lineup turnover to prompt that performance). Scenario 2 averages 56 points. Out of 1000 simulated seasons, Seattle achieves 72 points or more in 15 (1.5%), and 46 points or less in 102 (10.2%, with a decent chance of missing the playoffs). Scenarios 1 and 2 essentially straddle fivethirtyeight’s preseason projection for Seattle of an average simulated season of 52 points, a 69% chance of making the playoffs, and a 4% chance at the Shield.

Fivethirtyeight makes no adjustment for key player injuries and absences as I have done, implicitly, with the four above (it would be difficult to make any such adjustment essentially objective or evenly-applied throughout the league). Given “average” (rather than 2019’s “dire”) player availability, but also given Seattle’s participation in multiple competitions, Scenario 2 looks like a fairly plausible model for the season.

56 points, playoffs and a likely 2nd seed from the west. A small chance of disaster and an even smaller chance of setting a regular-season record or claiming the Shield.

But, if either of those unlikely events were to happen, what would they look like, and what shall we look for during the season to tell us where Seattle stands?

The median points are plotted for the 15 simulated seasons in which Seattle meets or sets the the point record (“record”), for the 1,000 seasons averaging to 56 (“full”), and for the 102 seasons reaching only 46 points or less (“playoff”). March through May is the relatively friendly portion of the schedule for Seattle — all three modeled paths start out relatively well, taking advantage of the Sounders’ six home games from their first nine.

Although May brings the first of 2020’s three-game road trips, the end of the month welcomes relatively soft opponents to CenturyLink in Montreal Impact and Real Salt Lake. The six road games in eight from June 13 to July 18 are the key test of the schedule (particularly when one considers the likelihood of losing players for Euro 2020 and Copa America). Even the two home games in this stretch present a significant challenge, facing NYCFC and Minnesota United. The Sounders earn 16 points from these eight games in the record pace, 12 points on average, and eight points on the playoff margin path.

After this dire summer stretch, Seattle plays 7 of its last 12 games at home, including key games for the Shield (LAFC home and away), for playoff positioning (Portland and LAG at home, Minnesota away), or for playoff contention (SKC and SJ at home). We should know — on July 18 in Toronto — which of those regular season goals is most realistic. After July 18, the record pace is 46 points (2.09 ppg), the middle road is 36 (1.64), and the bubble is 27 (1.23).

A Statistician’s 5 Home Games to Watch

May 31 vs. Salt Lake

A must-win-game for Seattle’s more ambitious scenarios, this is the last game preceding the doom stretch of June and July.

July 5 vs. Minnesota

Minnesota receives a great deal of dark horse noise, having improved so much from 2018-19 (largely on the presence of Ike Opara at CB and Romain Metanire at RB). This is Ozzie Alonso’s 2020 visit to Seattle, and 1 of 2 key home games in the summer doom stretch. The Sounders need this win.

July 25 vs. LAFC

LAFC’s record-breaking 2019 sets the standard for Shield ambitions in 2020.

Aug. 8 vs. Kansas City

Seattle lost both of its 2019 matches with SKC, conceding three goals in each. The two 2019 home losses — SKC and Portland, once again visit Seattle in the summer in 2 of 3 games (bracketing a visit to Minnesota, and following the back-to-back home-and-away with LAFC). One could easily mention the Aug. 22 Portland game here (or have it replace the fixture below). A team with Shield ambitions in 2020 must aim nearly every home match. This stretch of five games from July 25 to Aug. 22 against key competition will do much to decide where Seattle stands in the conference at the end of the year.

Oct. 8 vs. Nashville

For this exercise, I gave Inter Miami a league-average statline and I gave Vancouver’s 2019 performance to Nashville. National punditry has not generally been kind to Nashville’s chances, but has praised its MLS veteran defense (personally, I don’t rate it so high — Dax McCarty is a solid possession midfielder but overrated on defense, in my opinion. Anibal Godoy’s strength has never been his range, and he is 30 years old coming off a muscle injury that held him out of a significant amount of 2019). In any of the scenarios I’ve described, expect there to be a good chance of the season finale meaning something to the team, and expect Seattle to need a win.