clock menu more-arrow no yes

Filed under:

What can expected goals tell us about RSL?

New, 2 comments

Attempts to accurately measure of the quality of individual performances in matches has lead to the creation of many new statistical measurements.

Joao Plata back at work John Engels

Expected Goals (xG)

Experienced soccer fans intuitively judge the relative quality of a shot on a number of factors. How close was the shot to the goal? Was the shot on goal or off goal? What kind of angle was between the location of the shot and the goal? Was the shot from a boot or a head? Was the shooter through one-on-one with the goalie or engaged with defenders? Was the shot blocked by a defender before reaching the goalie or goal?

About five years ago expected goals (abbreviated as xG) was introduced as an attempt to quantitatively measure the quality of each individual shot in a soccer match. Originally, the offensive side of the pitch was divided into six zones, and a numerical value assigned to each of the areas based on the percentage of shots for those locations that resulted in goals. Over time the distance of each shot from the center of the goal mouth measured in yards was factored in. But distance alone is not complete without the angle of the shot. So shot locations measured by angle off center was added, and then converted into the width of goal mouth available to the shooter. A straight on shot allows the player to "see" all eight yards of the goal width, whereas another at an extreme angle, leaves the player only a yard of goal to work with.

The result of these calculations is a single numerical value for each shot taken which can be added together for individual players xG or collectively for an entire teams xG. You can also use any team opponents xG as expected goals against xGA and calculate an expected goal differential xGD for a single match. Another frequent comparison is the actual goal differential GD minus the expected goal differential xGD (GD-xGD).

What are expected goals?

xGoals Explanation

Different xG Models

It is important to note that there is no single universally accepted value for expected goal events resulting in slightly different values depending upon the developer of the statistical model. OPTA, the official provider of MLS statistics, has a xG model which remains shrouded in mystery and the resulting values are not publicly available although MLS has begun to release some of this information in articles which include the 2016 season xG for each MLS team and the 2017 opening week xG for each MLS team.

American Soccer Analysis is a public internet blog which includes their own values for the 2016 season xG and the 2017 opening week xG as well as much more. This site also tracks an xA (expected assists) value and a xG value for goalkeepers based on completely different criteria, archives their stat values for all MLS seasons since 2011 and player salary releases since 2007. More details of their methodology and model is available here.

Finally it should also be acknowledged that there are many who discount the entire concept of xG as demonstrated by an article by Michael Bertin here.

What Does xG Say About RSL

While it is obvious that RSL cannot score either 1.33 (OPTA) or 1.78 (ASA) goals in a game, the xG does allow a quantitative measure of the quality of chances the team saw in this game. During 16 home games in 2016 RSL had a maximum xG of 3.23 for the August 27th match against Colorado which they won 2-1 and a minimum xG of 0.59 for the June 23rd match against the NY Red Bull which ended in a 3-3 tie. The average xG for the 17 home matches in 2016 was 1.53 so the effort this week against Toronto was slightly above 2016 home average as to quality shot attempts.

Last years game in Toronto which ended in a 1-0 loss saw RSL with a 1.10 xG and Toronto with a 1.21 xG. Again relatively even but with slightly less quality opportunities that the game this year.

Last years 3-1 victory against Chicago at home had RSL finish with an xG value of 1.85 and Chicago with a 1.23. During week 1 of 2017 Chicago tied 1-1 in Columbus and wound up with a 1.14 xG while Columbus had a 1.81.

Week 1 2017 xG Values

Team GP GF xGF (OPTA) xGF (ASA) GA xGA GD xGD GD-xGD
Team GP GF xGF (OPTA) xGF (ASA) GA xGA GD xGD GD-xGD
Portland 1 5 2.68 2.54 1 0.58 4 1.96 2.04
San Jose 1 1 1.50 1.67 0 0.1 1 1.57 -0.57
L.A. Galaxy 1 1 1.65 1.64 2 0.58 -1 1.06 -2.06
Houston 1 2 1.29 1.62 1 0.99 1 0.63 0.37
DC United 1 0 1.60 1.59 0 0.97 0 0.63 -0.63
New York City FC 1 0 1.18 1.10 1 0.48 -1 0.62 -1.62
Salt Lake 1 0 1.38 1.77 0 1.35 0 0.42 -0.42
New York Red Bull 1 1 0.68 1.36 1 1.02 0 0.34 -0.34
Columbus 1 1 1.81 1.44 1 1.14 0 0.3 -0.3
Colorado 1 1 0.95 0.77 0 0.52 1 0.25 0.75
Vancouver 1 0 0.59 0.80 0 0.57 0 0.24 -0.24
Philadelphia 1 0 0.37 0.57 0 0.8 0 -0.24 0.24
New England 1 0 0.37 0.52 1 0.77 -1 -0.25 -0.75
Chicago 1 1 0.76 1.14 1 1.44 0 -0.3 0.3
Atlanta United 1 1 1.13 1.02 1 1.36 0 -0.34 0.34
Toronto 1 0 1.35 1.35 0 1.77 0 -0.42 0.42
Orlando City 1 1 0.57 0.48 0 1.1 1 -0.62 1.62
Kansas City 1 0 0.72 0.97 0 1.59 0 -0.63 0.63
Seattle 1 1 0.71 0.99 2 1.62 -1 -0.63 -0.37
FC Dallas 1 2 0.45 0.58 1 1.64 1 -1.06 2.06
Montreal 1 0 0.13 0.10 1 1.67 -1 -1.57 0.57
Minnesota United 1 1 0.48 0.58 5 2.54 -4 -1.96 -2.04
OPTA and ASA xG value comparisons MLS, OPTA, American Soccer Analysis

Do you Like These Stats?

Please feel free to comment below if you would like to see more of these type of stats. Use data at American Soccer Analysis it is possible to break out xG for individual players (Plata had 1.04 of RSL’s 1.78 xG) and xGA for goalkeepers (Rimando’s 0.73 xGA was far below Bendik’s 1.25 and Melia’s 2.31 among keepers who kept clean sheets). What would you like to see?