MLS 2020: Reviewing Week One of the MLS Season

I plan to do one of these posts for each week of the MLS Season. I’ll include the result of each game, expected goals and win probabilities, and then key stats or other thoughts where something stands out.


All Expected Goals amounts below are from MLS (which they source from Opta). I’ve then turned these into win probabilities. Opta calculates expected goals based on the type of shot, type of pass leading up to the shot, and where the shot is taken from.


Colorado 2 – DC United 1


Expected Goals: Colorado 1.63 – DC United 1.28

Colorado Win Probability: 46%

Draw Probability: 24%


Key Stat: Shots per Minute of Possession


I like using this stat because it helps to provide context when teams play in different styles; a team can counterattack and have lower possession numbers, but then they need to be more efficient in generating chances with the possession that they do have. In the case of this game, that stat is useful for another reason. Both teams had similar possession numbers (about 50 minutes to 45 in favor of Colorado), but Colorado generated .42 shots per minute compared to .27 for DC. DC’s new attackers will need to gel quickly and start turning their possession into more opportunities.


Other Thoughts:


Related to DC’s lack of efficiency in possession, Ben Olsen will need to figure out how to get spacing with his new forward line. Edison Flores primarily played in the center in his last appearances with Monarcas and on the left wing over the course of his career, but played on the right on Saturday. Conversely, Julian Gressel has more often played on the right in his career. On Saturday, this resulted in their heat maps overlapping as Flores came towards his preferred left foot and Gressel drifted out to the right. I’d like to see Olsen either flip these two or move Flores over to the left wing going forward.


Montreal 2 – New England 1


Expected Goals: Montreal 1.67- New England 1.47

Montreal Win Probability: 43%

Draw Probability: 24%


Key Stat: Saves


While not an advanced stat, one of the keys here was Clèment Diop making 5 saves compared to 1 for Matt Turner. Turner’s save was on a good chance, but he was also chipped on a play where he likely should not have come so far off his line. Each team scored on a corner in this match, which isn’t surprising since 16% of all MLS goals last year came from set pieces. The difference here was Diop making the saves on the other opportunities that New England created.


Houston 1 – LA Galaxy 1


Expected Goals: Houston .72 – Galaxy .74

Galaxy Win Probability: 32%

Draw Probability: 37%


Key Stat: Chicharito’s Touches


Chicharito had the fewest touches of anyone in either team’s starting 11. While it’s not surprising to see forwards with a lower number of touches, it’s especially concerning that very few of Chicharito’s came inside Houston’s box. Given how dependent his game has typically been on scoring inside the box, the Galaxy will need to find a way to get him the ball more in positions to present a danger.


San Jose 2 – Toronto 2


Expected Goals: San Jose .74 – Toronto 1.82

Toronto Win Probability: 63%

Draw Probability: 22%


Toronto was unlucky here to lose two points in the last minute. While San Jose dominated possession and the teams had similar shot totals, Toronto did a far better job of generating their shots from dangerous areas – 7 of their 10 shots came from inside the box.


Orlando City 0 – Real Salt Lake 0


Expected Goals: Orlando City .91 – Real Salt Lake .24

Orlando City Win Probability: 54%

Draw Probability: 39%


Very little to discuss in this game. Orlando was the better team overall and generated a number of crosses, but they were unable to convert. Real Salt Lake dominated the aerial duels, which was enough to prevent Orlando City from capitalizing.


FC Dallas 2 – Philadelphia 0


Expected Goals: Dallas .75 – Philadelphia .71

Dallas Win Probability: 33%

Draw Probability: 37%


Key Stats: Tackle Success and Clearances


This game was more even than the score suggests, but Dallas played well defensively. Tackles are clearly key to any defense, but missing them can put the team in bad position and open up quick attack opportunities. Dallas was successful in 25 of their 31 tackle attempts, compared to just 13 out of 23 for Philadelphia. Dallas also had 36 clearances compared to 17 for Philadelphia.


Atlanta 2 – Nashville 1


Expected Goals: Atlanta .24 – Nashville .82

Atlanta Win Probability: 10%

Draw Probability: 42%


As the win probability shows, Atlanta created very few chances but was able to convert and get the three points. The big story here is clearly the injury to Josef Martinez; Atlanta still has a number of skilled players, but it’s hard to imagine them overcoming the loss of Martinez.


Sporting KC 3 – Vancouver 1


Expected Goals: Vancouver 2.16 – Sporting KC .52

KC Win Probability: 8%

Draw Probability: 17%


This is an example of some stats being limited – in this case the expected goals metric based only on shots. Taking other actions into account (as FiveThirtyEight does) would result in a win probability more in line with what was actually seen.


Columbus 1 – NYCFC 0


Expected Goals: Columbus 1.74 – NYCFC .35

Columbus Win Probability: 72%

Draw Probability: 21%


Columbus dominated most of this game, though they should have done more with their opportunities. The Crew had three shots inside of the six-yard box, but their lone goal was not one of them.


NY Red Bulls 3 – Cincinnati 2


Expected Goals: Cincinnati 2.08 – Red Bulls 1.97

Red Bulls Win Probability: 38%

Draw Probability: 21%


Kyle Duncan was excellent for the Red Bulls, generating two shots and a goal in addition to strong defensive play.


Seattle 2 – Chicago 1


Expected Goals: Seattle 3.7 – Chicago .52

Seattle Win Probability: 92%

Draw Probability: 6%


Key Stat: Expected Goals


This is another example of the difference in different methods of calculating expected goals. Opta gives Seattle the large advantage, while FiveThirtyEight’s non-shot based method makes the gap only 2.0 – 1.0. This demonstrates why advanced stats can’t tell a whole story, but also can let you know where to look for answers: did Seattle get lucky by generating more shots than their play would imply, or does their skill and style of play generate more chances that the metrics would indicate. It will be interesting to track this over time to see if there’s a consistent gap between the two metrics when either Seattle or Chicago are playing.


LAFC 1 – Miami 0


Expected Goals: LAFC 2.75 – Miami 1.61

LAFC Win Probability: 62%

Draw Probability: 17%


Unsurprisingly, LAFC was the stronger team; while that would be expected, in this case that’s notable due to their strong Champions League performance in mid-week as well. Miami was impressive in their own right, and easily could have stolen a point without a strong game from Kenneth Vermeer.


Minnesota 3 – Portland 1


Expected Goals: Minnesota 2.48 – Portland 1.68

Minnesota Win Probability: 55%

Draw Probability: 19%


Portland was stronger in possession, but Minnesota was strong on the counter-attack in creating strong chances when they did have the ball.



#MLS #WeeklyReview #WinProbability

1 view0 comments

Recent Posts

See All

Revisiting Mikel Arteta's Statement on Win Probability

Mikel Arteta received a lot of attention in December for his comments about analytics and win probability, most of it focused on opinions that it was either a manager of a struggling team looking for