Welcome to Lowered Expectations! Here we review some of the weekend’s best open-play shot attempts which did not quite live up to expectations. We break down the GIFs through the lens of our expected goal model and discuss each one evaluating the results, and also the process.
This week, we have a couple of shots by Houston, which gives us opportunity to discuss their overall team as they leave Week 7 undefeated. Columbus makes a surprise appearance lead by Pedro Santos. Philadelphia gives me an opportunity to rant about crosses. There is a Will Johnson imposter in Orlando, we talk about Big Chances in expected goal models and lastly we take on goals shot from crazy angles! LET’S GET TO IT!
5. Mauro Manotas, 78th minute, Houston Dynamo
Pass from: N/A
Keeper: Daniel Vega, San Jose Earthquakes
Expected goal value: .367
Shot Distance: 7 yards
Body Part: Head
Alberth Elis and Mauro Manotas are just criminal here. Anibal Godoy, what are you even doing? Elis just puts a pin-point pass nearly vertically into and through zone 14 for Manotas. Manotas in turn provides a simple set up with Harold Cummings unable to do anything about it.
Daniel Vega just can’t hold to this... and we’ll talk more about him and his inability to hold on to shots, giving Manotas an easy shot on goal and he just heads it a bit too hard.
I’m mostly #TeamHeadersAreHard. Trying to redirect a ball on frame is hard enough and to do it with a surface like you head that isn’t flat leads to unexpected results but this is about as easy as they come and it’s a bit wince inducing that he misses this.
Almost as wince inducing as Jungwirth who leaves Manotas behind and joins in the onslaught that isn’t going to catch or stop Elis before he launches a shot. Jungwirth is then left ball watching as the rebound goes right to Manotas for an easy put back.
Houston really missed out on some gimme chances this weekend but they still came away with a big three points and their defense is ... well, I cover that below. Did I ever mention that I write this column backward? I do. More Houston goodness to come.
4. Gyasi Zardes, 39th Minute, Columbus Crew
Pass from: N/A (Pedro Santos?)
Keeper: Evan Bush, Montreal Impact
Expected goal value: .382
Shot Distance: 9 yards
Body Part: Right Foot
While the pass and shot are kind of meh (both could be better, but aren’t necessarily bad in themselves), the things which stand out most in this GIF are the press and transition by Columbus and specifically Pedro Santos.
Looking more at the impressive pressure, Pedro Santos does work. He comes back to the ball and impeding Samuel Piette forward progress and dictating his dribble into space where nothing can really be done. Piette in turn makes a bad pass allowing Santos to immediately attack space and help provide a double team with Zardes.
Santos dispossesses the ball from Diallo here and Pipa is there with an easy one touch pass into space where Zardes gives it the weird and awkward dummy creating a quick 3-on-2 counter in the opposition box. The end isn’t great but the process still leads to a solid opportunity being created.
As I implied, the pass by Santos isn’t great. It leads to a shot which also is almost entirely reactionary by Zardes, who makes the best of a tight window and a lack of space. He puts a quick flick on the ball and barely pushes the shot wide. It’s a tougher chance than it looks and in turn means the value itself is probably lower than the model associates.
In general, Columbus has lacked an attacking punch early on and has mostly gotten by on their defense being very good. Obviously Zach Steffen is a difference maker in the defense with some big moments but the question should be raised as to whether or not the defense can continue to carry this team when Steffen is gone.
3. Alejandro Bedoya, 54th Minute, Philadelphia Union
Pass from: Cory Burke
Keeper: David Bingham
Expected goal value: .397
Shot Distance: 8 yards
Body Part: Right Foot
This cross is not great and the fact that they are rewarded with a high value shot because of the Galaxy’s complete lack of ability to clear the ball irritates me.
I may not like it, but this play was effective at creating a high leverage chance. Kai Wagner gets down to the end line and the looping pass crosses the goal face which creates some problems with positioning for Bingham and his defenders who are desperately trying to keep track of oncoming attackers.
However, the cross to Alejandro Bedoya isn’t good, he doesn’t have great positioning and it’s made all the worse by the fact that if Wagner picks up his head and cuts towards the box he has Brenden Aaronson at the near post with a good pocket of space underneath to create a better opportunity.
Aaronson has been so impressive since the start of the season and failing to get him the ball in a position like this where he can impact the game is a huge missed opportunity, the high value of Bedoya’s acrobatic twist and shoot notwithstanding.
I think you can see the shot opportunity created by Bedoya and recognize that because of how off-balanced he is coupled with his scoop-n-turn devalues the expected value of his positioning and the shot that comes as a result. That’s not to say the run into the box isn’t good just that the shot which comes about as a result isn’t likely as valuable as it looks at first glance.
2. Will Johnson (but not really), 11th Minute, Orlando City
Pass from: Ruan
Keeper: Nick Rimando
Expected goal value: .411
Shot Distance: 6 yards
Body Part: Head
I got two things on this:
First, while Will Johnson is credited with taking the shot, it’s not actually Will Johnson. It’s actually a young Will Johnson look-a-like named Joao Moutinho. The former-number one draft pick and left back for Orlando City, who was traded from LAFC this past off-season.
As my astute friend, podcast mate and GIF provider, Ian, pointed out to me, he does resemble a young Will Johnson. His short hair and arms in the air with a look of displeasure make him look enough like Will Johnson that it’s understandable that Opta would make this mistake. Maybe, check the jersey number next time though.
Second, this specific chance is awarded with the “Big Chance” qualifier. It’s been a minute since we discussed Big Chances so let’s talk about ‘em real quick using non-mathematical equations or jargon.
If you need or want a break down of how Big Chances work within a mathematical concept of expected goals, there are many posts out there which both postulate and discuss the concept and their incorporation within the respective models.
As for Big Chances, the simple takeaway is that they’re coded by Opta and defined as “situations where a player should reasonably be expected to score usually in a one-on-one scenario or from very close range.”
These are situations which aren’t determined by a model but rather through the observation of someone scoring a game. Situations such as this specific chance, with Joao Moutinho above, where someone felt the opportunity required a higher xG count than what the model gave.
The problem with using this information to help both provide further insight is we rely on someone interpreting a specific event the same way each time, which is unrealistic, and also agreeing with the interpretation which they make. The reality is that bias exists and it can influence the data set in new and challenging ways all their own.
It’s fair to say that we already know that the expected goal model is flawed. Heck, this column mostly exists to talk about some of its trappings. But the point of expected goals, aside from placing numerical value of quality on a specific shot, is the repeatability and predictive power of the metric both of which in the past have shown to decrease their our models value.
That’s not to say models which leverage Big Chances are bad. They aren’t. We made a decision which we think is best for us. But that’s not the end of the story. We’re constantly discussing these types of topics and challenging concepts as a group here at ASA.
These conversations aren’t specific to our group either. There are plenty of hot topics across the soccer analytic community right now and it’s probably not going away any time soon. Dan Altman mentioned on twitter plenty of topics which are both good for bar room and dinner parties discussions but also have dominate the type of ASA’s own slack channels.
A few debates in soccer/football analytics: 🤔— Dan Altman / smarterscout / North Yard Analytics (@NYAsports) April 11, 2019
1. Does a foul by a defender interrupt a possession? What about an unsuccessful touch?
2. After how many touches does a set piece become open play?
3. How do you calculate expected goals from two shots in the same sequence?
4. Can subjective measures of chance quality be part of expected goals models?— Dan Altman / smarterscout / North Yard Analytics (@NYAsports) April 11, 2019
5. How many touches must occur to begin a possession?
6. Are defenders always responsible for the same patch, even if they leave it?
7. How do penalties affect underlying result probabilities?
We made the decision to not use Big Chances in our expected goal model and this specific chance is just one example of why. The chance in and of itself has potential but the combination of Moutinho being double teamed and the requirement of him needing to redirect it on frame makes it such a difficult chance to convert.
I think we could have a good conversation concerning whether or not it’s deserving of an increase to its expected goal value. Perhaps as more data like defender tracking becomes available, we can start making adjustments to our model to account for this sort of thing.
1. Marlon Hairston, 8th minute Houston Dynamo
Pass from: Oscar Boniek Garcia
Keeper: Daniel Vega
Expected goal value: .667
Shot Distance: 3 yards
Body Part: Right Foot
Once again, Daniel Vega can’t hold on to a ball to save his life. The dude gives me grey hairs just watching the replays. Props to Oscar Boniek Garcia for jumping on the rebound created by Daniel Vega on the poor save. Those types of saves which roll back into traffic should be punished be and need to be at the very least turned into additional opportunities. But I’m growing tired of seeing these drab lazy saves where the ball comes free. It really can’t continue.
This whole GIF has a lot to unpack, not just in the lead up and creation of the shot for Mauro Manotas but also in the chaos when OBG gets to the rebound.
Starting with the entry pass from Tomas Martinez to Garcia, it wasn’t anything brilliant but it’s the kind of pass that instantly puts you in great situation in possession, improves your chances of creating a chance and puts the defense on it’s heels. I’m all kinds of shocked to watch Magnus Eriksson just let it happen. It’s super weird.
As soon as that pass happens you start to see pieces across the board spring into action and the man marking comes becomes more evident and almost highlighted. You see four orange jerseys start to converge towards goal with four white jersey’s trying to keep up.
When OBG receives the ball he immediately picks up his head and plays an easy square to zone 14 and into Manotas (again) who receives it. I think the lack of ball pressure is a result of the man marking by Florian Jungwirth. Jungwirth positions himself to intercept a bad pass but can’t completely vacate his area in case he doesn’t successfully win the ball when jumping the passing lane.
Elsewhere, Alberth Elis was death to Marcos Lopez through most of the game but Lopez does a pretty good job here keeping between him and the goal. However, as soon as Manotas gets on the ball, and with Jungwirth coming out of the 18 yard box to apply pressure, Elis tucks underneath the newly available real estate and posts up on Lopez. This gives OGB a path to jog into with space all his own.
Elis and OBG both have position on their marks even before the shot is taken and had Manotas passed it to Elis a series of combination passes could have been had to easily create a much higher leverage shot. But the lack of any control following the save by Vega gives the chance a second life.
The pure amount of scramble in this frame by San Jose you’d think it was at Denny’s. And don’t “at me”. Denny’s is the home of the Grand Slam Breakfast and I dig that. But Denny’s is for winners and San Jose weren’t that. They managed to push back a bit and even made the game more interesting than they deserved up until the 60th minute, but ultimately it wasn’t enough to trouble Houston.
Houston, if you didn’t realize, is undefeated. Oh, you heard? Well let me to continue to up the ante. They’re not just undefeated... they’re actually really good.
At this stage in the season expected goals is really only mildly predictive but their expected goal difference is the second highest in all of the Major League Soccer. But if you’re like, “Harrison that isn’t super useful until at least week 10”. Well, let’s hit up TSR which long term use is limited but has more predictive capabilities early on in the season. They do drop a bit further down the rankings at ninth, but still sit solidly at third place within the Western Conference.
I get being apprehension about their team. While their attack has been absolutely vicious at times, it was their defense which repeatedly let their team down in the closing moments of games last year despite strong team performances. But so far it’s been rock steady coming in first in all of MLS in home-adjusted expected goals against.
If the defense can continue to hold up they’re easily a playoff team and if they get into the playoffs, they have the potential to wreck some teams with their attack.
ONE TO GROW ON!
Yordy Renya, 53rd minute, Vancouver Whitecaps
Pass from: Unassisted
Keeper: David Ousted, Chicago Fire
Expected goal value: .002
Shot distance: 24 yards
Body Part: Right foot
While it was further from goal than Cristian Espinoza shot last week what makes this shot even more difficult is the angle. There have been multiple goals, just this year, scored from a -70 degree angle but the only shot with less available goal mouth scored was the sliding goal by Brandon Bye.
Going back to 2018, Jefferson Savarino scored a goal from a similar angle and was about 14 yards out. Maxi Urruti’s goal against Travis Worra in 2017 was an even crazier angle. Then in 2016 Diego Valeri had a goal against Seattle from a vicious angle from almost 18 yards out. But going back through the data I don’t see any goals from a -70 degree angle from over 20 yards out.
I did briefly explored the topic of goals from beyond the 18 yard box with tough angles last year while looking at Jimmy Medranda’s goalazo. While Medranda was a combination of finesse and power. Renya’s goal was about chaos and luck. It took a weird set of circumstances to occur enabling Renya to even take the shot, let alone strike gold.
My biggest question with this whole string of plays is what in the heck was David Ousted thinking here? The guy obviously over committed here and he realizes it somewhere between the balls second bounce and right before he dives for the ball. You almost feel the whole “record scratch, I bet you're wondering how I got here” meme, and believe me if there was a close up on his face as he slapped the ball to Renya, I would have made one.
I have to admit these types of shots interest me greatly. Not just because they’re low leverage but because they’re equally low volume. People just don’t frequently try to “bend it like Beckham” and while I get that, I do sometimes wonder if floating a shot at the back post wouldn’t be better than a cross on some occasions.
Top-10 individuals in Expected Goals from the last week (excludes PKs)