News reports suggest Everton are on the verge of acquriring Juventus starlet Moise Kean this week. What can the data tell us about him?
Bare basics tell us he still won’t be 20 years old until next February. He’s played less than 1700 minutes in Serie A during his short career between parent club Juventus and Verona where he went on loan in 2017/18. That’s less than 19 full games to you and me. Not much to go on but he’s packed a fair whack in already.
Everton fans want a goal scorer. Well he’s that. Roughly 9 xG with 10 goals to actually show for it all told. He shoots in decent volume in decent positions. Here’s his shot map for last season at Juventus:
A quick look at his radar shows you that beyond that shooting profile he carries the ball pretty well to beat opponents but beyond that we’re struggling. Moise isn’t one for setting up colleagues, can’t head it, and like Cristiano Ronaldo, left all that boring final third defensive work up to Mario Mandzukic:
Quite how that sits into Everton’s current forward line is a difficult one. Richarlison, Gylfi Sigurdsson and Dominic Calvert-Lewin all put shifts in when they pull on the blue shirt. The Goodison Park crowd won’t stand for anything else these days and manager Marco Silva’s set-up requires it too.
You also have to consider that Everton’s favourite way of creating chances under Silva is the humble cross – mostly launched from the boots of Lucas Digne and Sigurdsson. Eyes back up to Kean’s shot map. Yep, not one header in there.
It’s also interesting to note that Kean only started getting real game time in March, netted 6 times from then on, but Juventus’s team xG did this:
But you want positives, don’t you? I reckon the eyes will give you more of those. Go and have a look at some highlight reels. Kean looks great receiving the ball on the deck with his back to goal. He gets on the half turn very quickly if defenders get too tight. He has two good feet, good close control and good acceleration to get away. His running style with the ball puts me in mind of a Samuel Eto’o/George Weah hybrid. I’ll watch that all day long.
Back in 2012, spurred on by my wife and a friend who out of boredom at work deliberately sparked daily debate with his outlandish football takes, I started blogging about football and using data to answer the questions it poses. Fast forward six years and what started as a mess about has become seriously surreal.
Twitter gets its fair share of criticism, but it’s enabled me personally to reach crazy great people. At first it was all about volume of traffic and gaining new followers. Soon, it became about more than that. I was invited back to my home town to visit Bolton Wanderers’ Head of Analytical Development, invited by the Scouting and Recruitment Co-ordinator of my beloved Everton to speak with him at Finch Farm. It’s still bonkers to me that people in the professional game speak to me, read some of the words I’ve written and listened to some of the words I speak. Following my StatsBomb come back in August the Swedish national team goalkeeping coach messaged me wanting to chat.
There’s a feeling that the new generation of football club analysts and coaches, routinely more formally educated than their predecessors, are embracing new ways of communicating their ideas about the game. Maths Elfvendal is definitely one of these guys.
“Wherever you’re working you need to be good in a lot of areas to be a good coach these days," says Elfvendal, “but especially the bigger the club or the bigger the national side. It’s getting complex. You need to handle the social relationships. With the national team you have one, two, three goalkeepers who think they should be playing. You need to handle them with care – both individually and as a group.”
Elfvendal has been a semi-professional footballer in Sweden. His father was goalkeeping coach for the Swedish U21 team, and coached in the Swedish top flight while Maths grew up looking on. Now the boy is grown up and goalkeeping coach for both IFK Norrköping and the national team. Elfvendal is still only 31 years old but has somehow also managed to fit in 5 and a half years of university - learning to be a secondary teacher in social studies. He sees his educational experience every bit as important as his footballing background: “I adapt myself to the goalkeeper I have. I speak to every goalkeeper in a different way. In one way I have those years learning how to teach. In another way I have years of experience in the dressing room and hear what footballers say. I need to adapt.”
The concept of adaptation comes up repeatedly in our two hour chat. Elfvendal believes the new wave of technology and information available now is changing the game rapidly. It’s a matter of keeping up and developing your own philosophy or being left behind. “I’ve been reading your blog and it’s helped me a lot to build my own philosophy, both from statistics and also experiences with goalkeeper coaches and players,” says Elfvendal casually. I nearly choke on my cup of tea. “A lot of football experience can be short-cut these days. With all the information we have you can gain the necessary experience and knowledge of the game much faster now.”
I have to admit, despite being flattered by Elfvendal’s comments I’m still a ‘football philosophy’ sceptic. If you have three keepers to look after with the national team and they’re all different how do you approach a training session they’re all involved in? “That’s interesting. They are different, definitely.” he says. “Physically, technically. They also play in different leagues. On international duty the standard is only three training sessions before a game. For me, to believe I can change them technically in that space of time and for them to then perform at their best would be naive. With the national team my training sessions are based on tactical aspects for the next game, for the next opponent. In that sense it’s easier. How do we want to build up our play, how do we need to protect the area from crosses or from cutbacks. These positional changes we can change a little bit. It’s more based on the game plan so the whole team is on the same page. At a club there’s more time to work on technical detail. Also, you have to respect that the clubs own the players, we just borrow them for the national team. I don’t want to change anything that will negatively impact them when they go back to their clubs. I don’t want to say something when the coach at their club says otherwise.”
The teacher is eager to test me using some of the presentation slides he uses when delivering lectures to other coaches. “What position do you prefer the goalkeeper to take up?” he says flicking up a still of an attacking situation. The goalkeeper has been removed from the image. “See the positions marked there? One, two, three, four, five. I want you to tell me in 5 seconds what position you want the goalkeeper in.”
We go through a couple more before this bomb is dropped on me:
I take longer than 5 seconds. “Er, five? I feel worried now, Maths.” He laughs at me.
I start bleating: “It depends on your goalkeeper and what you want the team to do I guess.”
“Ah,” he says switching to the next slide. “It’s in Swedish there but it says ‘What information do you need to have to answer the question of what the best position is?’.”
“Are you asking me?”
“Yeah, you taught me one of them so…”
This statement doesn’t help. I start rambling about what your coaching style is (are you leading as coach or are you allowing the player to lead you). Is the keeper good at moving? Can he move his feet?
“Ok, that’s two. Team tactics and goalkeeper style. Four more.”
“Four?!”
"You don’t need to answer, it’s just a fun game.” He sounds disappointed. I try and up my game.
“Ok, so I want to know where my defenders are. Do I want to be aggressive and hit the opposition on the break if I can quickly gather it?”
“That’s still tactics. I don’t have all the answers here…” My brain has gone.
“Tell me what else?”
“Expected goals! What is the expected goal value of a shot from here?”
“Next to nothing."
“Right so what is the expected goal value if he is assisting one of the forwards? He can cross it here. There’s a defensive line four versus two. As a keeper you can either be aggressive and come and get it, say at position 2 or position 4 on the picture.”
I remark that the Premier League I see more and more that keepers are taking up position 7 or even a 9 as if they’re terrified of being beaten on the near post.
“Yeah, you have a big problem there if the header comes.” Elfvendal modifies the picture an adds an extra attacker: “How does your position change now?”
“Possibly between 5 and 7?”
“This is really interesting”, he says. “Do you think if the ball is played into him where the attacker has moved to now, do you have enough reaction time standing on his line?”
“Yes.”
“Yep. So if the ball is played in front of him to five and a half metres…ah…to the six yard line, I am talking to an Englishman now. Do you think you have reaction time?”
Elfvendal is well aware of my comfort zone preferences. He’s fishing. Images of David De Gea saves flash through my mind. I take the bait: “Yes.”
“I think you overestimate a little bit,” he laughs. “From my point of view you will have difficulty reacting in a good way from there. A guy called Scott Peterson is doing some science about this. What’s the likelihood of saving the ball depending on the distance between the ball and the goalkeeper? Imagine you are in position 8. You will be one to two metres away from the striker, less reaction time but close enough to block the sight of the whole goal with your body. Back on your line you have more time and reach but are not covering so much goal. Also Scott’s research shows that the conversion rate for a goal when the distance between ball and goalkeeper is 2-7m is significantly higher than any other distance.”
I start protesting about sample size, the possibilities and permutations of what could happen as things stand in the still image. Anything could happen from here. Is Elfvendal teaching physical cues? He flicks another image on screen. “I showed this to one of the participants on my course,” he says.
Yeah, by the way, Elfvendal is a course instructor on the UEFA goalkeeping A License. “We’re just looking at the biggest threat,” says the Swede. “From my point of view I’m trying to prevent a high value xG chance here. From position 2 or position 4 I can gather a cross to stop the header every time. I can also move to position 8 quickly if there is a striker coming to the near post like the second picture.” He shows me several video clips of keepers doing just that. Then he shows me another still. Pretty much the same as the first two but the man with the ball is just wider, out near the touchline about 25 yards out.
“What if the keeper is super aggressive? What if he is out like near the penalty spot leaving his goal open? How does the xG change?”
I’m still in stupid mode: “There isn’t enough sample to model it.”
“I was expecting a deeper answer here,” he says sounding disappointed again. He perks up immediately and laughs. “Use your imagination! You can be aggressive here. I showed the goalkeepers, I put them in the position on the ball near the touchline that far out. I stood myself near the penalty spot. Try and score against me guys. Left foot from there, score on me now. They didn’t score a goal I can tell you that. It’s harder than people think. By being higher, you are not risking too much being beaten from there, but you’re helping the team by protecting a higher xG value chance being made from the cross.”
And it dawns on me that right here is the value of theory meeting a real life practitioner. A practitioner influencing real outcomes, taking it to the next level with a logical step. I’ve been looking at this stuff for years, I’ve never even thought of it this way and this guy is here crediting me for the inspiration. “I see all kinds of rubbish on the internet about evidence based coaching,” I say. “This is the real deal.” “Yeah,” says Elfvendal, satisfied.
After two hours, I think the coach is finally happy with me. He’s used every trick in the teacher’s book – serious voice, disappointed voice, gentle mocking, laughter and praise to bring it out of me.
The first professional women’s match I attended in person was the 2010/11 Champions League Final between Lyon and Turbine Potsdam.
A few rows in front sat Trevor Brooking and Michel Platini.
Yeah, I didn’t slum it for my first game.
It was a great match with the technical French side winning 2-0 against the more physical Germans. Three things stood out. Firstly, Potsdam No 10 Fatmire Bajramaj, still one of the best technical players I’ve ever seen in the flesh. Secondly, Lyon substitute Lara Dickenmann’s half-time warm up that would’ve given a Ronaldinho trick show a run for its money. And lastly, but not erm, leastly, Sarah Bouhaddi, Lyon’s goalkeeper.
Man, was she good.
She effortlessly plucked every cross out of the air and was comfortable with the ball at her feet. Bouhaddi made a Champions League Final look like a training session. The last professional women’s match I attended in person was last week’s FA Women’s Super League (FAWSL) game between West Ham and Chelsea. The Irons’ keeper Becky Spencer was given player of the match and certainly deserved it having made three great saves in the first half alone. Watch them here.
It’s all a little bit odd as every time I speak to any male friends (who barely watch women’s football) they immediately raise the state of the goalkeeping as being terrible…
The guys at Statsbomb have started collecting data on the FAWSL and the National Women’s Soccer League in the United States this year. Now, the data set is still waaaaay too small to conclude anything but imagine my delight on discovering that so far, when you lay a shooting model from the professional men’s’ game over the women’s data, the goalkeepers in women’s football are performing better than their male counterparts. They’re breaking the model.
As the data gets gathered I felt it was a nice time to get some views from within the game itself. Andy Elleray is goalkeeping coach at Birmingham City Women and is England Women’s Youth International goalkeeping coach too. Do his coaching sessions for women look different to ones he coaches for men? What differences must be accounted for?
“The sessions don’t look amazingly different to be honest,” says Elleray. “It’s all based around the keepers you’re working with and the way the club wants them to play. In terms of attributes, female keepers are not as powerful so lots of work is done on jumping mechanics, speed to cover the goal and defend the area. Distribution is a big area for us and the goalkeepers must be able to play with variation - especially in our half of the pitch. The length that females can strike a ball is more often than not shorter than males so we look for our goalkeepers to play what we call PACE (positively, accurately, consistently and effectively) rather than rely on hitting areas and turning the opposition defence."
The data collected so far backs these statements up. Only Sophie Baggaley at Bristol City and Megan Walsh at Yeovil relentlessly go long. Here’s what Birmingham keeper Ann-Katrin Berger’s distribution looks like:
“With the role of the goalkeeper ever evolving I can see more statistics based around distribution,” says Elleray. “I have developed my own distribution analysis that I use at various points to assess how well we are retaining the ball or penetrating the opposition.”
Lloyd Yaxley, goalkeeping coach at NWSL club Orlando Pride is in agreement. “I think there will definitely be statistical differences between the men's and women's data. With distribution, the male goalkeepers’ range of passing is for the most part significantly longer than in the women's game. I think if you were to take something as simple as a long goal kick and compare the average length of a Major League Soccer goalkeeper to a NWSL goalkeeper there would probably be at least a 15-20 yard difference.”
Despite this, Yaxley states that like Elleray, his sessions for men and women don’t look that different. “I try and keep my philosophies towards the role of the goalkeeper as set as I can and how I want my goalkeepers to perform wouldn't change all that much. I would want them to be aggressive in certain situations and then in other areas less aggressive and maybe play deeper in their goal. With shot stopping I like my goalkeepers to be deeper so to allow for more reaction time - especially when there are a lot of bodies in front - this I would keep the same for males and females.”
Yaxley thinks having the ability to parry into safe areas is a significant attribute to have and is working to bring this to the women’s game as standards continually rise. “Men are used to the ball being struck at greater pace and therefore with greater movement too. We work on this with the Pride goalkeepers - especially if it’s a reaction exercise or I connect with a strike really well then we look at manipulating the body to best control the rebound. For example, when the ball is struck at pace and straight at you, rotating wrists and leaning with the shoulders late forces the ball into a wide area. If the lean of the shoulders comes too early, the control of where the ball ends up is often lost.”
How much does data influence training at present?
“I don’t really use data on a daily basis, but the higher you go at international level there is lots of work done on goalkeeper specific stats throughout the teams to analyse, review and compare keepers,” says Elleray. “I have written lots of research on goalkeepers from various view points and am a massive advocate of using video analysis throughout the coaching process. Recently there has been clubs using GPS data to measure goalkeeper training load and different types of movements and this is an area that I will be moving more towards. At any turn statistics should be used to aid the process not take it over."
Yaxley uses numbers in a different way.
“The main data I analyse is where the opposition is shooting from and scoring from and on the flip side, where we are conceding shots from and conceding goal from. We look at opposition trends and their style of play, danger players’ traits - do they like to cut in from the left and shoot with their right, are they getting to the line and cutting it back, for example. That will all help shape my focus for the weeks training.”
The level of detail StatsBomb collects at goalkeeping level now would aid such a process, and enable quicker analysis of your own players too. You’ve all seen the shot maps on site by now but there’s some cool new stuff deeper under the hood:
We’ve never been able to dive down and get dirty to this degree until now. As one keeping coach I follow on twitter is fond of saying…I’m #alloverit Header image courtesy of the Press Association
I’ve spent an hour or two (or three, or four) on the beaches along the Dorset coast over the last few years. And I reckon Alum Chine at Bournemouth is my favorite.
Which is odd because it’s almost smack bang in the middle of my personal football hell.
A few miles down the coast lies Sandbanks, home of Harry Redknapp, maybe one of my least favorite managers of all time. The game’s Donald Trump equivalent, a man who once said:
“Whether it is 4-4-2, 4-2-3-1, 4-3-3, the numbers game is not the beautiful game in my opinion.”
Before uttering in his next sentence: “It's 10% about the formation and 90% about the players.”
A few miles in the opposite direction stands the Vitality Stadium, the soccer home of Eddie Howe, a manager whose name I’ve physically winced at every time someone mentions it being on the list for the latest Everton boss vacancy.
Unlike Harry, when Howe opens his gob, nonsense doesn’t pour forth. He comes across really well and speaks intelligently. I just don’t like the way he sets up his football team. Eddie Howe couldn’t control a game to save his life.
As a neutral, watching Bournemouth in the Premier League has been a great experience. The Cherries play with great energy both in and out of possession and their games are end to end. Unfortunately for Bournemouth, it’s mostly been at their end.
The expected goals numbers against Howe’s side have been fairly horrendous every season in the Premier League. To the extent that a certain James Yorke of this parish perennially picks them to get relegated. I’ve never been that down on them myself, but I’d never wanted Howe anywhere near my beloved Blues either.
But something changed.
First up, let’s have a look at the shot maps from last season after 10 games (left) and this season after 10 games (right):
Okay, so they’ve had an easier set of opening fixtures this season but the reduction in expected goals against is pretty crazy. And okay, they still concede an above average amount of shots per game, but look at the reduction in shot quality against per shot: 0.104 last season, 0.073 this. That’s enormous.
Not only that, but looking at the team’s rolling expected goal trends at both ends you see that they’re doing better in attack too:
Numbers wise, it’s clear to see why Bournemouth were 19th in the table this time last year and they’re 6th in the table now. But what about what’s happened on the pitch?
First up, with Callum Wilson fully available this season, Howe has committed to having two center forwards on the pitch as much as possible. Bournemouth have regularly played in a fairly traditional 4-4-2 set-up. It’s a bit too simple to say attackers attack, defenders defend and midfielders, er, midfield but you get the gist.
Fullbacks are more and more important in the modern game and Bournemouth’s are no exception. They’re charged with getting forward a lot. How their forays are covered seems to have changed somewhat.
Last season, the side was terrible at preventing the opposition getting into the box in open play. This was particularly the case on the left hand side to cover Charlie Daniels bombing forward. Andrew Surman got the easy job. However, Dan Gosling’s defensive pressure map was nuts considering he played as a center mid for pretty much all his minutes:
Partnering with Jefferson Lerma this season, Gosling has kept shape far more, no longer having all the responsibility to cover Daniels:
Bournemouth’s wide midfielders Ryan Fraser and David Brooks are getting lots of praise for their attacking prowess this season. When I’ve watched them, my pragmatism takes hold and I love seeing them tuck in, making a decent show of protecting their fullbacks. They’re not even that great at it, but just that extra presence makes a difference. Flanks are protected.
The middle is protected. Fraser now mostly on the left, performs a nice solid block in the middle third. A bit less aggressive further up field than last season, he drops off more into a conservative position:
On the other side, Brooks provides a more consistent steady approach to defending his flank than the likes of Ibe gave last season. With the steadier central pairing in place, Brooks isn’t having to drift inside as much to cover the middle either:
Not only do Bournemouth have a more straightforward organisational structure off the ball now, they also have the most fun central striking pairing in the Premier League. In Joshua King and Wilson, the Cherries have the icing on the cake. Two willing workers and runners to harry opposition defenders who dally on the ball. Wilson is off the scale:
With the rest of the team in a more orthodox and conservative shape, any quick transitions when opponents moves break down are ‘dip yer bread’ territory for the pacey front two, often with half a pitch to run into. It’s a sight to behold. You should definitely make room for more Bournemouth in your life, despite the oncoming winter chill.
‘Tiki-taka’ died for the umpteenth time in Russia this summer. What with Spain and Germany’s possession game going home earlier than usual from the World Cup, there could be nothing else to conclude. Both these nations had been found out.
Rio Ferdinand said: "Spain have been so successful with that style of possession football but there comes a time when you have to get the ball into the strikers. You have to change it up a bit. Spain got what they deserved.”
Gareth Southgate said: "It has been unusual to see them (Germany) struggle as much as they have but the level of all of the teams is strong and they have played teams who have been tactically very good against them.”
Since then, in the UEFA Nations League, Spain have beaten Southgate’s England and whooped World Cup finalists Croatia 6-0. Germany have drawn with World Champions, France, and beaten Peru.
What. A. Crisis.
The problem is that most of football - the professionals in it, and the fans watching it, still really struggle to look beyond the score line to back up their takes.
In terms of shots, expected goals and creating chances, Germany battered everyone they played in the World Cup:
Germany dominated the ball too, completing three times as many open play passes as its opponents. It wasn’t all sideways and backwards either.
Germany proportionately played more balls vertically into the opposition box from the middle of the final third than any other team. No other side crossed the ball into the box more than Germany either.Maybe as Gareth said, the level of all teams is strong now and all the teams Die Mannschaft played were tactically very good against them. The only numbers that back that up are the score lines. They’re the only numbers that matter! scream the real football guys. As usual, here at Statsbomb, we beat the drum that the underlying numbers count for more in the long term and give a better reading of performance now and in future. The score lines since back that up.
Spain didn’t actually lose a game in normal time during the World Cup. This is how the game Spain went out to versus Russia looked on the shot map:
That big red square there on Russia’s map was a penalty. After Artem Dzyuba scored that spot kick on 40 minutes, Spain proceeded to have 24 shots to Russia’s 3. They also passed them to death for the entire game.
Spain only allowed seven shots on target during its four game tournament but David De Gea conceded goals on six of them. Two were penalties, one was a peach of a free kick from Ronaldo and one he fumbled in from outside of the box.
This was not a team in crisis (despite the loss of its manager on the eve of the tournament), it was a team whose path crossed with misfortune.
Maybe it’s like Rio said. Spain had to change it up and get that ball into the strikers more. The only numbers that back that up are the score lines. They’re the only numbers that matter! scream the real football guys. As usual, here at Statsbomb we beat the drum that the underlying numbers count for more in the long term and give a better handle of performance now, and are a better predictor of it for the future.
We repeat that message. A lot.
Spain only did enough to get one or two goals versus Croatia last month according to expected goals. However, Luis Enrique’s men still totally controlled the ball with 70% possession. Keep dominating games as Spain do, and the cliff you occasionally fall off will, at the bottom of it, have a dinghy in the water to handily catch you before you get wet. The rain in Spain falls mainly on the opponent’s goalkeeper.
A quick search of Google throws up all manner of newspaper articles looking at the secret of Croatia’s success both during and after the country’s run to the final. Much of it surrounded the wealth of character the people possess after what the nation’s been through over the years. I’m #justsaying the underlying numbers had them as a side dominating possession and shot counts.
Presumably, a 6-0 drubbing precipitates calls for a change of style and play from our English soccer (yes, soccer) pundits. A quick Google search throws up nothing. It wasn’t a World Cup game after all, and it’s only Croatia. They’ve never won anything anyway so we’re not desperate to find a chink in their Balkan armor.
Back in the Premier League, our pundits currently wax lyrical about the teams at the top of the table: Manchester City, Chelsea and Liverpool – the teams that, funnily enough, dominate the ball and most importantly, the shot count. But the pundits will never frame it that way.
Remember Gary Neville questioning Guardiola (who’s won loads) and City last season? “Every single team that’s won the league, barring none, has had power and strength at the heart of them - that spine. I just wonder whether they can play that way. That’s the fascinating thing over the next 12 months. Can you play that way, with those players and win this league? That will be the real test.”
Forget all the other nonsense, and get on board. City absolutely smashed everyone for shots and in the expected goals table last season. It’s the actual real test and measures the strength, heart and spine of any good side.
At 19 years old, Thibaut Courtois was Diego Simeone’s first choice keeper at Atletico Madrid. After three full seasons in La Liga he was being touted as the best young goalkeeper in the world. Four seasons on, people were more likely to tout David De Gea, the man Courtois replaced at Atletico, as the best keeper in the world. Then, at the World Cup, over the span of a few games, De Gea has a nightmare, Courtois did well, won the Golden Glove, and suddenly he, and not De Gea, who’s on his way to Real Madrid. Welcome to the football merry-go-round.
In the last two seasons at Chelsea, Courtois has conceded one more goal in the Premier League than models expect for the shots on target he’s faced. In Russia, his shot-stopping was worth one full goal during his seven games and he had the most work to do out of all the keepers out there (Here’s lookin’ at you, Bobby Martinez!). In the middle of it all he even managed to get into a mini spat with tiny, little, Jordan Pickford.
Given time to pluck the ball out of the air from a long looping cross, there’s barely anyone better than gigantor, Courtois. He dominates aerially. He faced just three shots on target from the centre of his six yard box. He saved two of them. Pickford faced 11 and saved one. Pickford does not dominate his six yard box. He doesn’t dominate aerially. Size is important. Thibaut was right!Pickford responded to the comments by bigging up his own power and agility and not caring if he was the biggest. Power and agility is important. Jordan was right, too!
Last season Courtois had a real problem in dealing with shots coming from central positions in the area around the penalty spot. Pickford didn’t.Whisper it quietly, but it’s almost like keepers have different make ups and different strengths and weaknesses.Because of his size, Courtois doesn’t often need to power from his set position into full length dives. He doesn’t do it often, and when Courtois does need to get power from his set position, he frequently fails to do so. Because of his size, Courtois is far too in the habit of just collapsing down in order to get to the deck quicker. He has to do this because he takes up fairly aggressive positions which restrict his reaction time.Courtois’ size also goes against him when situations develop quickly. He reacts once the thing has happened. He keeps by numbers, and is reactive rather than proactive. He decides to close out shooters when the ball is already at their feet rather than anticipating the play early. He is s-l-o-w.
Nit-picking? Maybe. But these are the small differences that make or break a goalkeeper’s season.
Courtois’ save % from this area was 50% and way below average. Pickford’s was 64%. And those shots from wide areas in the penalty box like Adnan Januzaj’s goal against this summer Pickford? Last season Pickford saved 19 out of 22 of those. Courtois saved 10 out of 17. But, given time to react from longer range shots outside the area centrally, Courtois swallowed them whole. Pickford struggled. These patterns were exactly the same for season 2016/17, even with Pickford plying his trade at a different club with a different set-up.
Football’s analysis is often confined by week to week constraints. There’s always a game around the corner during the season that needs to be prepared for. But, if you don’t break down a keepers’ numbers it’s difficult to break down their game. You could watch every game in real time and not note these patterns. The record of what actually happens, the information in the long term data highlights the problem. Long term success needs long term analysis and planning. Football has more flashes in the pan than a wok chef with pyromania. The need to ignore short term narratives is huge.
Courtois’ weaknesses also highlights another keeper coaching bugbear of mine. Look at all the publicly available coaching videos. There’s more repetition in the drills than a Jive Bunny track. The keepers know what’s coming. Footwork steps, dive, get up, speed back to start position, footwork steps dive, get up, speed back to start position. Yes, developing some muscle memory is important, but how are you teaching visual clues as to what’s going to happen during the actual match if it doesn’t match your rigid training exercise?Saving a shot involves footwork, it involves diving technique, agility, power, handling. But, first of all, it involves making dynamic decisions about how to handle each situation as it develops. Coaches are taking this away in their sessions.
Every recorded shot on target in a model has all this information built in by virtue of it simply happening in thousands and thousands of top level matches over many years. Long term data tells you what’s working and what’s not. Use it.Or don’t.
At window’s close, with Courtois off to Madrid, Chelsea spend a whopping £72 million to replace him with Kepa Arrizabalaga. Pickford, who was briefly rumored to be on Chelsea’s list to replace Courtois stayed at Everton for at least another season.Kepa is, as yet, unproven in the Premier League, but to prove himself better than Pickford he’ll have to clear a very competent bar. In the last two seasons, Pickford has saved three more goals in the Premier League than models expect for the shots on target he’s faced.Meanwhile, De Gea, sitting pretty above this summer’s transfer fray, he’s saved about five or six times that.
Well...one great season, one bad season so far for Roberto Martinez. Which means the jury's firmly out on the man who had a dream to build a football team.
A 72 point haul followed by a 47 point haul. It's a monumental drop-off, but the underlying xG numbers suggest that Everton under Martinez is a 55-60pt team. Which is exactly the same level the club was at under David Moyes.
It shouldn't be surprising considering 10 of the 15 players that played over 1000 Premier League minutes last season played under Moyes too. The differing styles perhaps means Moyes steady, structured, percentage football regularly got you those steady 55-60pt seasons, whereas Martinez' ocassional football jazz mixed with sterile possession is capable of searing heights but also some minging lows.
Numbers
The graphic below shows all the shots on target Everton took last season (blue) as well as the ones they conceded (red). The bigger the circle, the bigger the chance of scoring:
The black rectangle pinpoints Everton's main problem last season - simply conceding more high quality chances than it created. As discussed here at Statsbomb last summer, Everton got away with conceding these high quality chances in 2013/14 but it was never going to continue. The 50 goals shipped last season was more like (although still not quite) the horror show some were fearful of when Martinez arrived at the club.
But the volume and quality of shots faced are pretty constant over the last 5 years - the Catalan has kept Everton's xG against numbers stable. I've simulated the shots conceded over the last two seasons 10,000 times. The most likely outcome was to concede 85 goals. Therefore 89 goals against all told is only slightly below par but does tally with Martinez's overall lack of efficiency in sorting out defence.
The big question is whether he can continue to keep the lid on the number and quality of shots against. I'm not sure he can. My reasons are not on tactical or style of play grounds. Ultimately, how Everton managers approach the game on the pitch will not define their stay. How they spend limited resources will be. My twitter followers will know that I've been a huge critic of Martinez' squad building since day one.
Defence
For several years, the centre back pairing Phil Jagielka and Sylvain Distin were excellent but relied a great deal on their mobility to hold a high line before dropping back to defend the scoreline late in games. Jagielka was approaching the age of 31 and Distin was 35 when Martinez arrived. You didn't need a crystal ball to think one of them might fall off a cliff sooner rather than later. You certainly didn't need one to see that a bad way to deal with this was to sign a slow, injury-prone, 30 yr old Antolin Alcaraz.
So now with both Distin and Alcaraz having gone to the great mobility scooter in the sky, Everton find themselves with just two first team-ready centre backs. And Chelsea want to get their mitts on the youngest one. Martinez has had two years to prepare for this but here we find ourselves.
On top of this you have 34 yr old Gareth Barry hubbing in front of them. He's signed up until he's 36. Again, Martinez has had two years to sort out this position for the long term but reports are that Darron Gibson (Barry's direct replacement) is in line for a contract extension. Gibson is possibly the best midfielder at the club and certainly the most progressive passer, but he is also completely unable to stay fit (again, no crystal ball needed to figure this out).
Muhamed Besic was signed last summer but has been unable to nudge out Barry or James McCarthy. He'd played just a handful of senior games as a defensive midfielder at that point. He'd played most of his football as a centre back or full back for Ferencvaros.
According to a Tim Lewis piece in The Guardian: 'Martínez, and his chief scouts Reeves and Brown, find the suggestion that they would buy a player because of their numbers pretty funny. "You need to see a player and fall in love with a player," says Martínez. "When you see a player, you'll watch his warm-up, the way he speaks to the referee, the way he speaks to other team-mates after missing a chance, the way he celebrates a goal, the way his team-mates react when he scores. Data might help you narrow the margin of error, but the decision is still a feeling. It's a gut instinct."'
Now, I'm no fan of judging players by bog standard on-ball stats at all, but wow, I hope Martinez was playing to the crowd with some of the sentiments in that statement. I'm not convinced he was. Had Besic been thoroughly scouted beforehand, or did Martinez fall in love with him on the back of one overrated performance against Argentina? Whatever the situation, Besic found himself out of favour at Everton by March. His manager at first club Hamburg reportedly tried to strangle him due to his indiscipline. Having continued to watch Besic play in midfield, I can see why. While centre back is an issue, why not try him out there again? His centre-back style fits what we think Martinez is all about.
To top it all off, the club is wrangling with James McCarthy, its most reliable defensive midfielder (and a solid Martinez signing) over an improved contract. Meantime, the media continually link him to Spurs. Once again, despite the ridiculous amount of money flowing into the Premier League from television, Everton don't appear to have a pot to piss in.
All in all, alarm bells should be ringing on the defensive side of things. It's difficult to see how those stable underlying expected goals against numbers continue to remain steady without new signings. Get yer cheque book out, Bill.
Attack
As Bobby Gardiner suggested a few weeks ago, Everton's attack went missing last year. In a Roberto-esque show of positivity I was confident in the preview last summer that Everton would continue to do well going forward. Despite all the evidence I was still thinking it would get better in November. Well, that'll teach me. Everton have a major creativity problem. I picked on Kevin Mirallas and Ross Barkley last year regarding their output. Time to do so again. Below are graphics (Everton attacking left to right) showing where every successful open-play pass made by these two last season in the Premier League ended up:
Between them they made well over 1600 successful passes, yet look at the lack of penetration into the box. Watch Everton for any decent amount of time and you'll see Romelu Lukaku arms outstretched pleading to be played in. He rarely is.
What he has signed is yetanother player who overruns the ball, doesn't look up enough and frustrates as much as he delights. To come up with a comparable graphic for Gerard Deulofeu I had to bolt both his seasons in senior football together:
You'll see that Our Ged is different to the other two. He's taken less than 400 passes to match Mirallas and Barkley's combined passing penetration into the box. Despite Deulofeu frustrating the hell out of coach Unai Emery at Sevilla, he actually started to walk into some increasingly successful output.
To the eye, this kid doesn't look up from his boots and he's the greediest player you've ever seen. But 8 assists in 1700-odd minutes of senior football across two top leagues is crazy for a 21 yr old. The probability of the average player doing that is miniscule. Like less than 1% miniscule. Over time, that's going to regress.
But when I watch him play I don't see a Barkley-esque 'never going to truly get' it lack of football intelligence. Deulofeu seems to display a willful lack of obedience, like he's conciously revelling in showing off what he's capable of. I'm not sure Barkley could conciously do anything on a football pitch if his life depended on it.
The other thing to like about Deulofeu is his shooting. His xG90 is 0.30 so far in his career. Last season, Mirallas was at 0.19 and Barkley was at 0.09. This is despite the fact it's harder for him to create good shooting angles due to being a right footed player mostly playing wide right.
Maybe even better still for the coming season, Deulofeu only has 4 career goals when xG says he should have 6. In fact, there's a 58% chance the average player would have scored 6 or more by now so I'm pretty hopeful of some 'good' regression here. If he was English, there is no way on earth Everton would have got their hands on a young player with these numbers at a price of around £4m.
Talking of chance creation and xG90s brings us to nicely to Romelu Lukaku. The Belgian scored 10 league goals last year and was only assisted for 3 of them. The other 7 were from penalties or shots that deflected into his path.
It seemed like Martinez was trying to turn him into a more all-round striker, and more of a target man. The graphics show the volume of passes Lukaku has received in each area of the pitch in the last two seasons. First 2013/14, then 2014/15. You can see the increase in Lukaku coming short into the left hand channel to receive the ball:
Lukaku's non-penalty xG was down from 0.58 in 2013/14 to 0.32 last season. On average, each shot on target has gone from a 1-in-3 chance of going in to a 1-in-4 chance. To put this in perspective, both he and Hary Kane took a similar amount of shots (105-112) and shots on target (42-47) but the simulation for each players set of shots say that Lukaku had 15% chance of scoring 15 goals while Kane had an 80% chance of scoring 15 goals. Clearly, Lukaku is suffering from spending the extra time in non-threatening areas and from the lack of creativity around him.
Conclusion
Based on last years numbers, the model says Everton will finish 7th. But this means the defence has to hold up its side of the bargain. As discussed above, I can't see how this happens unless the squad is improved. Cleverley and Deulofeu are decent business - good ages, sell on value, and first-team ready but they don't help here. They should help improve that creativity problem, though.
My best guess is that Everton finish about 10th or 11th again. Depending on ins and outs between now and the end of the window, it could be a bit better or a whole lot worse.
It's the start of a new era in L4 as Everton prepare for life without David Moyes. It's almost impossible to write a season preview without continually referencing the Scotsman's work. Moyes was Everton. Fans seem pretty split right now. They're either looking forward to a breath of fresh air from Roberto Martinez who could finally make the push the club has often threatened to in recent years, or they're dreading an inevitable decline because Moyes milked every last penny of the money available.
A lack of cash defines Everton, and, as is now tradition, speculation is rife about who the club is going to sell rather than buy. Marouane Fellaini looks to actually really be on Moyes' Manchester United radar now after failure to capture first choice targets. The Old Trafford outfit may also still come back for Leighton Baines after a failure so far to flex their financial muscle in the transfer market.
The Numbers
Everton 2012/13
League Finish: 6th
Goal Difference: 7th
Chance Creation: 4th
Chance Conversion Efficiency: 19th
Chance Prevention: 5th
Preventing Goals Efficiency: 11th
With a new manager on board who espouses a different philosophy from the pragmatic Moyes, it's necessary to look at what the new encumbent did at his previous club. Going forward, Wigan improved last year on previous by posting fewer shots overall but taking them in better areas more central and closer to goal. Throughout Martinez’ time the Latics were pretty efficient in front of goal, scoring at the rate they should have. Their output stretched from around a goal a game to a goal and a quarter.
Everton also improved going forward last year both increasing shot numbers overall and in better areas to boot. However, not once in the last 3 years did Moyes’ side even get close to being efficient in front of goal. Last year they were particular inefficient. However, the Blues still outscored everyone below them bar Liverpool and posted 63 points, their joint 2nd best total since 1988.
The graphic below shows how Everton created their chances last year in the “wheat” zone (the zone where goal productivity is best):
As you can see, Everton were big crossers of the ball from high, wide positions. Over half the chances sent into the wheat zone were made this way. The Toffeemen also worked a pretty good number of chances to the wheat zone from within the box itself. What Everton didn't do under Moyes was thread the ball from "the hole". A lot of Everton's play over recent years has been to methodically work the ball wide to create overloads on the left side of the pitch.
Much is made in the media and amongst fans of Martinez' attractive, progressive football. The numbers suggest this is a myth. Wigan created around half the chances that Everton did in the wheat zone. But the proportions of how they made them were almost exactly in keeping with how Everton made them. Lots of crosses, a fair amount of chances created in the box and very few through balls. Wigan played a lot of their football away from danger areas and proportionately took a lot more shots from 'chaff' zones - from wide angles and long distance pot shots. Everton preferred to hem opponents in - no team has controlled the football in the opposition half more than Everton over the last few years. That's no team - not United, not City, not Arsenal, not Chelsea.
And this could be one of Everton's problems with efficiency in terms of converting chances. The pace of Jagielka and Distin enables them to play a high line. With Fellaini able to hold up the ball, Everton can then rely on a small squad of decent footballers to keep the ball there. But being squeezed in allows defences to get set. And if a defence is set, it doesn't matter much if you have the best crosser of the ball in the league in Baines and the best converter of headed chances in the league in Fellaini - it's still going to be difficult to score goals. Unless defences are pulled away from their own goal, space in the box is at a premium.
Everton relentlessly bombarded well defended boxes in a footballing war of attrition. Over time it worked and outcomes were pretty good in terms of points and goals over the season. However, there's an underlying feeling around Goodison that this team is still capable of more.
Nikica Jelavic suffered a lot last season.
The fact is that last year, Jelavic was getting far fewer touches of the ball and having to deal with far more aerial chances arriving his way (13% first season compared to 40% last). This was mainly due to Fellaini's presence up front being so prominent. It's plain to see the kinds of chances the Croatian thrives on - ones played in front of him on the deck, but his confidence was so shot by Christmas that even when they arrived, he started fluffing those too. Jelavic has looked in decent nick in his fleeting appearances during pre-season and scored an absolute belter against Real Madrid last week. Everyone at Goodison is hoping he can regain his goalscoring touch.
Maybe Martinez' less in your face approach will suit Everton and knocking the ball about in less threatening zones will draw those defences out. It may also provide Kevin Mirallas with the kind of open spaces in behind that he thrives on.
Unsurprisingly, Everton's defensive numbers were pretty decent. One of the best sides at preventing chances, the Blues were also pretty efficient conceding 3 goals less than expected for the types of chances conceded. This is no mean feat considering how Everton push up. It's testament not just to the team's organisation but again to defensive pair Jagielka and Distin's ability to hold a high line and provide adequate cover for not just the marauding Baines but Coleman's pacy forays too. Here's the graphic showing the types of chances Everton conceded last season in the wheat zone:
Conversely, defence cannot be said to be Martinez' strength. Wigan's combined goal difference in the league during the Spaniard's 4 years at the helm was a whopping -109. Last season, Wigan conceded 13 goals more than could have been expected. This made them the least efficient defensive unit in the league by some distance. I have numbers for many things but how this is going to play out next year is difficult to determine exactly. Everton have a better goalkeeper and better defenders than Wigan do, but the difference in numbers here is alarming and is the main source of worry to Martinez sceptics.
Incoming Transfers
Four players have arrived so far and it’s difficult to see any further bodies coming in at this stage without one leaving. If Martinez is going to survive at Everton then how he handles limited cash reserves is going to be THE most important factor during his tenure. Moyes understood this and from the off treated the club's cash like his own by gradually replacing older, clapped-out players with younger, hungrier talent.
The striking thing about Martinez’ two main bits of business so far – Arouna Kone and Antolin Alcaraz - is that Moyes wouldn’t have touched either with a barge pole if he was still in charge. Kone will be 30 before the end of 2013 and had a major knee operation a few years back. Alcaraz turned 31 last week, spent last season’s run in injured and is currently injured again. This pair may be useful additions to the squad for the upcoming season, but beyond that, they’re a leak of resources that Everton simply don’t have.
To balance the age thing, 23 year old Spanish goalkeeper, Joel Robles, another buy from Wigan, comes in to replace the outgoing 30 year old Slovakian international, Jan Mucha. It’s unlikely Robles will replace Tim Howard this season. The youngster edged out Ali Al-Habsi towards the end of last season at the JJB, but Wigan continued to leak goals like a sieve.
The player everyone wants to see is Spanish ‘prodigy’ Gerard Deulofeu, who top-scored for Barcelona B with 18 goals last year in Segunda Division. It’s pretty unclear how much this kid will be used during the course of the season. It’s taken until this week for him to feature in Everton’s pre-season and only then for a few minutes. Also, Everton only have him for a season on loan. To me, the (dreaded) youtube reel of Deulofeu looks more like an advert for how bad second tier Spanish football is rather than an endorsement of his talent. Let’s hope he actually is something kind of wonderful.
What we’re left with then, is Kone probably being the new player who will get most minutes next season. I have studied the Ivorian’s shots and goals going back to 2010 when he started playing again regularly after cruciate knee ligament surgery. Remarkably he has retained a decent turn of pace but what’s in store goals-wise probably isn’t what Evertonians may expect having seen him terrorise John Heitinga last season.
Kone may occasionally skip past some challenges and motor down on goal before smashing home, but he only does this a few times a season. His bread and butter goals are often close- in scrappy affairs and he’s not a stylish finisher. Still, the last two and a half years have been decent for him. For the type of shots he takes, he’s scored more than expected in both seasons. However, last year saw a little decline and other work done on this site shows that at his age, the trend is unlikely to reverse. At most, Everton fans can expect Kone to just about reach double figures goals-wise next season in the league.
Conclusion
If Everton can get to the 1st of September without losing players, there's no reason to worry about any huge decline in the coming season. Only rivals Liverpool were within touching distance last season and Everton have a decent level of talent to call upon in most departments. The main problem for Everton continues to be lack of squad depth. The first choice XI has shown on many occasions that they can match the big boys (even if they don't actually beat them very often). The bench, however, is often made up of spark-free senior players of similar or lesser ability to the ones they're replacing or kids who have yet to make the grade.
Unlike most fans, if one of Baines or Fellaini had to be sold, I'd prefer that one to be Baines. He's getting on, he'll lose value and there's a ready made replacement for him in situ in Bryan Oviedo. Coleman could pick some attacking slack up on the right hand side too. Fellaini leaving would leave a huge gap in a not-so-strong central midfield area and the likes of James McCarthy simply won't fill such a void.
It's unthinkable that Everton's attacking efficiency won't get better next season as the play becomes less direct. Any increases here though look likely to be offset by a greater defensive frailty. It will take time for Everton's players (especially the centre backs) to adapt to life under Martinez.
7th place looks a realistic target for The Blues next season.