Real Madrid and Liverpool met in Kyiv for the 2018 Champions League Final, providing fans across the globe with excitement from the get-go: Ramos brought down Salah early on causing the Egyptian to leave the pitch in tears, Bale came on to make the difference in the final half hour, winning Man Of The Match as he scored, arguably, the greatest ever Champions League Final goal and Karius' two blunders cost Liverpool the game, but his poor performance was credited to an apparent concussion a few days later. Madrid won the game 3-1.
I have decided to dive into the stats using Statsbomb's free data once more to assess the shooting abilities of both sides on the day and try to see where Los Blancos made the difference or was it simply a bit of luck (and a concussion) on the day?
Firstly, I took to R to assess where the shots were taken from for each team and what their outcome was, with the quality of each chance being shown through the xG value of each goal. Opta's definition of the Expected Goals metric states that it "measures the quality of a shot based on several variables such as assist type, shot angle and distance from goal, whether it was a headed shot and whether it was defined as a big chance".
Next, I looked at where they ended up in relation to the goal. This is quite useful in understanding if a shot was like Bale's second or third. In other words, if the goal was down to great finishing or poor 'keeping. Obviously not all scenarios are as black-and-white as this, but it is helpful in gleaning some insight into where the game was won and lost.
Moving over to Tableau, I had a look at how the game changed over time by creating a cumulative xG plot and another visualisation showing the xG value of each shot and the time it occurred. This is a great way of seeing the ebb-and-flow of the game, and could be done in a way to show sustained periods of attack from one side before further exploring why they didn't come away with a goal, perhaps.
Finally, I downloaded a YouTube video of the match highlights and made a basic buttons template in Nacsport to have a better look at the major opportunities that shaped the game and how the technical and tactical differences lead to Madrid winning their third consecutive European Title. From this, I was able to annotate some of my thoughts onto still images of the game using KlipDraw.
The Story of The Game
Using Tableau to look at the simple shooting statistics for both sides, as well as the advanced xG metric, we see that Real Madrid were most likely deserving winners based on the quality of their shooting positions. Both sides managed 14 shots during the game. I am immediately drawn to the 8 Liverpool shots that were blocked by Madrid - a huge 57% of their shots. I would question if this is due to Liverpool shooting in poor quality situations where there are lots of bodies in between them and the goal, or if this was perhaps due to the Madrid players defending intensely and quickly closing down any shooting chance.
Using the cumulative xG total for both sides over time shows that Liverpool started the game well, having plenty of shots in the first half hour. Madrid, on the other hand, are kept extremely quiet. However, in the 30th minute, Liverpool's talisman Mo Salah was forced off after injuring his shoulder in a challenge with Ramos, and the tide begins to turn. By half time both sides have created almost equal quality of shooting chances and go into the break at 0-0.
Early on in the second half both sides create high quality chances and convert them; 1-1. In the final half hour of the game Madrid appear to create a higher overall quality of shot opportunities, but in reality both goals would come from low xG values.
To improve this plot I would want both lines to start at 0 for xG and minutes and finish at 90 minutes, but this does not affect the messages being relayed.
The quality of all shots from both teams can be seen in the next visualisation.
Now we can see that nine of Liverpool's 14 shots came in the first half hour before Mo Salah's exit. Both sides converted their best opportunity of the game. Bale's two goals that sealed the win had extremely low xG values, showing that the frequency of similar shots that would have been converted from these situations would be very low.
Shooting is the final part of the attack, and although these visualisations suggest that the loss of Salah greatly limited the Liverpool attack, data can be used to assess if this was the case. For example, a rolling average of the number of attacks would show a large dip after 30 minutes if this was the case. Deeper analysis could be used to show if Liverpool's build up play was affected by losing Salah, or perhaps Lallana (Salah's replacement) was not a limiting factor in this, but simply acting as more of a creator than the finisher, like Salah, attempting to craft out a better opportunity to shoot that never came.
Real Madrid's Attack
Data Visualisation
The statistics used above can be better visualised through a shot location map. Madrid's shots that missed the target were all taken from difficult angles (the two to the right), or from distance (the two ~25 yards out to the left). The other shot from outside the box was Bale's 30 yard strike which slipped through Karius' hands.
When looking at the end location of these shots in relation to the goal there are only 11 shots seen on the visualisation. This is because blocked shots don't reach the goal. There is no point in including colour for the shot outcomes in more detail than if it was a goal or not as it is obvious; the large red dot on top of the bar is clearly the shot that hit the woodwork, those outside of the rectangle representing the goal are obviously off target, meaning the three red dots inside of the rectangle must be those saved by Karius. When making this visualisation one comment I would make is ensuring the correct dimensions are used for the goal; i.e. using yards in this instance as that is the unit of measurement which Statsbomb uses for their location data.
From this image, we can see that the goal scored to the bottom left had a high xG value, this was Benzema's goal which he scored after Karius attempted to roll the ball out for his defence to build up play with, but instead gifted Benzema an open net. The goal scored in the top corner was Bale's overhead strike - a fantastic goal with a low xG value as so many other players attempts at acrobatics would end up in row Z or with a lesser player being carted off injured. The third goal is very central and also has a low xG value. This was Bale's long range effort, and Karius' second mistake which led to a goal.
Without consulting the footage, one would expect Bale's first to be a good finish given its final location, whereas his second is so central - a comfortable height and position for a goalkeeper, so it makes you wonder what happened for that goal to go in, of course we all already know that it was Karius' unfortunate mistake.
Two other opportunities hold large xG values; the shot that hit the bar, and another from about 8 yards out which was saved. In both instances, it looks like they may have been poor finishes given how central their end locations are. We would then consult video to see if this was actually the case, or perhaps another factor caused the shots to be placed centrally.
Henderson covering at centre back
Upon consulting the video, there are two aspects of Madrid's attack/Liverpool's defence that I noticed where Madrid had an opportunity to exploit.
Firstly, Liverpool's marauding full backs Alexander-Arnold and Robertson are an extremely important part of Liverpool's attacking play, pushing high up the pitch to set up the front three from wide. No one can dispute that the pair also show excellent work rate when coming back to form the back four once the ball is turned over. Henderson plays as the deepest of Liverpool's midfield three and slips into the back four if need be to cover his full backs. However, Madrid's pacey attack did reveal slight tendencies for Henderson to be attracted to the ball, rather than maintaining good positional sense.
In this attack, Marcelo is carrying the ball towards the Liverpool box whilst TAA works hard to get back. Henderson has slipped into centre back and pushed Lovren out to right back to cover. Lovren sees Ronaldo as the obvious passing outlet so is about to push onto Marcelo to try to cut this out. If a good centre back like VVD were in Henderson's position here, they might be more likely to keep backing off, ready to move out right to close down Ronaldo if the pass is slipped through, but more importantly to mark up Isco for now.
However, Henderson instead does not back off but steps up with Lovren, one pass would take both players out the game, and give Ronaldo and Isco space behind inside the box. Robertson is keeping both players onside. Marcelo reacts to Lovren's pressure and tries to play inside to Ramos who had joined the counter attack from deep, the pass is cut out, but in doing so deflects it into Isco's path.
Isco is now inside the box with no defenders to beat, but his shot misses the target. If Henderson hadn't stepped up with Lovren, you would expect him to be here to at least challenge Isco for the lose ball. Isco had an excellent opportunity to punish Henderson here.
Henderson's will to win the ball here could possibly be used by opposition players. If this is something that he does frequently, then attempting to force Henderson to commit during counter attacks when he is covering the full backs could lead to more attacking players getting the ball free in dangerous areas.
Madrid's crossing game
Like Liverpool's full backs; Carvajal and Marcelo get great joy from joining in with attacks. Madrid overload the wide areas to get into good crossing positions and put in some great balls. In the next image, Carvajal has an opportunity to bypass five Liverpool players with just one ball, and the Madrid striker in the image has a good amount of free space between him and the defender to get an uncontested shot if the ball is better.
Playing this cross into the front post is dangerous as it doesn't give time for Karius to come out and claim it, with the Madrid player running onto it he is a clear favourite to get onto the end of the cross. The Liverpool defender should be touch tight so that he can follow the front post run, and if he does not win the ball, at least he might put the striker off and prevent a goal by contesting with him.
This was potentially a great opportunity, but the ball ended up being played behind the attacker, so his shot on the turn was difficult to convert and produced a good save from Karius.
Again, looking at another crossing opportunity, we see that the defensive shape is questionable. In this case there are two lines with the midfield three and back four clearly visible, which is better.
When defending, Klopp required his two wingers Mane and Salah/Lallana to drop deep and form a midfield five as Madrid's full backs push on. Lallana is doing this, putting pressure on Isco's cross, however Marcelo has caught Mane out and is free at the back post. Although it is a difficult ball that could well be cut out by TAA, or Mane could possibly be fast enough to get back in time, he makes no real effort to get back in once he sees the ball being played more centrally towards the front two. On another day, Marcelo would have been rewarded for continuing his run, with the second ball spilling to him to finish.
In this instance, Ronaldo won the header in the middle, Karius palmed the ball out and Benzema put the ball into the empty net. It was ruled out for offside against Ronaldo, as this image rightly shows. In my opinion, I like the position that VVD takes up, ready to clear the ball if the cross does not clear him, and preventing any Madrid player from coming in front of him. TAA also stays high, but Lovren is already backpedalling towards the goal to track Benzema's run. The unit has to work in unison, they almost came undone here, and should instead either all stay high and play the offside trap or follow the men, goalside (TAA to Ronaldo) and tough tight (Lovren to Benzema).
Liverpool's Attack
Data Visualisation
Instantly, we see how well Madrid must have defended the central areas, with all the blocks coming in this area, it seems to suggest they weren't as passive in their defensive style as Liverpool appeared to be in the examples above. It is quite telling that the only two saves that Navas made were from wide right in the box; extremely difficult angles to score from.
However, credit to Liverpool they are obviously doing something right to create shooting opportunities in such central areas within Madrid's box. A clear limiation of xG is that it only gives value to an actual shot that a player takes. It would be useful to assess the number of other attacks from both sides and how they were ending. It appears here that Liverpool were intent on crafting an opportunity from around the penalty spot, the low xG values are most likely due to there being defenders between the shooter and the goal. It would be interesting to see if Liverpool worked other attacks into similar positions at other points in the game, but did not pull the trigger due to the likelihood of not finding the target. Liverpool play quick, short passes in the final third which may close down their own space, as a result of drawing players into the central areas, which is why so many shots may have been blocked.
Missing the eight blocked shots, the shot location map contains just six Liverpool shots. Unlike Madrid's, most of Liverpool's shots that made it to goal seem to be struck with intent to find the corners, of course this is always good practice, but especially so given Navas stands at only 6'1". The only shot struck centrally is Mane's goal, which he poked home on the volley from less than six yards, after Lovren climbed well to beat Ramos at the corner.
Ramos, Varane and co.'s intense on-ball pressure
Ramos and Varane's partnership is one of the best defensive partnerships in footballing history, and the trophy cabinet reflects this. One might want to use buzz phrases such as "they wanted it more", when looking at how Madrid defended. They swarmed the ball whenever Liverpool worked a shooting chance and this work rate is great to see for any fan or neutral, however, there is much more to it than simply working hard - they risk leaving other Liverpool attackers free at times but increase their likelihood of winning the ball in doing so.
Liverpool seemed intent on getting to the byline and pulling the ball back towards the penalty spot for one of the front men to finish. There are plenty of articles out there listing countless xG facts on why this may not be the most productive way of playing football - ultimately, plenty of these shots are missed as there are defenders between the ball and goal who can get in a block, as we see during this game.
In the first image, Ramos, Varane and Carvajal close down Salah to block the shot. This wasn't actually from a pull back, Salah received the ball facing away from goal, and instantly turned to craft the shot. The Madrid defence is narrow, suffocating this central area exactly for this reason, Varane and Carvajal both shift in to join Ramos in blocking the shot. Two Liverpool players are unmarked to the left, but this is not an issue, for one, Salah is only shaping to shoot here, but two, both Varane and Carvajal have come from this area and positioned themselves well on their movement across so that a pass could not be played straight to the left. When Salah takes the shot, they are touch tight to him and would be able to prevent any pass or shot from being played successfully.
In this next image, Carvajal and Varane again work together to shuffle Firminho away from goal. They are intense in their defending, forcing Firminho to eventually pass out and start again. When the ball goes back they can return to their defensive shape. This means that TAA being unmarked is not an issue as he will be picked up instantly by Marcelo or the left midfielder as they reset their positions.
Again, Liverpool craft an excellent shooting opportunity as Robertson pulls back for Firminho, however, there are so many bodies between him and the goal it was always a difficult opportunity, as shown by the low xG values. However, Ramos and Varane make it almost impossible to get it past them as they charge out, but importantly don't oversell themselves, risking a foul or being beaten. Their defence is always intense but controlled.
Modric here shows that this intensity is well drilled into the whole Madrid team, not limited to the defence as he charges down Salah's shot from the free kick. It is a well worked move from Liverpool. Salah pretends to take the free kick, instead stepping over the ball, and Milner reverses the pass back into Salah's path. Milner's body shape would fool plenty into thinking he is crossing the ball. Against another team this is a free kick routine that I would be interested to see, as I think it could be pulled off if the man closest in the wall doesn't close down the shooter quickly.
Conclusions
Madrid came away as 3-1 winners, and their greater xG values suggest that this was deserved, however, I think that a lot of credit should go to them limiting the quality of Liverpool's chances by getting men behind the ball, defending the central areas well and not allowing Liverpool's attackers unchallenged shots on goal - the only real shot which came unchallenged was Mane's goal where he won the second ball at the corner. At the other end of the pitch, Madrid worked the ball into some good positions out wide when building up slowly, and attacked the central areas well on the break. Liverpool had some nervous moments where they could have been caught out, but were not. Ultimately, it was a game lost through two unfortunate mistakes.
This was a project that I really enjoyed doing, and was pleased to use a few different types of software. I came at this thinking I would solely focus on Madrid so it appears that I seem to compliment their defending a lot when assessing Liverpool's attack! This bias was not intended. I was solely aiming to produce some new visualisations relating to shooting that I have not done before, then decided to go one step further with some video analysis to back up my thoughts.
These Liverpool and Madrid sides are two of the most exciting teams in Europe at the minute, so it is no surprise that they produced such a good final for the neutral viewer, nor that they have both stormed their domestic leagues this season.
The visualisations that I have produced give clear messages and can be communicated very easily. There are obvious limitations to using statistics such as xG but for now it seems to be very popular within British football, especially with it being given so much broadcast time. I would like to look at different ways of assessing attacks rather than just shots to add another layer to my analysis going forwards.
Comments