Defensive Tunneling in the Premier League: ft. Liverpool’s Left Side

Image result for sadio mane tackle

I clustered all Premier League passes this season into 100 “pass types” for analysis using their x and y start and end coordinates. Then I looked at which teams allow proportionally more of each type of pass, the idea that teams either proactively or reactively allow opponents into certain “tunnels” of buildup and passing play. Today we will focus on a couple well-defined tunnels: Man City and Liverpool, both of whom have certain pass types that are allowed among the 10 most often and 10 least often. First the types of passes Man City opponents play the most often:

Unsurprisingly, it’s long balls from the box area and then back-passes in buildup. In fact Man City allow more of those two long ball types than any other team in the league. This is sort of astonishing as they allow the fewest passes overall so to find any pass type allowed with them at the top is surprising.

They do not do anything special once that pass has been played (allow 39% completion rate, higher than the league average) it’s simply forcing that pass in the first place that means the defense worked. Getting the ball into this tunnel is a win in the first place.

On to Liverpool and the more interesting duality of tunneling. First the passes that opponents play often:

Liverpool allow short passes back and forth in your own buildup area, which is something you would not have been able to predict before the season with the Heavy Metal Pressing reputation of the Klopp side. No team has allowed more of these dinky 92% success rate passes. The three teams who allow the fewest of these passes: Tottenham, Man Utd, Man City. Chelsea is also at the bottom. The teams behind Liverpool on the other end: Burnley, Southampton, Cardiff, Bournemouth, Brighton. It’s an odd one. 4 teams have played 43% of the total against Liverpool: Tottenham (51), Man City (45), Arsenal (32) and Huddersfield (27 in one game). On the other end, Spurs have not allowed any team to play more than 14 of these passes in a single game and Wolves are the only team to top 20 total.

This clearly seems to be a strategic choice by Klopp, to essentially allow short passes in buildup without pressure and possibly to shunt opponents toward a certain side of the field.

No other team has such a pronounced shift in where passes flow to. So:

  • Liverpool allow short buildup passes more often than any other team in the league proportionally, more totally than Burnley, Cardiff, other low-possession teams
  • Liverpool see a higher % of these passes go to attacking teams right side than any other team has to either side

Maybe it has to do with the personnel? A quick watch of a selected game didn’t jump off the page video wise but the numbers are what they are: they are stark and point to a clear tactical choice by Liverpool.

Looking at the tunnels and pass types they keep opponents away from could help shine some light on that tactical choice:

So while they are the league leaders in right sided buildup passes allowed, they are on the other end when it comes to passes in the Liverpool half on their left side.

They face almost no pressure from the left side of their box after shunting so much buildup play to that side. It feels like these two might be related somehow. Of course in general Liverpool faces almost no pressure on their box, but Man City faces even less and they have faced 32 more “Pass 52” passes. The difference is quite large.

I think tunneling and pass types opens up analysis in a way that can benefit anyone trying to understand the game better simply due to the fact patterns emerge almost instantly instead of through endless hours of tape grinding.


Scouting a Defense, Part Three: Early vs Late In Possession

Image result for weston mckennie schalke

Welcome to part three of basic prep for an analyst looking at an opposition defense. Our goal is to marshal all the numbers together and give a good glimpse at how the opponents have defended.

In part one we looked at how to find yardage.

In part two we looked at where shots have been generated.

Now we look at early vs late in possessions to see if we find anything interesting. We look at the first three passes of a possession as Early Possession and then passes 4 and on as Late Possession. My numbers are just good, not perfect as far as determining when possessions begin and end.

First we see where teams allow yardage early vs late:

To get a feel for how many yards this is in the “right-back” square (-.5 early and 2.6 late) it’s about 1600 early and 800 late. Leverkusen seem to be gettable down their right side late on in possessions, opponents have gained almost 4x as many yards late in possession on that side of the pitch compared to the other. All these charts are created looking at yards gained per possession loss, using that instead of % of passes leading to a shot for sample size reasons.

Schalke have a sort of weird midfield balance switch between early and late. If you strike quick, the area right in front of the box seems open before they close it up later in possessions.

Stuttgart seem to recover their defense a bit after 3 passes and look average-ish at stopping teams from gaining yardage unimpeded but early on they are about a standard deviation easier to gain on than average. So a forward looking approach after winning the ball back vs Stuttgart might be a good basic plan of approach.

Wolfsburg are sort of the opposite, later on their final third defense weakens. Throughout the entire attack, going down their left side seems to be the best way to go.

So now we’ve got the basics down: where to find yardage, where shots come from and early vs late in possession. We might try to combine the first two in part four or we might now. This is already a good step forward for your weekly preparation and it’s done in 2 minutes. Those hours of tape study can now be much more focused instead of looking for broad trends.

Scouting a Defense, Part Two: Finding Shots

Image result for valentino lazaro

In part one of this series (Finding Yardage) we figured out how to find defensive weaknesses where teams allow lots of yardage without forcing turnovers. We also learned I deadlifted more than Jean-Phillipe Mateta, now we will move into where teams allow chances. Again this is what a base report on how opponents defend should generate, one click of a button and we know where teams are weak yardage wise and chance wise. We can slice this up into different segments of the game (early/late possession most importantly but also into different game states and time periods) but that will come later. We are just doing basic charts right now…and ok, maybe a few Bulgarian split squats and Turkish get-ups. Wouldn’t sound as raw and powerful if it was Danish split squats and Malawian get-ups for some reason. Anyway…

Now we will look at how often teams allow a shot within 10 seconds of a pass in a zone compared to league average. So even if a pass is incomplete, it could conceivably lead to a shot soon. Now this is simply shots, not xG, not goals. Let’s look at the maps:

Hertha don’t let you create right down the pipe but Lazaro’s side of the pitch is a good path if you want to get shots off.

As we saw in the previous article, Bayern are very gettable if you can get the ball deep. Here they don’t show great shot suppression ability in the left/center of the pitch as well.

Leverkusen have leaked shots all year, allowing 13 per game from just 45% opposition possession. They do oddly seem to shut you down in a pretty clearly specific part of the pitch, buildup should maybe be shunted to the attacking teams right side.

These two prep tools take minimal time to prepare and condense a ton of information (probably a good 5-6 hours of tape watching from an expert) into bite size packages that should be the base level of any sort of plan of attack going forward. We’ll dive further into phases of play and maybe try to combine the yardage/chance creation maps in further parts. Now to go to Bulgaria on one leg…

Scouting A Defense, Part One: Finding Yardage

Image result for hard yards
Image result for wolfsburg press

I’m imagining being asked “How do we attack this team with the ball?” as an analyst preparing for a match in a week. I pause my heavy deadlift that I’m doing pointedly in the middle of the gym to show I can hang with the team, gaining the respect of these generally meat-headed athletes and ingraining myself deeply into the team culture so my recommendations will be taken and I’ll get a positive chapter in the inevitable Das Reboot 2: Reboot Harder that comes after our Mainz team wins the Europa League. After I pause that, I head to my office, open R, run a summary, then minimize it so I can open it quickly and open Excel.

Step one: finding how previous teams have exploited our opponents defensive setup. Should we use yards gained? Of course we should, but we need to normalize it for how often teams try to attack through an area. Per pass? Maybe. Per ball loss? Yes, that’s where we will start. How many yards teams gain per ball loss in each portion of the field. Normalize it to league average and pull up a chart…say we are playing Bayern

Bayern’s midfield is generally tough to move the ball through, and their left side has been death to opposition attacks. Down the right side however looks somewhat promising and if you get the ball in front of their goal, it becomes quite soft. This is good preliminary info to have.

Let’s look at our match after Bayern, Dortmund.

Dortmund are drilled pretty well vertically: they let you pass it around from your own goal, then in the extended midfield areas they are basically equal across the midfield while getting tougher in the areas in front of their own goal. No obvious imbalances or weaknesses really (that one green box isn’t enough for me to say anything other than dead center is open) and a well-drilled defense.

Then my bosses are dreading playing the brutal pressing of Leipzig and ask for the hit graphs I’ve been churning out:

Does this mean we don’t want to have the ball in midfield so often as they totally snuff out play there? Well we will need to look at the chance map that is coming in part 2 to know more, but it does tell us that we can’t plan on moving the ball or finding yards anywhere across the midfield at all.

Schalke tried to be a Leipzig type defense but if you get past that first line you can find yards:

Wolfsburg love to press, best way to beat the press is through our right side, good to be aware of:

So now I’ve started to help my team find those hard yards, now excuse me I can go back and finish deadlifting and impressing Mateta with that last set of 3 that I might drop pretty loudly. Can analysts get protein shakes from the cafeteria?

Where In The World Is The Ball (In Shot Buildup)?

Image result for where in the world is carmen sandiego

This was maybe my favorite game show for about a year as a child. Gave me a lifelong love of geography and now Bundesliga Ball Geography.

First let’s look at where teams have the ball in the 3 seconds leading up to shot buildup:

Bayern have the ball closest to the goal, with Dortmund right behind them. Dortmund have the ball the highest % of the time in the center (color of bar). Passes from the center vs the wide and the effects on getting a clean shot off is something I hope to look at tomorrow*.

The most common passing combo for Bayern for these passes is Müller-Lewandowski compared to Schalke’s Caliguri-Uth

*realize these vague promises come true about 10% of the time^

Next we see how Dortmund has slipped further back when you look at 4-6 seconds:

and much wider (see the color of the bar) when it’s 7-9 seconds pre-shot

Looking through the differences in player usage in these various segments between Bayern and Dortmund you really notice Dortmund’s big pieces getting involved in the possession deeper than Bayern. Reus and Sancho are Dortmund’s 2nd and 4th most common passer 7-9 seconds from a shot while none of Bayern’s top 5 are attacking players.

A few interesting tidbits looking through these charts:


  • start wide (7-9 seconds out)
  • stay somewhat deep (4-6 seconds out)
  • quickly come central and close (0-3 seconds out)


  • start very, very wide (7-9 seconds out)
  • stay very, very wide (4-6 seconds out)
  • finish “just” wide (0-3 seconds out)

Schalke, Hannover

  • pretty much are further away from goal than almost everyone else during every segment of shot approach


  • oddly are extremely close to goal during the 4-6 second segment after being further from goal at 7-9 seconds, lack of quality to make the tougher passes?

Hoffenheim, Bayern

  • basically central and close the entire sequence

The Nagelsmann to Leipzig move will be a fascinating one as there will be a pretty large clash in how they’ve played this year to how he wants to play. I’m guessing there will be no more Laimer starting wide next year and I’m on record predicting the Mukiele Explosion next season. Nagelsmann and Leipzig approach shots about as differently as you can this season.

^Cleanliness Of Shot

Ok a quick and dirty look at shots from just this year with blocked shots removed

% of shots that go on target from 5-9 yards away:

Coming after a pass from the center: 63%

coming after a pass from wide: 48%

% of shots that go on target from 10-14 yards away:

Coming after a pass from the center: 69%

coming after a pass from wide: 55%

Bundesliga Monthly Checkup

After a bit of time away, it’s time to check in on what players and teams have been doing interesting things in the past month. Games starting the weekend of Feb 8 are included here.

James Show in Bavaria

-Only three players have progressed the ball into the Danger Zone more often than he has over the past month: Sancho, Ribéry and Herrmann. Ribéry and Herrmann have done it in minimal minutes.

-Of the top 25 DZ progressors…

  • only 2 gain more yards per minute: Piszczek and Kimmich.
  • only one gains more yards per ball loss: Kimmich (who is 24th in DZ progressions and does it less than half as often and from wide areas.)
  • James is doing it via the pass: only a couple playes (Amiri/Eggestein/Sabitzer) have a higher % of yards gained coming from passes.
  • How he does in the Liverpool game will go a long way to defining his public image, which has taken several blows from Honigstein for some reason and many of the Bayern fans seemed to truly turn against him after the last Liverpool performance, but right now he’s probably the best player in the Bundesliga.

The Return (To the State Fair)

Tell me this isn’t the most State Fair Looking guy you’ve ever seen

Peter Bosz at Leverkusen has triggered a massive compilation of passing numbers (see vs Düsseldorf)

and Mr Aleksandar “State Fair” Dragovic has surged to the top of the monthly yards gained leaderboard on the back of it. His 9.2 yards gained per minute over 248 minutes is more than Jerome Boateng and David Alaba…combined. He basically was the beneficiary of Leverkusen having ~80% possession in the Freiburg/Düsseldorf games and this doesn’t really tell you much about Dragovic I don’t think, more about Bosz and Leverkusen. About 1% of his passes led to a shot within 10 seconds so it’s not like he’s breaking lines here.

The One Man Attacking Machine

  • Vincenzo Grifo has 13 of Freiburg’s 28 Danger Zone Progressions.
  • No other Freiburg player has as high a % of their passes turn into shots with 10 seconds
  • No other Freiburg non-defender/GK gains yards at a per-minute clip of even 70% of Grifo.
  • Freiburg have lost just 1 of their past 6 while scoring 12 goals.

Schalke Are An Embarrassment

I usually think soccer coaches get fired a bit quicker than other major sports and probably a bit too far, but Tedesco at Schalke is hanging on far beyond any seemingly reasonable decision should be made.

I don’t quite remember a team as “big” as this free-falling to an extent like this. At the start of the year they at least defended well, which didn’t make up for their complete abdication of trying to move the ball outside of CHAHs (Caliguri Hit-And-Hopes) but now the defense is collapsing also, only hapless Nürnberg have allowed more DZ progressions. Of course the 7-0 vs Man City made this even more clear.

A Worry For Frankfurt?

No team has allowed more yards gained per pass over the past month than their 8. With the fight for the final two CL spots have heating up massively, every little bit will count. They pulled a forward I’d never heard of in Gonçalo Paciência in and he’s scored twice and dropped an assist in the past 10 days. Lots seems to be going right for them this year, but this ease of pass-through could be worrying.

See If You Are Managed By Pál Dárdai With One Cool Trick; Bonus Nagelsmann

I was messing around in Tableau looking at passing clusters by possession length as one is wont to do on a Tuesday night and found this fascinating tidbit I knew I had to rush out to my readers. I was looking at what types of passes teams make when they are deep into a possession: 7 uninterrupted passes or more into a chain. I have 80 pass types out there and basically every one is “led” by Bayern or Dortmund, with a sprinkling of Gladbach. This makes sense as they are the 3 teams with the best completion% in the league so the most likely to have long possession chains. But there are 2 pass types where Hertha have the most “late-possession” completions, which is odd because they are 11th in pass completion% and 15th in possession, they really shouldn’t lead any of these categories.

Then you see the two pass types:

This is Pál Dárdai’s Defensive Possession Revealed With One Cool Trick. Late in possessions Hertha passes backwards more than anyone in the league. This is just 53 total passes, but no one else has cracked 40 yet. Reading this, I’m sure Ralf Rangnick is tearing out his hair, how can you pass backwards 8 passes into a possession, he is yelling. The king of quick progression has his teams playing that way: Leipzig play these late-possession retreating passes less than anyone else with just 11 on the year.

Bonus Nagelsmann

I’m not sure whether to mark this down as a manger tic to follow but it’s one to note as a potential: Hoffenheim try the Money Ball more than anyone else overall, and especially late in possessions.

Overall they have tried this pass type 144 times this year, second place is Bayern way back at 115.

Interesting Bayern essentially never try this pass late in possessions: they’ve tried just 18 7 passes or more into a possession while Hoffenheim have tried 44:

Interestingly Bayern lead the league by a huge margin in these late-possession crosses, with Kimmich and Müller taking a ton of them. Maybe this Money Ball Nageslmann tic is just a result of personnel, with Demirbay/Kramaric and Joelinton fancying this pass, or maybe he really focuses on getting to the center in established possession. Hard to know, interesting to see.