Welcome back to the second installment on why I think British pundits don’t know what they are talking about. As one commentor astutely pointed out on my previous piece, pundits are often pundits precisely because they’d never be able to make it as coaches... so I don’t know what that makes me as a blog writer, but I digress.
The turmoil hasn’t stopped at Tottenham Hotspur, with a disappointing draw against Wolverhampton Wanderers following the loss to Nottingham Forest. More squad issues have abounded, with an injury to Destiny Udogie and suspension for Rodrigo Bentancur, adding more fuel to the flame as ex-Liverpool, Manchester United, and Arsenal players gleefully assert the unsuitability of Spurs’ tactics from the safety of the commentary box. I did think the commentary team for the Wolves fixture were largely kind to Spurs and unusually insightful as to the context for some of the recent results - but that won’t stop me from addressing another regular critique: the “suicidal” high line.
Leaving aside the obvious point that the use of suicidal in this context is very much in poor taste, let’s discuss another regular talking point permeating the #discourse:
The high line is too risky, reckless, and responsible in large part for Spurs’ defensive issues
A couple of caveats before I launch into this one. Firstly, this is going to be less of a tactical analysis and more of a look at the numbers - so feel free to turn off if that’s not your thing (you shouldn’t, though). Secondly, numbers around defensive positioning are quite hard to come by, so the data compiled in this piece have been largely cobbled together from various sources (but those that I believe have a decent level of veracity).
So, let’s start with the elephant in the room - we can probably all agree that Spurs’ overall defensive record has not been the greatest this season. There were positives early: Spurs actually had the third best defensive record in the league (goals conceded) going into the Liverpool match, and that was AFTER conceding four to Chelsea. The underlying stats, however, tell a slightly different story.
Tottenham had conceded 19 goals up until this point; expected goals against (xGA) metrics had Spurs as worse off, with xGA of 25.21, 7th in the league, and non-penalty expected goals against (NPxGA) of 23.69, 8th in the league.1 A drop-off could arguably have been... well, expected, with Spurs’ defense really tracking at closer to a mid-table than title-winning level.
Is the high line to blame, though?
Let’s get our bearings. It’s hard to define a high line, as mentioned above, there are limited data available with regards to defensive positioning. Defensive Line Height, the average distance from a team’s goal to where defensive actions take place, gives some indication; but is strongly related to Field Tilt, the areas of the pitch at which play is occurring, meaning it can’t really be used in isolation as a measure of high line. Typically, lower table sides will have less of the ball, and therefore sit deeper on the pitch, especially against a “Sky Six” side. You could also incorporate possession into that discussion as well, where you again would often (but not always) see increased possession with the big teams. That said, we have to take what we can get, so I have charted this below2 (sans-possession stats):
An interesting thing to note here is that Spurs’ defensive line doesn’t appear especially high via the Defensive Line Height measure. It’s up there among the more aggressive sides, but is only 8th in the league. You can marry this up as well though with Spurs’ play on the ball to help contextualize further the positioning of Spurs’ defense. Below, I have pulled touch data for the starting center backs (those who have started the majority of matches according to WhoScored) from nine teams: all teams with Defensive Line Heights above 46 m, and teams in the Sky Six below that mark (i.e. Manchester United).3 I have excluded the opposition penalty area from this breakdown, as touches in this area are often due to attacking set pieces, and not reflective of defensive shape or style. You can see Defensive Line Height charted against these possession numbers just as an additional reference:
Here’s where the picture starts to become clearer. Ange Postecoglou’s positioning of his defensive line is aggressive; but not really anything out of the ordinary when it comes to modern, fluid, attacking football. Manchester City, for example, have the majority of their actions / possession in the middle and attacking third; Arsenal are not quite as forward-thinking, but still position themselves quite high up the pitch. Then you have United, stinking up the joint even more than Spurs, who have played closer to a mid-block, and have also probably had some of their numbers colored by the shift to a back three.
Spurs though (and to be clear, the series labelled “Spurs” refers to data covering Cristian Romero and Micky van de Ven) are actually reasonably conservative in comparison; still aggressive relative to a lot of the league, but compared to other sides that try to play front-foot football not exactly deserving of the superlatives that have been thrown around as to their recklessness. An even more interesting point starts to emerge when you look at the data for Spurs’ other center backs (labelled Spurs other on the chart): Radu Dragusin, Ben Davies, and Archie Gray - with these numbers probably look slightly more aggressive than reality, due to Gray spending some time at fullback as well as CB.
There is a clear shift towards the defensive third, with 58% of touches occurring in Spurs’ end with the backups as opposed to 52% with Tottenham’s preferred pairing. That signifies a potential shift to a slightly deeper line, data that back up my own tactical observations of recent fixtures (for instance, against Wolves, Spurs were utilizing more of a mid-block). I dive into this a bit more below, but without fixture-by-fixture data, it’s hard to really quantify this. I will say though that when I first looked at my source for Spurs’ Defensive Line Height stat, Spurs were sitting around third or fourth (which was based on data up to December 12th); there is a definite drop in our defensive line positioning, but unfortunately in this particular case - you’ll have to take my word for it!
For all the criticisms around high lines, however, it is clear that it is a robust defensive strategy. If we dig into the best defensive sides,1 you will see that in general, the high line is not the issue. Five of the top six best defenses utilize aggressive defensive setups, with Nottingham Forest, who play with a low block, the lone outlier. Fulham (not in the below table) are the other side of note, utilizing a mid-block and sitting seventh by goals against / 90, but jumping up into fourth if we look at underlying stats.
Thus far, we have established the following:
Spurs play a high line
Spurs’ high line isn’t especially risky or reckless with regards to positioning
The majority of the best defensive sides in the Premier League play with a high line, so there’s nothing wrong with the tactic, per se
Spurs have dropped their line slightly deeper over the last month
So what’s the problem?
In my view, there’s two main issues: defending in transition, and personnel, with the two being heavily related.
Defending in transition
The high line always leaves space in behind, and opposition teams will look to exploit this by sitting deep, winning possession, and quickly countering. A side’s ability to diffuse these situations can be measured in part by assessing the number of shots they concede from these breaks4 (as with a number of parts in this piece, you could delve deeper and look at shot quality as well, but I’m not looking to write a novel here). If we look at the sides playing with a high line and compare their Defensive Line Height to Fast Breaks Conceded per 90, the story told is rather illuminating:
It turns out the sides with the best defense in the league all concede fewer shots from fast breaks! Who woulda thunk it? Arsenal, Bournemouth, Chelsea, Newcastle, and Liverpool all play with high lines that opponents struggle to exploit, with all of those sides conceding less than one shot per match off a fast break. Spurs, Brighton, and Manchester City, however, are all less successful when stopping opponents in transition (please note - I had to pull these shot data individually, so I may have miscounted - in one case I definitely have, missing one against Spurs which means the Spurs data point should be aligned with Brighton and City with all three conceding 22 shots from fast breaks across their matches).
There are I think a few reasons for this (at least in Spurs’ case). One is the tiredness of the squad; we’ve seen on multiple occasions the likes of Pedro Porro and Yves Bissouma cover ground like they’re running through treacle, pushing without success to get back and make necessary defensive interventions. This flows onto the second reason, which is ineffective individual performances and questionable decision making (which is often exacerbated by said tiredness). Both Rodrigo Bentancur and Bissouma have been quite poor in the defensive midfield role at times, with Bissouma particularly struggling in recent weeks. We saw as well against Liverpool Radu Dragusin make a couple of notably bad decisions which led to goals.
The third is game state: of 22 fast breaks leading to shots, 20 of them were with the score tied or Spurs trailing. That’s in quite strong opposition to the talking points from the pundits, who have regularly referenced Spurs’ high line as a reason for any inability to hold onto a lead; instead, Spurs are really only getting hurt on the break when they are searching for goals (and thus three points). Ideally, Spurs need to score first.
The final reason? Well, that’s also the other issue with Spurs’ high line I referenced above:
Personnel & injuries
Look, it may seem reductive, and the very obvious thing to talk about... but it turns out injuries are a thing that kinda hurt the way a team can play! Somehow the commentary teams don’t tend to be aware of this, as when you lose three of your back five (make that now 4/5) it does have an impact on your defensive structures. There are a couple of ways that has played out.
One that has probably gone under the radar is the loss of Guglielmo Vicario. Though Forster has stepped in and done a solid to decent job at times, there are some severe limitations with his game that have really hurt the way Spurs play, and more specifically, defend. A high defensive line requires a keeper comfortable with coming off their line to cut out the space between the defense and the goal, and this is something Vicario is adept at. It’s also something that Forster struggles with, with the Englishman often glued to his line and affording opposition attackers ample space and time to attack the Spurs goal.
Vicario averages 1.67 defensive actions outside his penalty area every match (#OPA/90), with an average distance for defensive actions of 14.1 meters; Forster’s #OPA/90 is 0.43, with an average distance for defensive actions of 11.1 meters3 - that’s a significant stylistic difference, and that 1+ extra action per match could be the difference between your goalkeeper intercepting a ball just before an onrushing goalkeeper can reach it, and being caught in no-man’s land, giving the attacker a one-on-one opportunity.
The other big impact has been due to the injuries to the starting center backs. I referenced this with some data above, indicating the Spurs defense had shifted slightly deeper following the injuries to Romero and van der Ven. Unfortunately, as we are all aware, that hasn’t helped much - and the gap in quality has become more and more clear as time has gone on:
The underlying stats get worse when Romero and van der Ven aren’t lining up together; and significantly worse when neither are available.3 Note that I have excluded the Chelsea match from these brackets - it was a bonkers match that saw injuries across the pitch including Romero forced off inside the first 15 minutes, with game state impact aplenty, thus making it an outlier (and it actually makes the Romero / van der Ven combination look BETTER in the above metrics if we include it).
This may seem like an obvious point, but it’s nice seeing it laid out like this. And remember those fast breaks leading to shots? Excluding the Chelsea match (which saw two occasions in which a fast break led to a shot), Spurs conceded eight shots from fast breaks in the 11 matches in which van der Ven and/or Romero were available (a rate which sticks Spurs between Newcastle and Chelsea on the chart further up). Only one of those resulted in a goal. That’s right, Spurs’ crazy high line had only been exploited for a solitary goal until snipers started targeting hamstrings.4
When Spurs’ starting CBs WEREN’T available? 12 shots from fast breaks in 7 matches, three resulting in goals, and all the matches in which those breaks occurred were matches in which Fraser Forster was the goalkeeper.4 That is a STARK difference and one that is worth noting. Is it Forster’s fault, the center backs’ faults, or Postecoglou’s fault? There’s probably enough blame to go around, but Postecoglou HAS made tweaks and the defense is still struggling.
Let’s sum up one more time:
Spurs play a pretty standard high line in line with other top sides
When Spurs’ starting goalkeeper and center backs have been available, their defense is very good (though not elite) and their attack amongst the best in the league by underlying metrics
There is a huge drop-off when these players aren’t available
Spurs are being largely hit on the break when chasing goals
Additionally, Spurs are probably struggling to defend in transition due to tiredness and some of the above personnel issues
So what can Ange Postecoglou do about it?
Well, like I said above, he has already made some tweaks. He has dropped the defensive line deeper somewhat despite claims to the contrary, and has been playing slightly more defensive midfield setups in recent weeks to assist the defense.
I would be reticent to drop into a full-time low block or mid-block; doing so can invite pressure, and we don’t really want 18-year-old fullback / midfielder Archie Gray spending large chunks of games defending. Additionally, though lower block systems at times can be hard to break down, they do invite alternative avenues of attack - like crossing. And Spurs (Dragusin aside) are pretty bad in the air across the backline, starters or no.
There could be changes made to defensive positioning in the attacking and transition phases, though that very quickly becomes a different system altogether; or there could be changes made to the way Spurs press, to help retain energy for recovery sprints. This is where I see an opportunity for further change.
Ange has eased up the press a little in recent weeks, but Spurs are still one of the most aggressive sides defensively without the ball. Pressing is necessary for a high line to function well as opportunities for balls to be played in behind need to be minimized - anybody here remember when Andre Vilas-Boas decided to stop pressing with his Spurs side while still playing a high line? Those were some times, that’s for sure.
That said, the way Spurs’ press has dropped off a little hasn’t been into more of a situational nature, and that’s perhaps the problem I see. Most pressing systems have a trigger, an action or scenario the opposition finds themselves in that then signals to the defense to spring the trap; Pochettino’s Spurs side would often press the fullback when they received possession close to the touch line. Postecoglou’s Spurs, on the other hand, have a pressing trigger of “yes”. As far as I can tell, that has now eased up to “almost always yes”. By identifying some more specific triggers, Ange could potentially ease the physical toll on the team which in turn could lend itself to better defending in transition and recovery runs.
So, once more, the pundits have it wrong. Spurs’ high line is not particularly reckless. In fact, Spurs have been dropping deeper in recent weeks to try and manage the issues with the squad. The funny thing though is that there is room for criticism of Ange here. All of the above carries with it levels of nuance. I mentioned tweaking the press. Ange has already done so, but could he do so more? What about some positional subtleties? Unfortunately, nuance and injuries don’t seem to make as interesting of a narrative - and pundits exist solely for narrative’s sake.
Join me next time for part three, where we’ll talk on more commentator crapshoots.
