Impotent set-pieces, squad turnover and no defensive midfielders: Why Oxford's defenders don't score anymore

Luke McNally's thunderous 95th minute header against Ipswich was a sublime moment, rescuing an unexpected point against a playoff rival. 

Whilst this late drama has been a staple of football at the Kassam this season, a goal from a defender has become something of a rarity. 

This was McNally's fourth goal of the season, more than double Oxford's next most prolific defender, Steve Seddon, and Oxford's whole defensive tally is only eight goals.

In a season when Oxford are scoring so many, why has the attacking output from the defence declined? 

... 

Over the past few seasons, Oxford have been gifted with some remarkably consistent goal scorers at the back, and this was clear for all to see last season. 

The emergence of Sam Long as a genuine attacking outlet, along with Atkinson's expansive style and Ruffels' impressive ability to ghost in at the back post helped make Oxford the most prolific goalscoring defence in Europe. 

No side in Europe's top five leagues plus the EFL registered more goals from defenders than Oxford last season (19). 

You will be hard pressed to find another graph with a top three of Oxford United, Inter Milan and Burton Albion, and the U's achievements becomes even more remarkable when factoring in the styles of the other teams on that list. 

Most of the sides favour a five at the back formation: Antonio Conte's Inter Milan adopted the coaches' favoured 3-5-2, whilst Burton also used a five at the back for most of their matches. Chelsea spent a large part of last season with Thomas Tuchel as their coach, lining up in a 3-4-2-1, and also spent time in a 3-4-3 under Lampard. 

All these teams have the benefit of wing-backs, also included on these lists, and usually offer their centre backs a greater licence to go forward. 

Robinson has shifted Oxford towards a five/three at the back this season, but for 2020-21, it was 4-3-3 all the way, making their achievement to top the table for defensive goals even more impressive. 

It speaks volumes to Robinson's expressive style that Josh Ruffels and Sam Long provided the output of wing-backs from their more traditionally defensive full back positions. 

Goal contributions from all Oxford's defenders last season (all comps)

It is even more impressive considering neither Ruffels or Long would describe themselves as naturally progressive full backs, and their attacking play was built more on desire than any intrinsic attacking ability.

Oxford also don't have the benefit of playing a direct style, as many of the other teams on the list do. 

Much of Burton's game is based off set pieces and long balls into the box (their 17 set play goals this season is behind only Rotherham in League One), whilst Inter Milan, Carlisle and many others near the top are also reliant on set plays to varying extents. 

Whilst Oxford were highly efficient from corners and free kicks last year, their game remained based off building attacks through open play, and it is surprising that they were able to compete with some of the more aerially dominant sides in Europe in spite of this. 

Additionally, none of Oxford's defenders are set-piece specialists, or even set-piece takers. 

Whilst Ruffels experimented with some left-footed free-kicks last year, United's dead ball opportunities were reserved for their attacking options, with none of Oxford's defensive goals coming from direct set-plays.

Perhaps this season's drop off is a return to the natural order, but it is still pertinent that there has been a clear decrease in the number of goals scored from the back for Oxford. 

Oxford's eight goals from defenders this season puts them roughly mid-table in League One, and they are not close to challenging any European records this time round. 

Again, the trend of direct teams (Wycombe, Fleetwood, Burton) or teams favouring a back five (MK Dons) seems to stand up as they dominate the table. 

With this in mind, it does not seem Oxford's drop off has been particularly serious. 

However, considering the number of goals Oxford score (only Fulham have scored more than Oxford's 74 in the EFL this season), it is surprising that they feature so low down. 

Converting it into the percentage of goals scored by defenders, Oxford drop down the list into 16th in League One. 

If you are wondering whether Crewe's tally is a mistake, it is not. Crewe are yet to register a single goal from a defender, which may be some explanation as to why they are bottom of the league. 

10.81% of Oxford's goals have come from defenders this season, which is less than half of last season's total of 24.68%; a clear drop off. 

The table does suggest it is nothing for Oxford to be majorly concerned by. Sheffield Wednesday, Ipswich, Sunderland, Rotherham and Wigan all have a lower percentage of goals from defenders whilst simultaneously performing well in the league, and Oxford's league position doesn't seem to be massively influenced by their lack of defensive goals. 

It does represent a very sudden change for Oxford, though, and one which Robinson, who has always been a strong believer in his defenders getting forward, is surely not particularly pleased about. 

...

The reason for this drop off is less clear.

Oxford have scored considerably more goals per game this season (1.90) than last year (1.67), so it is mysterious that the defence appear to be dropping off as the rest of the side improve in front of goal. 

This season has seen a large increase in the number of goals scored from midfield which, at 32, is already more than triple last seasons tally with seven games still to play. 

Much of this will be down to Brannagan and Sykes, who are both enjoying their best campaigns in front of goal, but there is also a greater tactical emphasises on midfielders getting on the end of cut backs which has led to this increase in goals from the centre of the park.

Oxford have also spent much of this season with attackers playing in defence, too, which you would expect to bring more attacking output. 

As would Oxford's recent shift to a back five, which is shown to be conducive to plenty of defensive goals, yet they remain relatively rare for the Us. 

One explanation for this could be a struggle to make use of set pieces, which is a great opportunity for defenders to get on the scoresheet. 

Only eight teams have scored less set piece goals than Oxford (11) in League One this season.  

It means set plays make up 14.86% of Oxford's total goals compared to 23.38% of their goals last season. 

2020-21 also saw Oxford register the fourth highest Expected Goals (xG) created from set pieces in League One, whereas this season they rank 16th. 

The 126 shots created from set plays this season (3.23 per game) is also down from last season's 191 (4.15 per game). 

Overall, there has been a far greater reliance on open play chance creation over set pieces. 

The percentage of Oxford's xG coming from set pieces in 2020-21 was 29%, the sixth highest in League One, but that has dropped considerably to 16%, which is the lowest in the division. 

There has also been a drop off in individual performances, notably the attacking output of Sam Long. 

The 27-year-old shocked everybody with his bursts of pace and forward threat last season, registering a career high ten goal involvements in the process. 

A combination of injury, form and playing style has prevented Long from hitting those heights this time round. 

The loss of Gorrin and lack of a natural holding midfielder has no doubt also prohibited him from driving forward as much, and United's whole defensive unit may have suffered from the loss of the cover offered by the Spaniard. 

Long's crosses per game has decreased from 0.4 to 0.3 this season compared to last, whilst his key passes per game has also dropped from 0.9 to 0.6. 

Whilst both Long's shots and dribbles per game have surprisingly increased, he is playing considerably more long balls per game (3.3 compared to 2.7 last year) which is indicative of him hitting more forward balls into the front line and maintaining his defensive position, rather than progressing it forward as he perhaps did more often last year. 

This is a trend for the whole of Oxford's backline, averaging 0.95 more long balls per game than they did last season.

Average Long balls per 90 of all Oxford defenders (800+ mins) in 2020-21 vs 2021-22

These figures may reflect a slight change in Oxford's playing style: on numerous occasions, Robinson has shown a willingness to play the ball in behind or directly into the frontmen quicker than in previous years, and this has been particularly important in earning them points against the better sides in the division. 

A by-product of this, though, is that Oxford's defenders do not get into the box as much, as they are often charged with playing the balls in behind rather than getting on the end of them. 

This ties in with the additional goals from midfield, as they are no longer carrying so much of the creative burden, allowing them more freedom to get in the box when Oxford are building attacks. 

Conversely, last season Oxford's attacks were almost entirely built off wing play, and the full backs getting into the box was key to this. This slight tactical tweak from Robinson has allowed United to be more dangerous in transition against stronger opposition, but may have led to a drying up of goals for the full backs. 

It is also notable that Sam Long's change in fortunes has come in tandem with a change in his teammates at the back. 

The strangely prolific Josh Ruffels left for Huddersfield and was replaced with the strangely not prolific Steve Seddon.  

Rob Atkinson also departed to Bristol City, and Elliot Moore has been in and out of the team through injury. 

It all combines to somewhat justifying United's slip-up in defensive contribution to goals, and whilst they continue scoring for fun, it won't matter. 

But Ipswich proved things are not always that easy, and sometimes goals will need to come from sources you don't expect. 

McNally is providing them at the moment, but he needs more of his teammates to get involved if Oxford are to consistently get results from tight games between the big sides. 

Stats valid from time of writing (22/03/2022)