On The Defence

Here is the second in a series of quick snapshots of where we are player performance wise, focussing on the defending data. I usually go with 10 games or 900 minutes as an acceptable sample size for data. Clearly not all players have played that amount, so please take care as small sample sizes may render some statistics as outliers at this stage in the season.

A reminder that unless stated (e.g. it is a percentage), the data presented here is actions per 90 minutes.

Defensive Action Success Rate (DASR)

Jargon and lack of transparency are not my objectives so let me break down what DASR is. Essentially it is a % ratio between Successful and Unsuccessful defensive actions.

On the “successful” side of the equation is:

  • Tackles won, even if possession is not won
  • Aerial Challenges won, even if possession is not won
  • Interceptions
  • Shielding the ball out of play
  • Blocking a Cross
  • Blocking a Shot

On the “unsuccessful side”:

  • Tackles lost, even if possession won
  • Aerial Challenges lost, even if possession won
  • Missing a Block on a Cross
  • Miss a Tackle
  • Commit a Foul

Defensive statistics are traditionally harder to define and use as explanation for performance as much good defending is about avoiding anything bad happening. Thus, by aggregating a number of actions, it should tell us more about decision making and execution of defensive actions. It still doesn’t tell us what was avoided by being in the right position, for example. But the converse – Defensive Errors awarded for being in the wrong position – start to build the picture. As always, consideration of many data points helps to form a rounded view.

But back to DASR:

Head Boy heads the DASR! (Said no headline ever). Small sample alerts abound here. Ajer has played 360 minutes. Still, this is a hugely promising start – he just doesn’t lose his defensive actions – 1.75 per 90 mins is nearly 1 action per match better than the next lowest, new boy Ntcham.

Miller has started well but his data is from 2 appearances against Kilmarnock so “early days” warning applies.

Ralston also comes out well – he is robust for an 18-year-old and, given Miller’s small sample, arguably second in number of defensive actions won after Sviatchenko.

One can perhaps start to understand why Sviatchenko maybe falling down the pecking order. He wins an inordinately high number of challenges – nearly 3 per 90m more than anyone else. But also loses 3 more than anyone else. Contrast with Simunovic, with a healthy ratio 21% better than Sviatchenko. I don’t have DASR data for Boyata yet so that will be an interesting addition.

Tierney’s DASR data is disappointing for a defender but it is rare he plays as an out and out defender. Given his propensity to appear in attacking position, he does lose a lot of challenges as a winger would. Tierney’s overall performance, therefore, must be considered with reference to the creativity data that will follow later in the week.

Of the midfielder, no surprise to see Brown up there, but Ntcham is tucked in just behind, with only 2.68 defensive actions lost per 90m. Both Armstrong and McGregor could improve this aspect of their games. I would not judge Kouassi on 36 minutes.

A word on Bitton – he has been judged harshly on this site (and see defensive errors below) but he wins (and loses) more defensive actions than most other defenders.

No surprise that the “creatives” are at the bottom. Hayes is a standout and tends to lose the ball in the tackle more than anyone.

Personally, I like DASR – the list “feels” right – it passes the sniff test! Griffiths at the bottom, for example, as expected.

Possession Won / Lost

A slightly different take on the data is to look at overall Possessions Won and Lost. Possessions Won is the sum of:

  • Tackles won outright (i.e. tackle won and possession won)
  • Aerial Challenges won outright (i.e. Aerial Challenge won and possession won)
  • Blocked Cross and possession won
  • Shield Ball Out and possession won

Possessions Lost is the sum of:

  • Tackles lost outright (i.e. tackle lost and possession lost)
  • Aerial Challenges lost outright (i.e. Aerial Challenge lost and possession lost)
  • Miss controls leading to loss of possession

None of the players at the top of the table are amongst the highest appearance makers – Kouassi has only played 36 minutes. Indeed, you have to go down to Brown before a player has managed over 50% of the possible game time. So, small sample alert.

But we do like Ralston! With his remarkable resemblance to Stoke City manager and Welsh hard man centre forward Mark “Sparky” Hughes, there is something refreshingly old fashioned about him. It may appear a lazy simile but his playing style and approach IS very similar to Tierney. Not as confident attacking yet, he has started soundly defensively with no one else near his 7.55 possessions won per 90m.

Another to watch is Ajer. He rarely gives anything away and is a player who appears to have less confidence than his abilities suggest. Trusted in Champions League qualifiers, an axis of Lustig and Ajer is perhaps a safer stop gap until some/any regular centre back is fit.

Bitton, much maligned, but loses possession only 1.15 time per 90m.

Care needed on Miler’s data – when recording the games against Kilmarnock although he played very well, things did tend to “go his way” in challenges, so will watch his developing numbers with interest.

Of the midfielders, no surprise Brown is leading the possession stats, with Ntcham following closely. Armstrong and McGregor demonstrate how midfield balance is key as they tend to lose possession more than the more defensive midfielders – the selections in big games will be interesting.

No surprises with the bottom of the possession leagues although Hayes has the highest dispossession rate.

Finally, a note on Tierney. He spends that much time attacking he tends to lose possession as a winger would, and therefore ranks lowest of the “defenders”.

Defensive Errors

No one’s favourite topic but…..

It has been a tough start for Bitton in the sense that despite not being near a starting spot as a central midfielder, he has been asked to fill in as a centre back. In domestic football, where he has time to play out from the back, this has generally been ok although equalising goals could have been conceded at Partick Thistle and Kilmarnock due to Bitton switching off. In Europe, is was almost fatal in Astana. 1.34 per 90m is way above Ambrose alert levels. Hopefully he can get back to fighting for a space in the midfield.

Brown is surprisingly second. He is nearly at a bad error every game. A sign the legs are slowing? Mainly it is being caught in possession central and the odd poor pass back. Th aberration in Astana seemed to knock his confidence for the rest of the game. Just little signs the captain will need careful managing throughout the season.

Note a new feature. I calculate the xG that results from any defensive error. Some errors do not result in an opposition shot on target.

For completeness, I judge key defensive saves. Not too many of these due to dominating games, and the leader has only played 111 minutes.

Blockers

More new stats for this season include the recording of crosses blocked and missed, and shots blocked.

Blocked Shots

A wee bit of cheer for Bitton as he is the king of the shot blockers, and gets the early polka dot jersey!

Blocked Crosses and Missed Crosses

Probably don’t want the centre backs blocking crosses – suggests the cross has missed the first man and / or the centre back has been dragged out of the centre. So, more reassuring to see Ralston and Lustig figure. And indeed Tierney – nice of him to defend occasionally too!

Dribbled Past

Another new stat is to capture the players dribbled past when a player successfully completes a take on. For attacking players, I can show both the number of players take out the game with a run, and the number directly run past. Defensively, I record the number of times the player allows someone to run past them directly – i.e. without effecting a tackle or block.

Miller has played only 2 matches, but the fact he is being ran past 2 times per match may indicate some learnings needed about defensive body shape.

Armstrong has wonderful attacking stats, but defensively, he can be taken out perhaps too easily. Note Ntcham does not figure on the list at all – all players on 0 have not been shown. Brown also at 0.6 per 90m – not ideal for your sitting midfielder.

Clearances

Now when dealing with young players, I would not encourage the aimless punt up the park. At professional level, it is sometimes necessary to clear the lines to alleviate pressure. But although listed high to low, I would not characterise high as “good” and low as “bad”. The decision to punt long but aimlessly depends on whether you think that was the best decision, which is a matter of context.

Bitton’s position at the top may be due to his unfamiliarity with the centre back role and therefore taking a safety-first approach. Similarly, Ralston and Tierney as young players tend to play it safe if in doubt.

A lot of data to sift through but I hope this gives you a flavour of the data points I gather. By piecing together all the strands above you can build a comprehensive picture of each players defensive contribution.

I will update as the season progresses.

2 thoughts on “On The Defence

  1. […] Here is the second in a series of quick snapshots of where we are player performance wise, focussing on the defending data. I usually go with 10 games or 900 minutes as an acceptable sample size for data. Clearly not all players have played that amount, so please take care as small sample sizes may … Read More […]

Leave a response

Your email address will not be published. Required fields are marked *