Letting The Side Down

Players are human, and mistakes will occur. Clearly you want your players to make as few mistakes as possible, but they do happen so what has been the impact of defensive errors this season?

The Glossary is concentration and sound positioning personified.

Of all the data points I collect, Defensive errors are identified the most subjectively. The risk is applying judgement after the event – e.g. a goal was conceded so who is to blame? Football is a dynamic game with 22 player simultaneously moving as well as the ball. Players will be under instructions we don’t know about. Events happen that can impact decision making that we do not pick up – an extreme example happened in the Champions League Final where the Liverpool goalkeeper Karius made two high profile and game changing errors. But what was the impact of the blow to the head he received from Ramos prior to his first error that no one picked up on at the time?

As it is me doing the subjective judging there should at least be a level of consistency applied! I tend to look at the more obvious things like defensive positioning, miss controls, giving the ball away in central areas in your own half, failing to pick up players for marking.

On occasions there is no obvious individual to blame. Some goals or chances to the opposition are down to systemic failure – the team is not set up correctly (see various European thrashings – this is more a coaching issue). Or often 2, 3, 4 small errors or misfortunes concatenate into a major chance. And of course some goals are down simply to good play from the opposition.

As an aside I have considered collecting Offensive Errors – for example; missing a clear chance; failing to pass / shooting when a team mate is in a better position; a dangerous move ending due to bad control etc. Football is primarily about decision making and this would provide an insight into the forward-looking decisions. Lack of scoring goals due to bad decision making / execution is as worrying as errors leading to goals being conceded. I’ll think about this over the summer.

All that being said then, who has contributed the most defensive errors?

Defensive Errors Total

Seasoned readers will know that whilst interesting this does not provide an accurate picture as it is skewed towards those playing the most minutes. Brown has played the equivalent of 55.46 matches – nearly 11 more than the next Gordon on 44.5 and Lustig on 44.4.

We need to work pout the per 90 minute average number of defensive errors for that.

Bitton stands above the rest of the squad with 0.83 defensive errors per 90m. Not quite reaching the Ambrose bench mark of 1.16 defensive error per game, it is a standout for this season. Bitton was repurposed as a central defender earlier in the season. As I showed in Defensive Action Success Season Review and Possession Stats Confirm Ajer Most Effective Defender his defensive performance metrics were impressive. In addition his innate ability to play good accurate passes from deep made his deployment fruitful especially in domestic football where Celtic dominate possession and actual defending is rare. But not surprisingly, for someone trying to learn a new job, there were mistakes, and many of them were in European matches.

Bitton struggled in two ways predominantly. The first was positioning – aligning with the rest of the defence and being in the correct place to avoid danger. This is perhaps the most important aspect of defending as correct alignment mitigates most danger through avoidance as it constrains space and angles for forward passes. Secondly, he often seemed to “switch off” to use pundit vernacular. This usually manifest itself as a failure to react to the developing opposition attacking picture. Sometimes because he was ball watching not ball and man watching, as against Partick Thistle when Doolan was left with a relatively simple header in a tight game, but missed. More materially there were numerous errors of this kind away to Astana when a 5-0 lead was almost wiped out. Kilmarnock’s winning goal in February was another example.

Injury curtailed his season and it will be interesting to see what role Bitton plays in the squad for the new season.

Otherwise the error rate of the defenders and midfielders was in the 0.5 per 90m and below range. This is average to low, which is good!

Ajer has the lowest rate of errors of the central defenders at 0.4 per 90m. And De Vries the lowest of the goalkeepers at 0.21 per 90m. Although as we will see, we need to look at context to complete the picture.

Finally, I gently mock Tierney for the lack of defending he does. But he does not make many errors – the lowest of any of the defenders and midfielders at 0.13 per 90m. Tierney’s defending is of the old school “if in doubt kick it out” style. He has more Clearances (4.81 per 90m) than anyone else. But it is another reason why he is such a highly rated talent.

Defensive Error Impact

The number of errors in itself tells you nothing about the impact of them. So, in one of the many improvements I made to the data collection this season, I record whether the error lead to an opposition chance, and the Expected Goals (xG) of that chance. We can therefore quantify the impact of the errors.

Firstly, how many errors lead to chances? Overall 52% of defensive errors led to an opposition shot at goal – 77 of 147.

Here we can see perhaps those that were unlucky – Simunovic, Ntcham, Tierney and Gordon all saw their errors lead to shots at goal well over the 52% average. Conversely Lustig, Bain, Armstrong and Hendry saw very few of their errors turn into opposition shots.

Although Ajer and Bitton are around the average, it is noticeable that Ajer’s errors have tended to come against lower placed domestic opposition and he has largely avoided errors in big games. Whilst Bitton’s errors have often been in the high profile European games.

It is one thing giving up a shot but how many errors resulted in opposition goals? In total 32 of the 54 goals Celtic conceded can be traced back to an individual error. They are attributed as follows:

Once again to give that context we need to look at the xG from the chances given by errors to get a truer picture not skewed by number of minutes played.

This shows the difference between the xG of the chances resulting from defensive errors compared to actual goals conceded. Again Simunovic seems to have been very unlucky in that his errors have been severely punished. Expected to concede 1.235 goals from his errors, Celtic actually conceded 5.

Gordon had the highest xG from errors of 2.087 but conceded 4. Boyata, Brown, Lustig and Bitton were all similarly punished with the ultimate sanction from their errors in manner exceeding the xG projection.

On the other hand Ajer’s great form has been helped in that his errors have gone punished to a lesser extent with only 1 goals conceded compared to a xG of 0.645. Hendry has avoided any goals being conceded but in his relatively few minutes had one of the higher errors per 90m rates. This is another yellow flag for me on his performances as discussed in the other defensive review articles.

Finally, I can show the average xG from the errors committed. This shows whose errors are potentially most damaging irrespective of the actual outcome. This is a more reliable predictor of future performance.

This shows the average xG from defensive errors per 90m. I would expect Gordon to be high in this list as obviously goalkeeper errors usually occur closer to the goal and with less chance of covering defenders. I’d have expected the central defenders to be behind the goalkeeper using the same logic I applied to Gordon.

However ahead of Gordon is Lustig whose errors expose Celtic to 0.252 xG per 90m. Or, once every 4 games a Lustig error will cost Celtic a goal. That sounds bad. And/or very unlucky.

Ajer again has the lowest xG of 0.108 per 90m from defensive errors of the central defenders. And Tierney the lowest of all the defenders at 0.086.

Team Impact

21 out of 83 defensive errors have occurred in European matches. That’s 25% of the errors in 23% of the games. 11 goals resulted.

28 of the 83 defensive errors occurred against the other top 6 sides in the SPFL, leaving 34 against the rest of the Scottish opposition.

Overall Celtic concede 0.78 goals per game in Europe from defensive errors and 0.44 goals per game in domestic football. Clearly better teams will force more errors and be more likely to capitalise on them with better chances and goals. These are usually tight games so every goal counts.

Opposition Errors

I’ve logged 33 goals for Celtic from 126 as attributable to defensive errors – that’s just 26% of opposition goals. This suggests I am less critical of other teams’ defences than of the Celtic defence. Additionally, my Celtic bias means that I attribute Celtic goals to good play rather than search for fault. So those 33 must have been really bad errors!

Here is the breakdown by team, by Expected Goals and actual goals conceded.

Hearts, perhaps surprisingly, offered up the most inviting chances and Celtic punished them 3 times. The Rangers conceded 8 goals attributable to defensive errors although Celtic did score rather a lot of goals against them. Motherwell did not enjoy their matches against Celtic over the season much either.

Summary

Most goals have some element of defensive errors associated. It can depend on how critical you want to be about analysing the goals scenario. However, at the individual level there are clear patterns. Most Celtic players (defenders and midfielders) are performing around the average that is emerging as I collect more historical data (four seasons now) of around 0,5 defensive errors per 90m. There is great variation in the quality of chances that result from these, and whether goals are scored. This can be a matter of luck although some players do seem to concede more errors in the bigger games.

Bitton is the only individual that stands out with an error rate well over the average. He has been adapting to a new central defensive position that has seen him perform excellently across several defensive metrics, but with a high error rate, especially in European games. In tight games against better opposition it is a high-risk strategy to adopt.