Stats review: Week 17 update

We start with the alternative Superliga tables before we go to our other story of the week, which will be about the quality of the goalkeepers.

Updated alternative tables:

We use the following measures:

TSR, Total Shots Ratio, is a team’s share of the total shots in their matches. This is defined as:

TSR=Total shots by team/Total shots by team and opponent

SoTR, Shots on Target Ratio, is a team’s share of the total shots on target in their matches. This is defined as:

SoTR=Total shots on target by team/Total shots on target by team and opponent

PDO, is a measure from hockey, that has been applied to football. Our use of it is to evaluate how lucky/unlucky a team has been during the season. PDO is defined as:
PDO=1000x(Saving%+Scoring%).

The table is sorted by SoTR, which is our favourite indicator for quality.

For detailed reviews and tables from previous weeks – click here.

Comments:

Frontrunners FC Midtjylland and AaB should be scared, because a group of hungry lions (FC Copenhagen) are chasing them. The reigning champions are the best by some margin on both vital measures of quality: TSR and SoTR. The only reason that you still find FC Midtjylland and AaB ahead is because of their better efficiency, highlighted by a PDO of more than 1050, which indicates that FC Midtjylland and AaB have had their fair share of luck. Some might argue it was the case in this weekend too, where both escaped with points despite being weaker than their opponents. Esbjerg are the new number two on the SoTR list, but a terrible saving percentage leaves Esbjerg at the bottom of actual table. Esbjerg should have room for improvement in the table; at least their chance creating abilities indicate so. FC Nordsjælland are a team on the rise and should have won this weekend against Randers (shots on target score line ended: 8-2). AGF had a lousy game against OB, while SønderjyskE improved since they met poor Viborg. SoTR now indicates that if SønderjyskE get some of the luck they have been lacking so far, AGF should be worried about relegation. Viborg should do all that they can to secure top scorer Thomas Dalgaard on a long-term deal. Without his efficiency Viborg would have been far below the relegation line and not in 7th position. We still believe that Viborg will drop once their luck/efficiency does as well; they are simply the weakest team in the league based on our stats.

Question: How much are the keepers to blame for a poor position?

We have claimed that the PDO was mainly a luck indicator. No club should be able to tally significantly more or less than a PDO 1000 over a longer period. We have however admitted that smaller differences can occur simply based on the quality of the strikers and the keepers. We wanted to test this, so we looked at the keepers. Is their any difference between the keepers with high saving percentage and low saving percentage, when it comes to the goals they let in? Although the sample is too small to draw strong inference, we decided to give at a go anyway.

We picked out the six goalkeepers that have played all seventeen matches this season. At the same time these goalkeepers are ranked pretty dispersed based on saving percentage with two from the top (Lössl, Hradecky), two from the middle (Peskovic, Mikkelsen) and two from the bottom (Skender, Rønnow). We have looked at every goal conceded this season and placed it into one of four groups:

Type 1: A big mistake – the goal is entirely to blame on the keeper

Type 2: Most keepers would have saved the shot

Type 3: A fantastic save could have stopped the shot

Type 4: Not even Peter Schmeichel in his heyday would have saved it

It must be admitted that these assessments are subjective and others are likely to evaluate the goals differently, so have this in mind. It could be fun to have a (former) professional goalkeeper evaluate the goals, but you have to settle with us this time.

Here are the results for each keeper. In parenthesis is the type of goal’s share of the total amount of goals. The cost of goals compared to Lössl is how many goals a mistake percentage (type 1 or 2 goals) like Lössl’s would have deducted from the total goals allowed.

Jonas Lössl, FC Midtjylland

Total saving percentage: 77.78% (rank: 1st)

Total goals allowed: 16

Type 1 goals allowed: 1 (6.25%)

Type 2 goals allowed: 1 (6.25%)

Type 3 goals allowed: 8 (50%)

Type 4 goals allowed: 6 (37.50%)

Type 1 or 2 percentage: 13%

Lukas Hradecky, Brøndby

Total saving percentage: 75.58% (rank: 3rd)

Total goals allowed: 21

Type 1 goals allowed: 1 (4.76%)

Type 2 goals allowed: 4 (19.05%)

Type 3 goals allowed: 11 (52.38%)

Type 4 goals allowed: 5(23.81%)

Type 1 or 2 percentage: 23.81%

Cost of goals compared to Lössl: 2

Michal Peskovic, Viborg

Total saving percentage: 74.56% (rank: 5th)

Total goals allowed: 29

Type 1 goals allowed: 0 (0%)

Type 2 goals allowed: 5 (17.24%)

Type 3 goals allowed: 17 (58.62%)

Type 4 goals allowed: 7(24.14%)

Type 1 or 2 percentage: 17.24%

Cost of goals compared to Lössl: 1

Thomas Mikkelsen, FC Vestsjælland

Total saving percentage: 71.59% (rank: 7th)

Total goals allowed: 25

Type 1 goals allowed: 0 (0%)

Type 2 goals allowed: 4 (16%)

Type 3 goals allowed: 11 (44%)

Type 4 goals allowed: 10(40%)

Type 1 or 2 percentage: 16%

Cost of goals compared to Lössl: 1

Marin Skender, SønderjyskE

Total saving percentage: 66.67% (rank: 11th)

Total goals allowed: 33

Type 1 goals allowed: 3 (9.09%)

Type 2 goals allowed: 5 (15.15%)

Type 3 goals allowed: 14 (42.43%)

Type 4 goals allowed: 11(33.33%)

Type 1 or 2 percentage: 24.24%

Cost of goals compared to Lössl: 4

Frederik Rønnow, Esbjerg

Total saving percentage: 64.00% (rank: 12th)

Total goals allowed: 27

Type 1 goals allowed: 1 (3.70%)

Type 2 goals allowed: 8 (29.63%)

Type 3 goals allowed: 11 (40.74%)

Type 4 goals allowed: 7(25.93%)

Type 1 or 2 percentage: 33.33%

Cost of goals compared to Lössl: 5

Lössl has done really well this autumn. In earlier seasons he has been criticized for poor keeping, but with only two goals allowed that can be labelled as poor keeping, he is a big reason for the success of FC Midtjylland. His total saving percentage of almost 78% is nearly 6% higher than the league average (72.06%). Although Lössl deserves heaps of praise, we would expect the PDO magic to work at some point and drag him closer to the average. If not, Lössl could be bound for bigger clubs outside Denmark.

Hradecky is the next guy in line. His mistake percentage (we define that as type 1 or 2 goals allowed) is 23.81%, which is more than 10% higher than Lössl’s (13%). Hradecky did not make the best introduction in Brøndby, where he committed several mistakes. In recent matches he has improved a lot though. If we subtract the two additional goals that mistakes by Hradecky has cost compared to Lössl, Hradecky ends up with a saving percentage of 77.91%, which is just as good as Lössl. An indication that the quality difference between Hradecky and Lössl is small.

Next up are Peskovic and Mikkelsen. Both have been doing a good job for their promoted clubs, Viborg and FC Vestsjælland, illustrated by the fact that none of them have made a big mistake (type 1). Compared to Lössl they have both let in a poor keeping goal too much and if you adjust for this, their saving percentages would have been 75.44% for Peskovic and 72.73% for Mikkelsen. So neither of them is as good as Lössl and Hradecky, when it comes to overall goalkeeping (but remember: small sample).

We go to the bottom, where SønderjyskE and Esbjerg are considerably poorer than the rest of the league when it comes to saving percentage. We first focus on Skender, the Croat, who got a lot of credit for his performance last season. His performance this season has not been particularly inspiring with three type 1 mistakes (most of the six goalkeepers in this analysis). We believe that Skender has allowed four goals too much compared to Lössl’s mistake percentage, which would have moved SønderjyskE from a goal difference of -17 to -13. A difference that probably would have been worth some points, although SønderjyskE’s scoring percentage and general SoTR are too poor to move them away from the bottom. Skender would have had, adjusted for the additional mistakes, a saving percentage of 70.71%. It is below the league average, so clearly not very impressive.

At last we go to Esbjerg. Esbjerg loaned youngster Frederik Rønnow for a season in Horsens, where he had been highly appreciated. He has been pointed out as a future keeper for the Danish national team, wherefore his poor saving percentage is surprising and concerning. The numbers from our research does not put him in better light. We record a mistake percentage of 33.33%, which compared to Lössl has had a cost of five goals. With these goals Esbjerg would have gone from a goal difference of +2 to +7 and probably moved into top six. It is especially Rønnow’s poor aerial ability that makes his record so appalling. If we adjust Rønnow’s saving percentage for additional mistakes compared to Lössl, it goes from 64% to 70.67%. It is a big improvement, but the level is still poor.

Although we have not analysed the keepers from the remaining six clubs we highly doubt that we will find poorer adjusted saving percentages than the ones for Skender and Rønnow. SønderjyskE and Esbjerg should strongly consider changing their goalkeeper in the winter transfer window in order to emerge from the relegation pit

With respect to the PDO, it could seem like there is a difference between a good and a bad keeper of around 7% in saving percentage (simply in form of more type 1 and 2 mistakes). A 7% difference in saving percentage is the same as a difference of 70 in PDO. Our conclusion is that even though the PDO should move towards 1000 points, a difference of up to 70 points could be persistent. And this is only for the keepers, an effect are likely to be observable for the strikers as well. But remember: small sample! It could be that the higher saving percentage of some keepers are due to weaker shots allowed by the defense, in this case the entire blame should not be put on the keeper, despite his weak saving percentage. We can however pick out Rønnow as the keeper that has performed the weakest this autumn, simply because he has the highest allowance of type 1 and 2 goals. These goals cannot be blamed on poor defending!

Efficiency table

Is there any better way to end this than with a look at the efficiency table?