Total Shots Ratio – TSR

Chris Smith discusses part of his trading strategy for this season; how Total Shots Ratio can be used to determine market value.

Categories: Data, Football, Professional, Statistical models, Tools

A basic total shots Ratio (TSR) has been the cornerstone of my attack and defence ratings this season. TSR assumes how much a team dominates a match. This has allowed me to produce more of comprehensive measurement of a team’s strengths and weaknesses. TSR = shots for / (shots for + shots against).

Looking at League 2 I have highlighted a TSR in italics if the top 12 team are above 0.5 and bottom 12 is below 0.49. Bold indicates where a team with a TSR of 0.50 or above is in the bottom 12 and a TSR of 0.49 of below is in the top ten.

Team P F A Tot TSR Pos
Chesterfield 30 12 8.5 20.5 0.58 1
Scunthorpe 30 10.5 10.4 20.9 0.5 2
Oxford 30 9.1 7.6 16.7 0.54 3
Rochdale 29 10.5 8.3 18.9 0.56 4
Burton 29 12.5 9.3 21.8 0.57 5
Fleetwood Town 28 13.4 9.5 22.9 0.59 6
Southend 29 10.4 10.6 21 0.5 7
Newport County 26 10.1 9.9 20 0.5 8
Dag and Red 30 9.4 11.9 21.3 0.44 9
Plymouth 29 10.8 9.8 20.6 0.52 10
Hartlepool 30 10.5 11 21.5 0.49 11
AFC Wimbledon 30 9.5 9.7 19.2 0.5 12
Morecambe 29 10.1 11.5 21.6 0.47 13
York 31 10.1 10 20.1 0.5 14
Cheltenham 29 10 9.1 19.2 0.52 15
Exeter 29 10.3 12.2 22.6 0.46 16
Mansfield 30 9.5 11.9 21.4 0.44 17
Accrington 28 8.1 11.7 19.8 0.41 18
Portsmouth 30 8.3 10.1 18.4 0.45 19
Bury 28 9.3 8.7 18 0.52 20
Bristol Rvs 29 10.3 10.5 20.9 0.5 21
Wycombe 29 8.4 7.8 16.2 0.52 22
Torquay 28 9.8 10.4 20.2 0.48 23
Northampton 28 8.6 11 19.6 0.44 24

Current League 2 table as of 10/02/2014

Based on the table Dagenham have overachieved so far this season from their TSR and therefore should be weaker in the betting compared those teams around them. Accrington had a disastrous start to the season but since October they have on excellence form. Based on their TSR they should be the weakest team in the league but this may be lower due to their extremely poor start. Bury and Cheltenham are two of the stronger teams who should be closer to the playoffs based on this table and their squads.

Sorting the table by TSR this is how the league would look.

Team

P  

F

A

Tot

TSR

Pos

Fleetwood Town

28

13.4

9.5

22.9

0.59

6

Chesterfield

30

12

8.5

20.5

0.58

1

Burton

29

12.5

9.3

21.8

0.57

5

Rochdale

29

10.5

8.3

18.9

0.56

4

Oxford

30

9.1

7.6

16.7

0.54

3

Plymouth

29

10.8

9.8

20.6

0.52

10

Cheltenham

29

10

9.1

19.2

0.52

15

Bury

28

9.3

8.7

18

0.52

20

Wycombe

29

8.4

7.8

16.2

0.52

22

Scunthorpe

30

10.5

10.4

20.9

0.5

2

Southend

29

10.4

10.6

21

0.5

7

Newport County

26

10.1

9.9

20

0.5

8

AFC Wimbledon

30

9.5

9.7

19.2

0.5

12

York

31

10.1

10

20.1

0.5

14

Bristol Rvs

29

10.3

10.5

20.9

0.5

21

Hartlepool

30

10.5

11

21.5

0.49

11

Torquay

28

9.8

10.4

20.2

0.48

23

Morecambe

29

10.1

11.5

21.6

0.47

13

Exeter

29

10.3

12.2

22.6

0.46

16

Portsmouth

30

8.3

10.1

18.4

0.45

19

Dag and Red

30

9.4

11.9

21.3

0.44

9

Mansfield

30

9.5

11.9

21.4

0.44

17

Northampton

28

8.6

11

19.6

0.44

24

Accrington

28

8.1

11.7

19.8

0.41

18

Five of the top six teams, with the highest TSR, are in top six positions in the league table. Scunthorpe in second place have 0.5 TSR, however, as the season has progressed they have improved. Wycombe produce so few shots their attack is one of the weakest in the division. They have only scored 0.95 goals on average in their last 20 matches. For this reason, their TSR and league position is acceptable.

Screen Shot 2014-03-06 at 08.00.27

The graph shows slight correlation to the league position and TSR.

League 2 is the tightest of all the English Professional divisions in every aspect. For example, points separating the teams, TSR and total goals. For this reason, I have difficulty backing a 0.5 jolly with high confidence in this division. For example, Newport vs. AFC Wimbledon on 11th February. The opening odds were a -0.5 line 2.0 / 1.92 on 16th February. On the 10th February the odds are now -0.5 2.11 / 1.82. FYI my supremacy is 0.42 on pure ratings, AFC Wimbledon have no less than six players missing, however.

Relative shot Ratio (RSR)

There is likely to be some under/over estimations based on previous performances. First I looked at how a team is performing against that average = (shots for – league average shots) + shots for. Then the same for shots against = (shots against – league average shots) + shots against.

For example, Accrington average 8.1 shots per match and concede 11.7 shots per match. Comparing this against an average team in the division an average team will have 10.06 shots per match and concede 10.05 shots per match.

Relative shots for: (8.1 – 10.05) + 8.1 = 6.14

Relative shots against: (11.7 – 10.06) + 11.7 = 13.34

RSR: 6.14/ (6.14 + 13.34) = 0.32

The table is now sorted on the RSR

Team

P  

F

RF

A

RA

RSR

TSR

Pos

Chesterfield

30

12

13.9375

8.5

6.941667

0.67

0.58

1

Fleetwood Town

28

13.4

16.7375

9.5

8.941667

0.65

0.59

6

Burton

29

12.5

14.9375

9.3

8.541667

0.64

0.57

5

Rochdale

29

10.5

10.9375

8.3

6.541667

0.63

0.56

4

Oxford

30

9.1

8.1375

7.6

5.141667

0.61

0.54

3

Cheltenham

29

10

9.9375

9.1

8.141667

0.55

0.52

15

Wycombe

29

8.4

6.7375

7.8

5.541667

0.55

0.52

22

Plymouth

29

10.8

11.5375

9.8

9.541667

0.55

0.52

10

Bury

28

9.3

8.5375

8.7

7.341667

0.54

0.52

20

Newport County

26

10.1

10.1375

9.9

9.741667

0.51

0.5

8

York

31

10.1

10.1375

10

9.941667

0.50

0.5

14

Scunthorpe

30

10.5

10.9375

10.4

10.74167

0.50

0.5

2

Southend

29

10.4

10.7375

10.6

11.14167

0.49

0.5

7

Bristol Rvs

29

10.3

10.5375

10.5

10.94167

0.49

0.5

21

AFC Wimbledon

30

9.5

8.9375

9.7

9.341667

0.49

0.5

12

Hartlepool

30

10.5

10.9375

11

11.94167

0.48

0.49

11

Torquay

28

9.8

9.5375

10.4

10.74167

0.47

0.48

23

Morecambe

29

10.1

10.1375

11.5

12.94167

0.44

0.47

13

Exeter

29

10.3

10.5375

12.2

14.34167

0.42

0.46

16

Mansfield

30

9.5

8.9375

11.9

13.74167

0.39

0.44

17

Portsmouth

30

8.3

6.5375

10.1

10.14167

0.39

0.45

19

Dag and Red

30

9.4

8.7375

11.9

13.74167

0.39

0.44

9

Northampton

28

8.6

7.1375

11

11.94167

0.37

0.44

24

Accrington

28

8.1

6.1375

11.7

13.34167

0.32

0.41

18

Screen Shot 2014-03-06 at 08.03.31

The overall RSR rankings are not much different to TSR. If I compare the RSR of Accrington and Chesterfield the ratio is 0.68:0.32. This looks more realistic if comparing their TSR of 0.58:0.41.

The drawbacks are is the quality of data, the quality of shots and where the shot was from. Also, some teams are set up to counter attack and therefore may have less shots as they undoubtedly soak up more pressure. Team news is an obvious one. Transfers and squad improvement or decline through injuries. The area I need to improve my work is now making sure I apply this correctly. I would be grateful if anyone can help with applying this to my attack and defence rating system.

About Chris Smith

Chris Smith began his career working with William Hill in Leeds before moving to London to work for WBX, an online betting exchange. He also gained experience with Sportingbet, Samvo Sport Radar and Skybet, within a range on trading, operations and supervisory roles. In the summer of 2013 he took the decision to begin trading his own book and is successfully building his trading account.
3 Thoughts on Total Shots Ratio – TSR
    Jonny
    13 Mar 2014
    12:42pm

    For articles on expected goal totals have a look at sotdoc.co.uk who highlight some games by keying the chance creation via recorded footage of games .

    Jonny
    13 Mar 2014
    12:46pm

    Hello Chris

    Where are you getting your shot data from ?

    The data provided by running ball is well off the actual shot and shot on target production in games

    Expected Goals v TSR(Total Shots Ratio) | The Ginger Analyst
    5 Jun 2014
    11:08am

    […] a blog piece by Chris Smith, the equation is TSR = shots for / (shots for + shots against) (Chris’ Blog, 2014). This does not have a squared value. In the piece, he uses the equation on league two teams. […]

Leave A Comment