A Total Shot Ratio analysis of Week 4 in the EPL

A Total Shot Ratio analysis of Week 4 in the EPL

By Mark Taylor Sep 17, 2014

Tweet

After four EPL games Swansea are 3rd and have taken the lead in all of their games. This article investigates the advantage of scoring first, and how teams like Swansea and Villa  – who have scored first in every game a goal has been scored in – are perhaps prone to being overvalued.

Week four of the Premier League saw a meeting between two sides with a previously 100% record. It wasn’t surprising to see Chelsea at the head of the fledgling table, but Swansea are typically a mid-table team and their three previous wins had each been achieved despite the Welsh side being out-shot in each game.

Saturday’s match therefore wasn’t a typical clash between the top two sides in the league and Chelsea continued their remarkable scoring streak in a 4-2 win.

Swansea also replicated their early season habits. They were again out-shot, this time by 30-13 and in keeping with each of their competitive games so far, they scored the game’s opening goal.

Unlike the previous three occasions, four if we include the Swans’ league cup tie with Rotherham, the opening goal wasn’t sufficient an advantage to help Swansea to a victory.

But both evidentially and intuitively, scoring the first goal is a tremendous advantage for a team and we can demonstrate this using a Poisson model of a typical Premier League match.

The impact of scoring first

It is well documented that a Poisson analysis of the goal scoring average for each side prior to kick off can be used to predict match outcomes – read how to calculate Poisson analysis here. For example, a match where the home team could be expected to score 1.5 goals, and the away team 1.1 goals on average, would result in the home team winning 46 times out of 100, a draw occurring 27 times and the first goal scored around the 37th minute.

In addition, we can chart the decay of each side’s goal expectation as time elapses and again use a Poisson approach to see how the win probabilities alter once a goal is scored, something extremely useful for live bettors.

If such an average home side scores under these circumstances, their chances of winning the game by opening the scoring after 37 minutes rise from 46% at kick off to about 74% immediately following the goal. The draw remains around 27%.

In short, a typical home team will take points from 91 out of 100 games if they score first after a typical 37 minutes. So scoring first in a low scoring sport, such as soccer is hugely influential.

Unsurprisingly, the best teams are more likely to score the opening goal in a game and there is a reasonably strong correlation between the share of goals a side is expected to score in a game and the proportion of first goals they claim in a contest.

For example, in 2008/09, Arsenal scored 65% of goals scored in their matches, and 64% of the opening goals.  Similarly, in the same season, Stoke scored 41% of the goals and 44% of the opening scores.

However, even over an entire season, random chance can lead to a team occasionally scoring the opening goal at rates that are out of sync with their overall scoring profiles. Between 2008 and 2011, Chelsea scored around 70% of the total match goals – involved in their games – and for the first two seasons the general relationship between overall scoring and scoring first in a game stood.

But despite scoring 68% of the goals in 2010/11, the rate at which they opened the scoring fell to 53% and this reduced their average league points per game – it had been in the previous seasons more than 2 points per game.

Newcastle’s sensational season in 2011/12 was fuelled by opening the scoring in over 60% of games, but had only a 53% share of the games goals. They went on to win 19 out of these 23 matches, a run that was unlikely to be sustained in the future.

This has implications for betting. Not only does an opening score tilt live odds heavily in favour of the side scoring first, but such an influential factor in deciding the outright result can also inflate a team’s results in the short term, especially if they run up a streak of opening scores.

Swansea have been dominated in terms of shot numbers in their three opening EPL wins, but have been fortunate to open the scoring in all three games. This has placed them in a position that favours a successful outcome, but is also a statistic that Swansea may not be able to control or influence to any large degree, especially in the short term.

When evaluating Swansea’s potential in the future it is wise for bettors to look at the basis for their previous, recent success and part of their 100% record which came to an end at Chelsea was down to a run of first goals.

The long term rates at which a side scores first rarely rises above 70% or falls much below 40% over a full 38 game season, so good or bad runs that are based on rates outside of these norms may be untrustworthy.

Another team that has been quick to score are Aston Villa, who now occupy Swansea’s vacated second spot and have scored first in all three of their EPL matches – where a goal has been scored.

Pundits are commentating on the positive effect Roy Keane has had on Villa’s defence, however bettors must be careful not to overplay his impact when assessing Villa’s apparent improvement and newfound defensive solidity, when chance and playing frequently with a lead is likely to have also played a role in their excellent start.

Likely match outcomes in Week Four of the EPL based on 10,000 shooting simulations

Home Team

Home Win %

Draw %

Away Win %

Away Team

Actual Score

Most Likely Score

Arsenal

37

24

39

Man City

2-2

1-1

Chelsea

73

13

14

Swansea

4-2

3-1

Crystal Palace

33

38

29

Burnley

0-0

1-1

Hull

30

28

42

West Ham

2-2

1-1

Liverpool

70

21

9

Aston Villa

0-1

1-0

Man Utd

73

17

10

QPR

4-0

2-0

Southampton

55

27

18

Newcastle

4-0

1-0

Stoke

73

19

8

Leicester

0-1

2-0

Sunderland

18

27

55

Tottenham

2-2

0-1

WBA

32

29

39

Everton

0-2

0-1

From the table above it is clear that most of the results were consistent with the respective shot numbers of both sides.

The two matches which saw the biggest disconnect between shots and final score involved Leicester and Aston Villa. Both sides were out-shot by their hosts, Stoke and Liverpool, respectively, but both claimed all three points.

The most likely current EPL position based on chances for & against

Position

Current Position

Likelihood of Top Five %

Upper Middle Five %

Lower Middle Five %

Bottom Five %

1

Chelsea

80

14

5

1

2

Aston Villa

15

24

34

26

3

Swansea

7

20

30

43

4

Southampton

41

32

20

7

5

Man City

25

35

26

14

6

Tottenham

5

26

35

34

7

Arsenal

71

17

9

3

8

Liverpool

63

27

8

2

9

Man Utd

39

32

20

8

10

Hull

4

13

32

51

11

Everton

10

22

32

37

12

Leicester

1

7

18

74

13

West Ham

28

39

23

11

14

Stoke

27

37

24

12

15

Sunderland

9

18

31

42

16

QPR

16

35

28

21

17

Crystal Palace

4

18

28

50

18

Burnley

9

26

35

30

19

WBA

26

29

30

15

20

Newcastle

28

28

23

19

In the table above, I’ve summarised the most likely current positions of each side based on the goal attempts they have faced and attempted so far this season. The table has been divided into convenient quarters to more easily demonstrate whether each team’s current position is in broad agreement or not with their shooting prowess to date.

Chelsea’s current position is consistent with their shooting ratios. They appear in the top five in 80% of the 10,000 simulations, although they are yet to be tested by a legitimate title challenger. That happens for the first time on Sunday, when Chelsea travel to play Manchester City, a side who have already faced two potential title rivals, in Liverpool and Arsenal.

Both Swansea and Villa, buoyed by scoring first, appear flattered by their lofty positions. Both have the chance creation profiles of lower half teams and bettors should consider betting against them, especially when they play opponents of similar strength.

The big five sides are all most likely to be found in the top half of the table, both in simulations and in reality, while Southampton, perhaps surprisingly given their summer departures also have underlying statistics which appear to justify their current lofty position.

The Saints do not appear to be riding on a wave of unsustainable positive events, unlike the two teams immediately above them in the table. And may be the most legitimate of the unfashionable contenders to chase home the big five sides. Their Saturday afternoon trip to the Liberty Stadium, therefore, has intriguing analytical undertones.

At the foot of the table, Newcastle are already under pressure to replace Alan Pardew, but they have conceded first in all games, apart from the goalless draw with Villa.. They are however almost certainly better than their current position appear to indicate and next welcome Hull, who may be flattered by their current 10th spot, but have played with a man less for around 20% of their season.

Click here for the latest Premier League odds.

Mark Taylor is a freelance soccer and NFL writer who, along with producing expert content for Pinnacle Sports, also runs his own soccer analytics blog, the Power of Goals.

If you have feedback, comments or questions regarding this article, please email the author or send us a tweet on Twitter.