A shot based appraisal for Week 1 of EPL games

A shot based appraisal for Week 1 of EPL games

By Mark Taylor Aug 22, 2014

Tweet

This soccer-betting article examines the opening fixtures of the EPL using a shot based analysis to identify which teams were lucky/unlucky. Read to understand why West Ham would have won 81% of the time against Tottenham based on this shot based analysis, highlighting how unfortunate they were to lose 1-0.

The use of shots rather than goals as an evaluation tool for soccer continues to rapidly evolve. Total Shot Ratio, for example, counts shot numbers to see which side may have dominated the match.

However, not all shots are created equally. A header from the edge of the box is generally less likely to result in a goal than a shot from the six-yard line. And so shot location models are becoming increasingly common, where historical data is used to assign the goal expectation for shots or headers based on where on the pitch the attempt originated.

It is then possible to run simulations based on shots to see if the actual result was a fair reflection of the type of goal attempts created by each team during the match. – Read this article, which explains how to calculate this process.

West Ham vs. Tottenham – an example of luck

It will be little consolation to West Ham fans, players and management, that their single goal defeat at home to Tottenham on the opening day of the season was extremely unlikely given the manner in which they dominated their visitors.

West Ham outshot Spurs by 18-10 and six of the home side’s goal attempts had at least a 10% chance of a goal, based on the location and type of attempt.

By contrast, the majority of Spurs’ efforts were from outside the box and therefore carried a low likelihood of scoring. Just one attempt from the visitors – Eric Dier’s 93rd minute winner – had a scoring expectation of greater than 10% based on its location.

In total, the cumulative goal expectation of West Ham’s 18 shots was 2.21 goals, while Spurs’ 10 shots carried a goal expectation just exceeding 0.5 of a goal.

Of course, statistical analysis should always be used to complement other forms of analysis and the missing ingredient in the match evaluation occurred in the 29th minute.

Red cards – read about the impact of cards here – on average benefit the side that retain 11 players. That side will, typically, concede fewer goals and score more than if no card was shown. The average time at which a red card occurs is just after 60 minutes. Similarly, a penalty kick presents an excellent opportunity to score, with conversion rates typically of around 78%.

These two high impact events are relatively rare and more often than not happen in favour of the better team. Superior teams tend to do more of the attacking, inviting more frequent challenges inside the box, resulting in more spot kicks and more opportunities for their opponents to be sent off.

Therefore, the red card shown to Spurs’ Kyle Naughton after 30 minutes, for conceding a penalty kick, presented West Ham with an unlikely, but excellent chance to overturn their status as outsiders at the start of the game.

Even when Mark Noble missed his penalty kick, the Hammers’ man advantage allowed them to create more frequent chances than Spurs, until James Collins received a second yellow card 34 minutes later to return the game to numerical parity.

A goal expectation of 2.21 to 0.51 goals hints at the shot superiority enjoyed by West Ham and by simulating the outcome of all 28 attempts 10,000 times we can generate score lines and ultimately how likely the Hammers were to win.

West Ham were fortunate to play against 10 men for over 30minutes, so their shot supremacy wasn’t necessarily an indication of possible improvement from last season or a decline from Spurs.

It was instead dictated by circumstances that are unlikely to be repeated with any regularity. However, the quality and quantity of the chances they created would see them win 81% of such games, draw 14% and lose just 5% of the time.

In the first, largely irrelevant, Premier League table of the season, the chances faced and created by West Ham, along with those of their 19 rivals spread across ten matches, were good enough to make West Ham the most likely table toppers – from 10,000 league simulations.

Over 20% of simulations resulted in West Ham leading the table on goal difference after week 1. However West ham’s failure to outscore, rather than just out shoot Spurs, sees them languishing in 18th position.

This highlights that despite most of the luck going West Ham’s way early on, and the odds favouring them to win after the early red card, they couldn’t capitalise on their superiority and ultimately paid the price as the pendulum of luck swung Tottenham’s way.

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

Home Team

Home Win %

Draw %

Away Win %

Away Team

Actual Score

Most Likely Score

Arsenal

73

21

6

C Palace

2-1

1-0

Burnley

14

26

60

Chelsea

1-3

0-2

Leicester

27

35

38

Everton

2-2

0-0

Liverpool

40

27

34

Southampton

2-1

1-1

Man Utd

55

31

14

Swansea

1-2

1-0

Newcastle

35

31

35

Man City

0-2

1-1

QPR

71

19

11

Hull

0-1

2-0

Stoke

28

39

33

Aston Villa

0-1

0-0

WBA

58

28

14

Sunderland

2-2

1-0

In the table above, we’ve listed the percentage of the three possible outcomes of each 10,000 simulations for all ten games played on the opening weekend of the EPL season based around goal attempts made by each side.

Hull share with Spurs the distinction of claiming an unlikely three points after being roundly out shot and also faced an unsuccessful penalty away at QPR, while Swansea made the most of their five goal attempts at Old Trafford compared to 14 attempts from their hosts.

In the Potteries, Stoke and Villa combined to shoot infrequently in the case of Villa and often from distance for the hosts, in a match where a draw wouldn’t have been unexpected.

Title hopefuls, Arsenal and Chelsea produced shot totals and shot quality which merited three points. Arsenal’s winner and each of Chelsea’s three goals were taken from excellent shot locations.

Both Liverpool and Manchester City were less impressive. City’s opponents, Newcastle, created attempts of similar frequency and location, but failed to test goalkeeper Joe Hart.

Liverpool vs. Southampton – Saints more than a match for Liverpool

The game at Anfield was intriguing as a Liverpool side without the departed Suarez hosted a Southampton team much changed from last season.

Twelve shots from each side implied a close match and despite eight shots from outside the Southampton box, compared to just four from distance from the visitors, Liverpool narrowly shaded the goal expectation by 1.25 to 1.21 goals – read this article to calculate a goal expectation based on shots. This was largely in relation to Liverpool creating three chances with goal expectations of 20% or greater, of which they scored their two goals.

Southampton appeared not to miss the departures and more than matched last season’s pace setters in terms of shooting volume. Despite leaving pointless, a shot based appraisal gives cautious optimism and the most likely positive Saints score line from the simulations was a 1-1 draw, occurring nearly 16% of the time.

At this stage of the season, records will be influenced by the quality of opponents.  Arsenal will find it easier to out shoot Crystal Palace than they will Manchester City. But as games and shots increase, this approach will begin to identify sides which may be benefitting or suffering from the results of unsustainable shot conversion rates, giving bettors a valuable tool in identifying teams who are over or underperforming.

Click here to see 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.