Season Review 2019/20

Welcome

Welcome to Stats Perform’s English Premier League season review for the 2019/20 season.

Interactive and showcasing a host of detailed performance metrics, this report provides insight into the league’s standout performers, applying innovative frameworks produced by our team of AI scientists.

Within this review, we share a comprehensive breakdown of performance at the key ends of the pitch, as well as sharing detailed insight on team style, both in and out of possession. Detailed player analysis also features, with key metrics ranked across different positions.

Notable additions to our reviews this summer include details on the teams most effective at generating opportunities from high turnovers, together with insights into how each team approached the changes to the goal kick rule. We also apply metrics to highlight the ball carrying players who were effective at generating goalscoring opportunities through running with the ball.

These new features reflect Stats Perform’s ongoing commitment to further explore how performance data can inform a club’s decision-making across performance analysis, recruitment and long-term strategic planning.

We hope you find some interesting insights from this review.

Team

Expected Goals (For)

Overview

Key Points:

• Ranked by league position, this table outlines teams’ performances in front of goal, from both open play and set piece situations.

• Manchester City’s ability to generate high quality chances from open play was demonstrated by amassing an xG output of 72.2 from 575 shots. They exceeded the xG output of champions Liverpool by nearly 18, who ranked second in the metric having generated an xG of 54.8.

• 36% of Burnley’s xG output came from set pieces, the highest ratio in the league. Tottenham sit at the other end of the spectrum, having generated 89% of their xG output from open play.

• Exploring overperformance against these metrics, five clubs scored at least two more goals than expected from set pieces, and in contrast eight clubs underperformed by at least two goals. The most notable of these were Watford, who perhaps would have expected to score nearly six more goals from the chances they generated from set plays. The Hornets also underperformed from open play, creating chances worth an xG of 32.7 but only found the net on 24 occasions.

Expected Goals For

		Set Play : Total	Open Play			Set Play
	Team	xG Ratio	Shots	xG	Goals	SP Shots	SP xG	SP Goals
	Liverpool	0.16	450	54.8	61	136	11.0	17
	Manchester City	0.14	575	72.2	77	159	13.2	17
	Manchester United	0.14	416	42.6	47	114	9.1	8
	Chelsea	0.17	459	53.5	51	159	12.1	11
	Leicester City	0.21	398	44.4	53	135	13.3	7
	Tottenham Hotspur	0.11	346	39.8	45	94	5.5	8
	Wolverhampton Wanderers	0.21	337	40.1	34	120	11.4	11
	Arsenal	0.19	292	36.3	41	111	9.0	12
	Sheffield United	0.25	254	32.1	29	98	11.0	6
	Burnley	0.36	256	28.7	25	128	17.4	13
	Southampton	0.25	346	38.8	38	148	14.5	11
	Everton	0.30	297	35.4	30	171	15.7	11
	Newcastle United	0.25	288	26.0	24	110	9.1	14
	Crystal Palace	0.28	249	23.9	19	124	10.0	7
	Brighton and Hove Albion	0.29	333	31.3	22	127	13.4	12
	West Ham United	0.28	288	32.3	31	125	14.1	14
	Aston Villa	0.30	323	29.3	24	132	13.4	15
	Bournemouth	0.31	235	28.4	19	150	14.1	15
	Watford	0.21	295	32.7	24	121	10.5	4
	Norwich City	0.21	302	27.4	20	105	7.6	3

Graphic

Definitions

Metric	Definition
Set Play	Chances occuring as a result of a corner, direct free kick, indirect free kick or throw-in.
Set Play : Total xG Ratio	The proportion of a team’s total xG that resulted from set plays.
Expected Goals (xG)	Expected Goals (xG) measures the quality of a shot based on several variables such as assist type, shot angle and distance from goal, whether it was a headed shot and whether it was defined as a big chance. Adding up a player or team’s expected goals can give us an indication of how many goals a player or team should have scored on average, given the shots they have taken.

Expected Goals (Against)

Overview

Key Points:

• Of the teams to qualify for the Champions League, only Chelsea conceded more goals from open play than their xG total would suggest. The Blues conceded chances at a similar quality level to Manchester United, who recorded the lowest output in the league, yet conceded 14 more goals than Ole Gunnar Solskjær’s men.

• For the second successive season, Newcastle’s xG conceded from set pieces was the highest in the league. The Magpies conceded 98 more shots than Manchester City, who recorded the Premier League’s lowest output.

• Nine teams conceded fewer goals than their xG output, with Crystal Palace being notable for conceding nearly eight goals fewer than xG would project. Together with Leicester, Tottenham and Sheffield United conceded the fewest goals from set pieces, however based on the quality of chances both teams conceded, we would have perhaps expected them to have conceded up to 50% more set piece goals during the campaign.

Expected Goals Against

		Set Play : Total	Open Play			Set Play
	Team	xG Ratio	Shots	xG	Goals	SP Shots	SP xG	SP Goals
	Liverpool	0.18	264	30.5	24	77	6.9	7
	Manchester City	0.17	203	29.0	24	74	6.2	7
	Manchester United	0.25	269	26.0	21	117	9.6	11
	Chelsea	0.29	202	26.1	35	118	11.3	14
	Leicester City	0.21	260	28.4	25	108	9.8	6
	Tottenham Hotspur	0.23	386	37.5	34	146	12.5	6
	Wolverhampton Wanderers	0.21	284	28.1	29	112	7.9	9
	Arsenal	0.24	387	36.5	26	158	13.4	15
	Sheffield United	0.24	298	36.5	32	130	12.1	6
	Burnley	0.21	389	36.2	38	145	10.5	7
	Southampton	0.20	324	39.8	45	124	10.5	11
	Everton	0.24	305	34.1	35	118	11.6	15
	Newcastle United	0.28	392	46.5	39	172	19.0	15
	Crystal Palace	0.29	363	40.5	39	148	16.7	9
	Brighton and Hove Albion	0.24	374	41.0	40	122	13.6	10
	West Ham United	0.17	368	48.7	44	127	11.1	10
	Aston Villa	0.21	440	49.2	46	162	14.8	15
	Bournemouth	0.17	436	48.1	50	117	10.5	10
	Watford	0.20	358	40.2	37	129	12.0	16
	Norwich City	0.23	437	47.0	51	163	15.3	17

Graphic

Definitions

Metric	Definition
Set Play	Chances occuring as a result of a corner, direct free kick, indirect free kick or throw-in.
Set Play : Total xG Ratio	The proportion of a team’s total xG that resulted from set plays.
Expected Goals (xG)	Expected Goals (xG) measures the quality of a shot based on several variables such as assist type, shot angle and distance from goal, whether it was a headed shot and whether it was defined as a big chance. Adding up a player or team’s expected goals can give us an indication of how many goals a player or team should have scored on average, given the shots they have taken.

Team Sequences (Style)

Overview

Key Points:

• In this section we explore team style using the Stats Perform sequence framework. Using sequence time and passes per sequence, we can assess a team’s approach in terms of how they move the ball. Using direct speed, we are able to identify who progresses the ball quickly.

• In relation to longer sequences, Manchester City are the Premier League’s standout team, with an average sequence time over three seconds longer then the next highest team, Liverpool, whilst recording over 100 more attacks derived from build-up sequences than the champions.

• At the other end of the table, Norwich City are the outlier. Amongst the bottom five teams, they were the only team to average more than three passes per sequence and their direct speed, 1.26 m/s, was the slowest in the league.

• According to this metric, Bournemouth were the team that progressed the ball quickest this season. Burnley’s direct style is also illustrated through sequences, with the Clarets recording the fewest passes per sequence and the shortest average sequence time.

Open Play Sequences

		Sequence Summaries			Sequence Styles
	Team	Sequence Time	Passes Per Sequence	Direct Speed	10+ Pass OP Sequences	Build Up Attacks	Direct Attacks
	Liverpool	11.04	4.28	1.43	682	148	80
	Manchester City	14.78	5.41	1.28	919	251	56
	Manchester United	10.30	3.85	1.45	471	111	82
	Chelsea	11.09	4.27	1.40	645	134	58
	Leicester City	9.81	3.80	1.34	485	105	72
	Tottenham Hotspur	9.39	3.54	1.42	420	73	72
	Wolverhampton Wanderers	8.71	3.24	1.55	326	68	76
	Arsenal	10.70	3.86	1.40	440	92	71
	Sheffield United	7.34	2.93	1.33	203	34	46
	Burnley	6.26	2.55	1.60	140	23	28
	Southampton	7.18	2.79	1.35	252	51	56
	Everton	8.27	3.10	1.61	291	62	56
	Newcastle United	7.15	2.70	1.59	209	32	56
	Crystal Palace	7.47	2.84	1.58	221	38	53
	Brighton and Hove Albion	9.47	3.52	1.48	377	72	55
	West Ham United	7.65	2.92	1.54	228	38	62
	Aston Villa	7.75	2.84	1.55	171	36	52
	Bournemouth	7.47	2.83	1.64	199	35	57
	Watford	6.90	2.69	1.61	186	35	50
	Norwich City	9.24	3.40	1.26	320	67	41

Graphic

Definitions

Metric	Definition
Sequences	Sequences are defined as passages of play which belong to one team and are ended by defensive actions, stoppages in play or a shot.
Possessions	Possessions are defined as one or more sequences in a row belonging to the same team. A series of passes leading to a shot which is saved and results in a corner kick would comprise one possession since the same team retains control, but more than one sequence, since the ball has gone out of play. A possession is ended by the opposition gaining control of the ball.
Sequence Time	The average time (in seconds) per sequence.
Passes per Sequence	The average number of passes per sequence.
Direct Speed	A measure of how quickly a team progresses the ball upfield (metres/second).
10+ Pass OP Sequences	The number of open play sequences that includes 10 or more passes.
Build Up Attacks	The number of open play sequences that contains 10 or more passes and either ends in a shot or has at least one touch in the box.
Direct Attacks	The number of open play sequences that starts just inside the team’s own half and has at least 50% of movement towards the opposition’s goal and ends in a shot or a touch in the opposition box.

Team Sequences (Pressure)

Overview

Key Points:

• The sequence framework can also be applied to assess a team’s approach out of possession. We can understand where on the pitch a team disrupts a sequence, and where they win the ball back.

• Southampton’s aggressive approach to pressing, which is reflected in their PPDA score of 10 (ranking them second in the league), paid dividends in terms of goals scored as a result of high turnovers. No other team scored more than the Saints’ total of seven.

• Newcastle were one team who did not adopt a pressing approach, allowing their opponents plenty of the ball and committing the lowest number of high turnovers in the Premier League. However unlike Everton and Crystal Palace, the Magpies did manage to score once from a turnover within 40 metres of the opposition goal.

• Selecting the ‘Graphic’ tab, we can see the high turnover pitch map of every Premier League team.

Sequence Pressure Stats

			High Turnovers
	Team	PPDA	Total	Shot Ending	Goal Ending
	Liverpool	10.3	252	52	5
	Manchester City	10.1	254	65	6
	Manchester United	11.1	146	33	4
	Chelsea	10.6	174	35	4
	Leicester City	9.5	171	38	5
	Tottenham Hotspur	12.2	144	27	5
	Wolverhampton Wanderers	15.0	112	26	4
	Arsenal	12.5	138	24	6
	Sheffield United	15.7	170	19	1
	Burnley	13.4	153	32	2
	Southampton	10.0	201	39	7
	Everton	11.8	153	18	0
	Newcastle United	19.5	106	21	1
	Crystal Palace	14.7	141	24	0
	Brighton and Hove Albion	11.6	134	21	2
	West Ham United	14.6	130	21	1
	Aston Villa	13.8	137	19	1
	Bournemouth	14.4	139	24	2
	Watford	13.8	177	37	5
	Norwich City	13.9	122	27	2

Graphic

Liverpool

Manchester City

Manchester United

Chelsea

Leicester City

Tottenham Hotspur

Wolverhampton Wanderers

Arsenal

Sheffield United

Burnley

Southampton

Everton

Newcastle United

Crystal Palace

Brighton and Hove Albion

West Ham United

Aston Villa

Bournemouth

Watford

Norwich City

Definitions

Metric	Definition
High Turnovers	The number of sequences that start in open play and begin 40m or less from the opponent’s goal.
Shot Ending High Turnovers	The number of shot-ending sequences that start in open play and begin 40m or less from the opponent’s goal.
Goal Ending High Turnovers	The number of goal-ending sequences that start in open play and begin 40m or less from the opponent’s goal.
PPDA	PPDA is the number of opposition passes allowed outside of the pressing team’s own defensive third, divided by the number of defensive actions by the pressing team outside of their own defensive third. A lower figure indicates a higher level of pressing, while a higher figure indicates a lower level of pressing.

Goal Kicks

Overview

Key Points:

• Following the introduction of the new goal kick rule, we can establish which teams have used the rule change to build-up attacks from the back.

• Manchester United recorded the highest proportion of goal kicks which ended inside their own box (60.1%), a tactic which resulted them gaining more territory compared to when they went long. United averaged nearly 50 metres ball progression upfieldwhen they went short, compared to 37.5 metres when a goal kick cleared their own box.

• Brighton were another team who regularly utilised short goal kicks. However, unlike United, they were more successful in progressing the ball when they went long, consistently going down the left hand side when they took this approach.

• Eight teams elected to clear their own box from goal kicks on over 90% of occasions, with Sheffield United not taking a single goal kick ending in their own box over the whole season.

• Selecting the ‘Graphic’ tab, we can see the end location of every goal kick taken by the goalkeepers for each club during the season.

New Goal Kick Rule: Who’s using it?

		Goal Kick End Location			Goal Kick Upfield Progression
	Team	In the box	Outside the box	% ending in the box	In the box (m)	Outside the box (m)
	Liverpool	98	141	41.0	53.6	50.0
	Manchester City	87	80	52.1	64.4	49.9
	Manchester United	161	107	60.1	49.6	37.5
	Chelsea	117	109	51.8	53.5	42.1
	Leicester City	96	145	39.8	51.2	43.1
	Tottenham Hotspur	108	184	37.0	42.9	43.3
	Wolverhampton Wanderers	10	273	3.5	34.2	45.4
	Arsenal	140	174	44.6	42.6	37.1
	Sheffield United	0	272	0.0	0.0	43.6
	Burnley	3	301	1.0	45.7	44.5
	Southampton	25	230	9.8	56.7	40.0
	Everton	96	179	34.9	47.8	43.4
	Newcastle United	15	325	4.4	54.7	43.6
	Crystal Palace	24	289	7.7	48.5	42.4
	Brighton and Hove Albion	146	128	53.3	45.3	49.9
	West Ham United	13	278	4.5	43.2	47.0
	Aston Villa	63	280	18.4	42.4	42.1
	Bournemouth	63	232	21.4	44.3	44.8
	Watford	9	286	3.1	62.4	40.6
	Norwich City	103	246	29.5	46.1	41.8

Graphic

Liverpool

Manchester City

Manchester United

Chelsea

Leicester City

Tottenham Hotspur

Wolverhampton Wanderers

Arsenal

Sheffield United

Burnley

Southampton

Everton

Newcastle United

Crystal Palace

Brighton and Hove Albion

West Ham United

Aston Villa

Bournemouth

Watford

Norwich City

Definitions

Metric	Definition
Total In Box Goal Kicks	The total number of a team’s goal kicks that ended inside the box
Total Out of Box Goal Kicks	The total number of a team’s goal kicks that ended outside the box
Average In Box Goal Kick Progression	Average distance in metres upfield a team reached while in control of possession following a team’s goal kick ending outside the box. Example (controlled possession): If Mustafi receives the ball from the goal kick inside his area and then attempts an unsuccessful long ball forwards that goes out for a throw in, then the progression was only controlled to the point of where he was in control of the ball in his own box.
Average Out of Box Goal Kick Progression	Average distance in metres upfield a team reached while in control of possession following a team’s goal kick ending outside the box.

Player

Expected Goals

Overview

Key Points:

• Ranked by xG per 90, this table illustrates who took the highest quality shooting opportunities throughout the season.

• In a season hampered by injuries, Sergio Agüero posted the highest xG per 90 output. He was the only player in the league’s top three to exceed his xG per 90, in both xGOT and goals scored, reinforcing the quality of his finishing.

• West Ham’s Michael Antonio was another player who managed to find high quality goalscoring locations, ranking fourth in the league for xG. Although he was unable to match his xG in terms of actual goals, his xGOT output, which marginally exceeded his xG, indicates he was somewhat unfortunate not to have scored more.

• From this list, Chris Wood scored the highest proportion of goals with his head, with 36% of the New Zealander’s 14 goals coming from headers inside the box.

Top 10 Players Expected Goals Per 90 Minutes

			Per 90 Minutes			Goal Method %				Goal Loc. %
Team	Player	Minutes Played	xG	xGOT	Goals	Header	Left Foot	Right Foot	Other	Inside The Box
	Sergio Agüero	1456	0.90	0.92	0.99	13%	13%	75%	0%	88%
	Gabriel Jesus	2027	0.82	0.76	0.62	14%	36%	50%	0%	100%
	Tammy Abraham	2221	0.68	0.52	0.61	20%	0%	80%	0%	93%
	Michail Antonio	1770	0.67	0.68	0.51	20%	10%	70%	0%	100%
	Marcus Rashford	2653	0.65	0.65	0.58	6%	18%	76%	0%	94%
	Raheem Sterling	2660	0.64	0.60	0.68	15%	30%	55%	0%	95%
	Mohamed Salah	2884	0.60	0.66	0.59	11%	79%	11%	0%	100%
	Jamie Vardy	3034	0.58	0.61	0.68	13%	39%	48%	0%	96%
	Chris Wood	2444	0.58	0.54	0.52	36%	36%	29%	0%	100%
	Danny Ings	2812	0.52	0.55	0.70	0%	36%	59%	5%	82%

Pressure & Clarity

Key Points:

• This section outlines these players’ shooting habits in terms of shot pressure and shot clarity.

• Recording the highest xG per shot output in this list, Chris Wood demonstrated an ability to get into good scoring locations inside the box. Over 40% of the Burnley forward’s attempts came from high clarity situations, with a third of his shots also occurring when he was under low pressure.

• Although Tammy Abraham was unable to match his xG for the season, we can establish that 50% of his shots occurred when under high pressure, meaning that opposition players were within tackling distance when he was looking to shoot.

• The Premier League’s top goalscorer, Jamie Vardy, again demonstrated good decision making in front of goal, which is reflected in his high xG per shot output. He also attempted the lowest proportion of attempts from low clarity situations from the players listed here.

Shot Pressure & Clarity Distribution

			Pressure Level %			Clarity Level %
Team	Player	xG per shot	High Pressure	Moderate Pressure	Low Pressure	Open Goal	High Clarity	Moderate Clarity	Low Clarity
	Sergio Agüero	0.19	27.6	43.4	28.9	1.3	30.3	59.2	9.2
	Gabriel Jesus	0.18	38.6	34.7	26.7	1.0	30.7	63.4	5.0
	Tammy Abraham	0.19	50.0	28.4	21.6	1.1	29.5	65.9	3.4
	Michail Antonio	0.19	45.6	39.7	14.7	0.0	30.9	66.2	2.9
	Marcus Rashford	0.20	18.9	31.6	49.5	1.1	25.3	55.8	17.9
	Raheem Sterling	0.19	26.3	39.4	34.3	1.0	28.3	58.6	12.1
	Mohamed Salah	0.15	22.7	41.7	35.6	0.8	28.8	57.6	12.9
	Jamie Vardy	0.22	37.1	38.2	24.7	2.2	36.0	59.6	2.2
	Chris Wood	0.24	36.9	29.2	33.8	0.0	41.5	52.3	6.2
	Danny Ings	0.18	34.4	33.3	32.3	1.1	21.5	68.8	8.6

Graphic

Sergio Agüero

Gabriel Jesus

Tammy Abraham

Michail Antonio

Marcus Rashford

Raheem Sterling

Mohamed Salah

Jamie Vardy

Chris Wood

Danny Ings

Definitions

Metric	Definition
Expected Goals (xG)	Expected Goals (xG) measures the quality of a shot based on several variables such as assist type, shot angle and distance from goal, whether it was a headed shot and whether it was defined as a big chance. Adding up a player or team’s expected goals can give us an indication of how many goals a player or team should have scored on average, given the shots they have taken.
Expected Goals On Target (xGOT)	Expected Goals on Target is a separate Expected Goals Model that includes the original xG of the shot, and the goalmouth location where the shot ended up. It gives more credit to shots that end up in the corners, vs shots that go straight down the middle, and is built on historical on-target shots.
Shot Pressure	Shot pressure is judged by the amount of pressure a player is under by opposition players when they shoot at goal. At the time of the shot, the opposition players must be moving towards the ball (even if only slightly) or trying to put the shooter off in order to add pressure on the shooter.
Shot Clarity	Shot clarity assesses the line of sight between a player and the goal as a shot is taken. It is determined by the number of players (opposition and own team included) obstructing the clarity of the ball’s path to goal.

Expected Assists

Overview

Key Points:

• Understanding that all assists cannot be measured in the same way (a through ball putting the striker one-on-one against a goalkeeper is not the same as a laying the ball off 30 yards from goal for a team mate), the expected assists metric can support in identifying the players who have made high quality passes throughout the season, and have contributed to their team’s chance creation.

• Kevin De Bruyne equalled the Premier League record for most assists in a single campaign and it will come as no surprise that the Belgian recorded the highest xA (0.41) and chances created per 90 (2.8) in this list.

• Adama Traore is the highest ranking player unattached to Manchester City amongst the league’s top ten. Playing in a team completing fewer passes per game than the league runners-up, it is noteworthy that the Wolves winger created the most chances per 100 passes of all the players listed.

• With five Manchester City players featuring in the overall top ten, team styles (and player position) can clearly influence the rankings here. Filtering by position allows us to better understand which players from across the pitch are creating high quality opportunities for their team-mates.

Overall

Top 10 Overall Open Play Expected Assists Per 90 Minutes

			Per 90 Minutes		Chances Created
Team	Player	Minutes Played	xA	Assists	Per 90 Minutes	Per 100 Pass
	Kevin De Bruyne	2798	0.41	0.55	2.80	4.78
	Riyad Mahrez	1940	0.28	0.37	1.95	3.94
	David Silva	1832	0.27	0.49	1.87	2.66
	Bernardo Silva	2029	0.27	0.22	1.95	3.28
	Adama Traoré	2606	0.22	0.28	1.35	5.46
	Ross Barkley	1104	0.21	0.33	1.55	3.13
	Mohamed Salah	2884	0.21	0.22	1.50	5.09
	Theo Walcott	1297	0.21	0.21	1.11	4.53
	Raheem Sterling	2660	0.21	0.03	1.56	3.96
	Paul Pogba	1205	0.20	0.22	2.02	2.69

Defenders

Top 10 Defenders Open Play Expected Assists Per 90 Minutes

			Per 90 Minutes		Chances Created
Team	Player	Minutes Played	xA	Assists	Per 90 Minutes	Per 100 Pass
	Trent Alexander-Arnold	3176	0.20	0.17	1.16	1.76
	Ahmed El Mohamady	1026	0.19	0.09	1.84	4.26
	Marcos Alonso	1432	0.15	0.13	1.38	2.44
	Andrew Robertson	3113	0.14	0.29	1.19	1.68
	César Azpilicueta	3230	0.14	0.17	1.09	1.62
	João Cancelo	1205	0.14	0.00	1.27	1.70
	Reece James	1513	0.13	0.12	1.07	1.79
	Ryan Fredericks	2237	0.12	0.12	0.72	2.11
	Kyle Walker	2395	0.11	0.15	0.53	0.73
	Ben Chilwell	2375	0.11	0.11	1.06	1.72

Midfielders

Top 10 Midfielders Open Play Expected Assists Per 90 Minutes

			Per 90 Minutes		Chances Created
Team	Player	Minutes Played	xA	Assists	Per 90 Minutes	Per 100 Pass
	Kevin De Bruyne	2798	0.41	0.55	2.80	4.78
	David Silva	1832	0.27	0.49	1.87	2.66
	Bernardo Silva	2029	0.27	0.22	1.95	3.28
	Ross Barkley	1104	0.21	0.33	1.55	3.13
	Paul Pogba	1205	0.20	0.22	2.02	2.69
	Emiliano Buendía	2462	0.17	0.22	1.79	3.57
	Andreas Pereira	1488	0.17	0.18	1.57	3.92
	Felipe Anderson	1505	0.16	0.24	1.20	2.27
	Christian Pulisic	1726	0.14	0.21	1.36	3.91
	Jordan Henderson	2244	0.14	0.20	0.76	1.04

Forwards

Top 10 Forwards Open Play Expected Assists Per 90 Minutes

			Per 90 Minutes		Chances Created
Team	Player	Minutes Played	xA	Assists	Per 90 Minutes	Per 100 Pass
	Riyad Mahrez	1940	0.28	0.37	1.95	3.94
	Adama Traoré	2606	0.22	0.28	1.35	5.46
	Mohamed Salah	2884	0.21	0.22	1.50	5.09
	Theo Walcott	1297	0.21	0.21	1.11	4.53
	Raheem Sterling	2660	0.21	0.03	1.56	3.96
	Willian	2601	0.18	0.21	1.49	3.06
	Michail Antonio	1770	0.18	0.15	1.07	4.86
	Sadio Mané	2753	0.17	0.23	1.70	4.75
	Bernard	1278	0.16	0.14	1.55	4.15
	Sergio Agüero	1456	0.16	0.12	0.93	4.44

Graphic

Kevin De Bruyne

Riyad Mahrez

David Silva

Bernardo Silva

Adama Traoré

Ross Barkley

Mohamed Salah

Theo Walcott

Raheem Sterling

Paul Pogba

Definitions

Metric	Definition
Expected Assists (xA)	A measure of pass quality, showing the likelihood that a pass will be a primary assist. The model is based on the finishing location of the pass, what type of pass it was and a variety of other factors. This model is not reliant on whether a shot was taken from this pass, so credits all passes, regardless of whether they result in a shot.
Chances Created	A measure of the number of times a player assists a shot (including goals).

Carries

Overview

Key Points:

• Identifying ball carrying players who consistently create chances for themselves and their teammates can provide valuable insights into the league’s most dangerous dribblers. The list below highlights the Premier League’s top 10 players for creating goal scoring opportunities as a result of a carry.

• Aston Villa’s Jack Grealish tops the list, closely followed by former Villain Adama Traoré. The two players, both known for their ability to move with the ball, created 75 and 67 goal scoring opportunities from their carries respectively.

• Manchester City’s Riyad Mahrez contributed the most assists following a carry, one ahead of teammate Kevin De Bruyne.

• Clicking on the ‘Graphics’ tab brings up the pitch map for each player, plotting the chances, assists, shots and goals resulting from their carries during the season.

Top 10 Players: Shot Ending Carries

			Carry Frequency		Carry End Product
Team	Player	Minutes Played	Carries per 90	Average Carry Distance (m)	Shot Ending	Key Pass Ending	Assist Ending	Goal Ending	Total Chance Creating Carries
	Jack Grealish	3234	22	12.84	30	45	3	4	75
	Adama Traoré	2606	22	14.03	32	35	6	4	67
	Kevin De Bruyne	2798	19	10.69	30	33	6	3	63
	Riyad Mahrez	1940	22	11.51	29	27	7	4	56
	Willian	2601	23	13.00	28	28	3	3	56
	Raheem Sterling	2660	17	11.44	30	25	0	4	55
	Gerard Deulofeu	2105	17	13.28	31	19	1	1	50
	Allan Saint-Maximin	1870	20	13.71	31	15	1	1	46
	Mohamed Salah	2884	11	11.16	32	14	1	4	46
	Wilfried Zaha	3280	18	12.62	26	19	1	2	45

Graphic

Jack Grealish

Adama Traoré

Kevin De Bruyne

Riyad Mahrez

Willian

Raheem Sterling

Gerard Deulofeu

Allan Saint-Maximin

Mohamed Salah

Wilfried Zaha

Definitions

Metric	Definition
Carries	The total number of carries where a carry is defined as the player moving the ball five metres or more.
Average Carry Distance	The average distance (in metres) that a player moves the ball per carry.
Shot Ending Carry	The number of carries that were followed by a shot (including goals).
Chance Created Ending Carry	The number of carries that were followed by a key pass/chance created.
Assist Ending Carry	The number of carries that were followed by an assist.
Goal Ending Carry	The number of carries that were followed by a goal.

Goalkeepers

Overview

Key Points:

• In this section we apply shot-stopping metrics that consider the quality of the shot that the goalkeeper faces. While shots faced and save percentage can often be misleading and favour those facing a high volume of shots, these metrics, particularly assessing by ‘goals prevented rate’, can account for that. All metrics here are non penalty and exclude own goals.

• Taking each goalkeeper that appeared most frequently for their team this season, Tottenham captain Hugo Lloris stands out. Despite missing nearly half the campaign through injury, the average Premier League goalkeeper would have been expected to concede nearly 10 more goals based on the quality of shots faced by the Frenchman.

• In his debut season in the Premier League, Dean Henderson, backed up by a resolute Sheffield United defence, conceded the same number of goals and faced only marginally fewer shots than the incumbent at his parent club, David de Gea, but the higher quality of these shots, according to xGOT, led to his goals prevented rate ranking him joint-second highest in the league.

• The ‘goals prevented rate’ metric can account for different keepers facing a different number of shots throughout the season. Sheffield United’s Dean Henderson and Crystal Palaces’ Vicente Guaita both have the same goals prevented rate (1.23), despite The Eagles’ goalkeeper ‘preventing’ more goals. Normalising for the volume of shots allows us to see that both goalkeepers were expected to concede 1.23 goals for every goal that they actually conceded.

Most Featured Goalkeepers Per Team

Team	Player	% Of Team Mins	Goals Prevented Rate	Goals Prevented	xGOT Conceded	Goals Conceded	Shots Faced
	Hugo Lloris	53%	1.53	9.5	27.5	18	96
	Vicente Guaita	92%	1.23	9.4	49.4	40	149
	Dean Henderson	95%	1.23	7.4	39.4	32	126
	Martin Dubravka	100%	1.16	8.7	62.7	54	193
	Ederson	90%	1.14	3.4	27.4	24	92
	Kasper Schmeichel	100%	1.13	4.0	35.0	31	125
	Alex McCarthy	74%	1.11	3.6	36.6	33	111
	Ben Foster	100%	1.07	3.5	56.5	53	169
	David de Gea	100%	1.03	1.0	33.0	32	128
	Bernd Leno	77%	1.03	1.0	34.0	33	144
	Mat Ryan	100%	0.98	-1.2	48.8	50	166
	Lukasz Fabianski	62%	0.97	-0.8	28.2	29	100
	Aaron Ramsdale	97%	0.97	-1.6	55.4	57	182
	Jordan Pickford	100%	0.95	-2.5	47.5	50	144
	Tim Krul	95%	0.92	-4.7	57.3	62	190
	Nick Pope	100%	0.92	-3.6	41.4	45	162
	Alisson	74%	0.92	-1.6	19.4	21	79
	Rui Patrício	100%	0.88	-4.4	33.6	38	126
	Tom Heaton	52%	0.84	-5.0	26.0	31	99
	Kepa Arrizabalaga	87%	0.75	-10.6	32.4	43	97

Graphic

Hugo Lloris

Vicente Guaita

Dean Henderson

Martin Dubravka

Ederson

Kasper Schmeichel

Alex McCarthy

Ben Foster

David de Gea

Bernd Leno

Mat Ryan

Lukasz Fabianski

Aaron Ramsdale

Jordan Pickford

Tim Krul

Nick Pope

Alisson

Rui Patrício

Tom Heaton

Kepa Arrizabalaga

Definitions

Metric	Definition
Expected Goals On Target (xGOT)	Expected Goals on Target is a separate Expected Goals Model that includes the original xG of the shot, and the goalmouth location where the shot ended up. It gives more credit to shots that end up in the corners, vs shots that go straight down the middle, and is built on historical on-target shots.
xGOT Conceded	The number of goals that a keeper was expected to concede, given the quality of the on-target shots he faced.
Goals Prevented	The number of goals that a goalkeeper was expected to concede compared to the number that they actually conceded, according to xGOT. Calculated as xGOT conceded from shots on target faced, minus goals conceded.
Goals Prevented Rate	The Goals Prevented metric adjusted to reflect the number of shots a keeper faced. It is the number of goals that a goalkeeper was expected to concede as a proportion of the number of goals they actually conceded. Calculated as: xGOT conceded divided by goals conceded.

Player Sequences (Defending)

Overview

Key Points:

• Ranked by players who initiate their team’s possessions in open play most frequently, we are able to identify players who win the ball back from the opposition and initiate a sequence for their team.

• Unsurprisingly, central defenders and defensive midfield players dominate this list, which is why positions can be filtered, allowing us to better understand which players are more involved from a defensive perspective.

• In the forward position some familiar names appear, reinforcing this approach to identifying players who perform well in this role. Despite starting fewer possessions, Chelsea’s Willian ranks extremely highly in regards to starting sequences, as well as being the forward recording the most ball recoveries.

Overall

Top 10 Overall Open Play Sequence Starts

			Per 1000 Opposition Touches
Team	Player	Minutes Played	Open Play Possession Start	Open Play Sequence Start	Tackles Won	Interceptions	Recoveries
	Wilfred Ndidi	2674	10.13	23.92	3.61	4.31	14.60
	N’Golo Kanté	1733	9.41	24.64	1.53	3.37	12.09
	Jan Bednarek	3058	8.98	17.78	1.62	3.29	6.88
	Fabinho	2074	8.92	20.82	2.12	2.06	9.47
	Virgil van Dijk	3420	8.61	22.83	0.57	1.90	10.46
	Çaglar Söyüncü	3037	8.50	18.73	1.93	2.08	9.41
	César Azpilicueta	3230	8.27	17.56	2.49	2.98	7.65
	Jonny Evans	3385	8.20	19.69	1.51	2.54	10.24
	James Tarkowski	3420	8.09	16.50	1.67	2.44	6.93
	Jorginho	2384	8.07	21.33	1.91	3.05	11.74

Defender

Top 10 Defender Open Play Sequence Starts

			Per 1000 Opposition Touches
Team	Player	Minutes Played	Open Play Possession Start	Open Play Sequence Start	Tackles Won	Interceptions	Recoveries
	Jan Bednarek	3058	8.98	17.78	1.62	3.29	6.88
	Virgil van Dijk	3420	8.61	22.83	0.57	1.90	10.46
	Çaglar Söyüncü	3037	8.50	18.73	1.93	2.08	9.41
	César Azpilicueta	3230	8.27	17.56	2.49	2.98	7.65
	Jonny Evans	3385	8.20	19.69	1.51	2.54	10.24
	James Tarkowski	3420	8.09	16.50	1.67	2.44	6.93
	Aaron Wan-Bissaka	3072	7.62	17.65	3.65	3.01	8.17
	Lewis Dunk	3231	7.45	18.72	1.05	2.37	10.96
	Aymeric Laporte	1104	7.43	17.07	0.70	2.32	8.13
	Jack Stephens	2453	7.32	14.77	1.23	1.41	6.36

Midfielders

Top 10 Midfielders Open Play Sequence Starts

			Per 1000 Opposition Touches
Team	Player	Minutes Played	Open Play Possession Start	Open Play Sequence Start	Tackles Won	Interceptions	Recoveries
	Wilfred Ndidi	2674	10.13	23.92	3.61	4.31	14.60
	N’Golo Kanté	1733	9.41	24.64	1.53	3.37	12.09
	Fabinho	2074	8.92	20.82	2.12	2.06	9.47
	Jorginho	2384	8.07	21.33	1.91	3.05	11.74
	Pierre-Emile Højbjerg	2750	7.95	22.08	2.24	2.10	14.86
	Etienne Capoue	2632	7.79	18.75	1.89	3.21	10.34
	Rodrigo	2486	7.67	19.90	1.59	1.38	12.07
	Declan Rice	3420	7.63	18.34	2.20	2.83	11.45
	Jordan Henderson	2244	7.18	18.57	2.34	1.71	10.43
	James Ward-Prowse	3420	7.18	18.29	1.96	2.45	10.16

Forwards

Top 10 Forwards Open Play Sequence Starts

			Per 1000 Opposition Touches
Team	Player	Minutes Played	Open Play Possession Start	Open Play Sequence Start	Tackles Won	Interceptions	Recoveries
	Roberto Firmino	3001	3.33	10.04	0.90	0.33	6.47
	Sadio Mané	2753	3.26	9.87	1.60	0.65	6.56
	Riyad Mahrez	1940	3.17	9.05	0.70	0.90	4.52
	Willian	2601	2.96	12.22	1.27	0.94	8.69
	Son Heung-Min	2485	2.88	9.39	0.76	0.96	6.41
	Ayoze Pérez	2019	2.85	7.22	1.25	0.62	4.59
	Gerard Deulofeu	2105	2.79	9.51	0.78	1.23	6.13
	Lucas Moura	2247	2.75	8.20	1.37	0.62	4.58
	Richarlison	3081	2.73	9.79	2.04	0.52	7.32
	Danny Ings	2812	2.57	8.78	1.02	0.37	4.94

Definitions

Metric	Definition
Sequences	Sequences are defined as passages of play which belong to one team and are ended by defensive actions, stoppages in play or a shot.
Possessions	Possessions are defined as one or more sequences in a row belonging to the same team. A series of passes leading to a shot which is saved and results in a corner kick would comprise one possession since the same team retains control, but more than one sequence, since the ball has gone out of play. A possession is ended by the opposition gaining control of the ball.
Possession Start	The number of times that a player initiates the first open play sequence in a possession. A player initiating the first sequence in a possession (open play sequence start) is regaining control of the ball from the opposition. A player initiating a sequence that isn’t the first in a possession will be recovering the ball in open play following the end of their own team’s sequence (such as a shot or an opposition defensive action).
Sequence Start	The number of times that a player initiates an open play sequence.
Interception	A defending player intercepts a pass between opposition players.
Recovery	When a player takes possession of a loose ball.

Player Sequences (Attacking)

Overview

Key Points:

• Ranking by players who are involved in the most amount of goals (per 100 open play sequence involvements), we can see that centre forwards dominate this list.

• Moving across to the ‘Build Up’ tab, the list shows players involved, but removes both the shot creator and the player who took the shot, providing a list that rewards players involved in earlier phases of the sequence. In his first season in England, Manchester City’s Rodrigo is one stand-out having been involved in 20 goal ending sequences from a deeper lying position.

Top 10 Goal Ending Open Play Sequence Involvements

			Per 100 Open Play Sequence Involvements			Open Play Sequence Involvements
Team	Player	Minutes Played	Goal Ending p100	Total xG p100	Shot Ending p100	Goal Ending	Total xG	Shot Ending
	Jamie Vardy	3034	4.54	4.02	22.14	25	22.17	122
	Sergio Agüero	1456	4.35	3.64	25.40	19	15.91	111
	Tammy Abraham	2221	3.30	3.60	20.51	18	19.66	112
	Gabriel Jesus	2027	3.25	3.72	25.70	22	25.16	174
	Mason Greenwood	1312	3.05	1.48	17.11	15	7.27	84
	Anthony Martial	2638	3.01	2.35	17.33	25	19.50	144
	Danny Ings	2812	2.96	2.19	16.47	25	18.46	139
	Riyad Mahrez	1940	2.86	2.54	19.13	29	25.72	194
	Harry Kane	2589	2.84	2.10	16.08	21	15.55	119
	Son Heung-Min	2485	2.73	2.02	15.92	25	18.49	146

Build Up

Top 10 Goal Ending Open Play Build Up Sequence Involvements

			Per 100 Open Play Build Up Sequence Involvements			Open Play Build Up Sequence Involvements
Team	Player	Nineties	Goal Ending p100	Total xG p100	Shot Ending p100	Goal Ending	Total xG	Shot Ending
	Riyad Mahrez	21.6	1.45	1.19	8.69	13	10.66	78
	Rodrigo	27.6	1.30	1.09	8.88	20	16.83	137
	Nicolás Otamendi	19.0	1.25	1.03	8.21	11	9.03	72
	Dani Ceballos	18.9	1.22	0.75	4.86	11	6.81	44
	João Cancelo	13.4	1.19	1.12	9.62	9	8.52	73
	Marcos Alonso	15.9	1.18	0.64	5.52	10	5.44	47
	Nemanja Matic	17.6	1.10	0.74	7.68	10	6.73	70
	Paul Pogba	13.4	1.10	0.67	5.90	8	4.85	43
	Dennis Praet	12.6	1.09	0.99	6.74	5	4.54	31
	Oleksandr Zinchenko	14.1	1.09	0.88	7.05	9	7.22	58

Definitions

Metric	Definition
Open Play Sequence Involvement	The number of unique open play sequences that a player was involved in. (Note: these are unique because multiple involvements in a single sequence will only count as one involvement)
Goal Ending Sequence Involvement	The number of unique goal-ending sequences in open play that a player is involved in.
Shot Ending Sequence Involvement	The number of unique shot-ending sequences in open play that a player is involved in.
xG Sequence Involvement	The total xG value of unique open play shot or goal ending sequences that a player was involved in.
Build Up Sequence Involvement	The number of unique build up sequences (shot and key pass removed) in open play that a player was involved in.
Goal Ending Build Up Sequence Involvement	The number of unique goal-ending build up sequences (shot and key pass removed) in open play that a player is involved in.
Shot Ending Build Up Sequence Involvement	The number of unique shot-ending build up sequences (shot and key pass removed) in open play that a player is involved in.
xG Build Up Sequence Involvement	The total xG value of unique build up sequences that a player was involved in that resulted in a shot or a goal.