Introducing EPL Effect Size Index

Effect Size Index (ESI) tracks how well a team performs when a player is on the park, as opposed to their scoring/concession rates when he is not. 

A player who plays every minute of every contest for his club - usually a goalkeeper or centre-back, will have an ESI of 0.000, as this statistic charts the difference between the team's performance when a player is on the pitch (in terms of frequency of goals scored/conceded) and compares it to his club's overall rates of goals scored/conceded. Thus, a player who plays every minute of the season - or who's Individual Plus/Minus per 90 minute matches that of his club - has an ESI of 0.

Take, for example, Jack Wilshere (please). Arsenal score less frequently when he is on the pitch, while conceding more often: Arsenal have GD this season of +7, but while Wilshere has been on the pitch the Gunners have been outscored by a total of 10 goals to 4 (making Wilshere's +/- a poor -6). While this might be bad luck, bad teammates or bad Jack, and keeping in mind that 0.000 is for want of a better word "baseline", his ESI is a phenomenally bad -1.527.

You can find the ESI of every player who has managed 450 minutes on the pitch this season by clicking this link. Player are grouped by position, and arranged in order of best to worst ESI.

On the impact of Bobby Zamora, Eduardo Vargas and Richard Dunne at QPR

The thoroughly entertaining Iain Macintosh today compiled an article suggesting how well QPR have been performing recently. He credits much of this resurgence to Harry Redknapp's re-utilisation of two senior Rangers, Bobby Zamora and Richard Dunne.

While he is of course correct - especially in a poetic sense - a certain set of numbers emerging from my ever-expanding Player Plus/Minus dataset - which you can find by clicking here - might help quantify the effect Zamora and Dunne have both have on the Hoops.

In their eleven matches so far, the Rs have managed a goal difference of -11, or being outscored by one goal for every 90 minutes they take the field. This means any player who has managed each of the 990 minutes Rangers played so far this season will have an Individual Plus/Minus of -11. If you will, think of Individual Plus/Minus as a player's "personal" goal difference, or the amount of goals his team concedes while he's on the field, subtracted from those his team scores when he plays.

However, players rarely play every minute of every match of the Premier League season - in fact, in last year's abbreviated (four-team) sample, only Brad Guzan and Steven Caulker managed such a feat. This means to compare a player to his teammates and thereby examine his impact on his club, we must standardize the time-frame in which player and team both score and concede. That is, a team's Goal Difference divided by 38 results in the average amount they score/concede more than their opponents per match - so a Player's Plus/Minus rating per 90 minutes allows us to compare a team's performance when a certain player is deployed against when he is not.

But technical definitions aside, this set of information details how much a team scores or concedes over 90 minutes when a player is on the field. And Zamora's, particularly, numbers are spectacular. When he is on the field, QPR have a Goal Difference per game of 0, a full goal-per-game improved over when he doesn't take the park. 

This is the single best "Impact Factor" for a forward in the league, and is shared (with teammate Yun Suk-Young) for the third-best rating of any position. Only Jack Cork and Charles N'Zogbia are better throughout the entire league; another Ranger with a similarly lofty total is forward Eduardo Vargas, whose Impact rating of +0.803 is fifth-best.

Dunne, the other cartworkhorse whose virtues Macintosh espouses, rates as the seventeenth-most impactful defender in the league of the 95 that qualified (must have played at least 360 minutes).

Other interesting points to come out of the data set include Adnan Januzaj's lack of ability to influence proceedings for Manchester United (with the Reds 0.71 goals per 90 minutes worse off with him on the field), Jack Wilshere's awful start to the campaign (Arsenal are nearly 1.7 goals worse-off with him playing, the worst ratio in the EPL) and the absolute ostrich egg laid by France forward Emmanuel Riviere.

For more information - and some pretty, relatively-informative charts - on Player Plus/Minus, you can read the primer here, or search this site for "Plus/Minus".

Six games in: Individual player Plus/Minus

Updated Individual Plus/Minus player stats are available by clicking here.

You can also find a link via the Room of Informational Illusions.

2014-15 EPL individual player plus/minus

Interested in which player contributes the most - or the least - to your favourite club's wins or losses? You can find out by checking out the 2014-15 EPL individual plus/minus stat pages.

You can also access this information via the Room of Informational Illusions.

(Suggested reading: A Plus/Minus Glossary).

What the numbers said: Cardiff City 2013-14

The following analysis was performed utilizing data from the Individual Plus/Minus series published on the site throughout the year. You can find the full data set in the Room of Informational Illusions.

Should you wish for a glossary of terms used in this article, it can be found here.

Overview

Cardiff City didn’t have a great 2014. After an encouraging start, the Bluebirds tailed away almost at the precise the instant owner Vincent Tan began questioning manager Malky Mackay’s job security. When Mackay was replaced by Ole Gunnar Solskjaer, the new manager began rotating their first team and team structure at an alarming rate – with the results you’d expect.

Some players didn’t suffer from their Welsh association. The two players for whom Cardiff City broke their transfer record in 2013 (Gary Medel and Steve Caulker) were perhaps the club’s best players, while two loyal Mackay men – David Marshall and Fraizer Campbell – had seasons that have or still might win them moves back up to the Premiership.

Teen millionaire comparison: Luke Shaw vs. Calum Chambers

The £12 million (rising to 16 million) paid for young Southampton right-back Calum Chambers might best exhibit the premium placed on potential in the English Premier League. The nineteen year-old joined the Gunners this week for a fee around half of that paid by Manchester United for line-mate Luke Shaw, who travelled north for a sum thought to be around £30 million.

There are a few subtle differences between the pair, however. Firstly, Chambers can’t possibly expect to earn the reputed £100,000 per week. This is probably in part because he hasn’t yet played for England, nor apparently interested the club he supported as a boy. And – perhaps – finally, while a quality player and precisely no grumbling has accompanied his transfer, Chambers’ performances for the Saints last year didn’t actually inspire a lot of success (he might be fitter, though).

What the numbers said: Aston Villa 2013-14

The following analysis was performed utilizing data from the Individual Plus/Minus series published on the site throughout the year. You can find the full data set in the Room of Informational Illusions. Should you wish for a glossary of terms used in this article, it can be found here.

Overview

Villa finished the season in fifteenth position on the EPL table, when, pre-season many expected them to challenge for a top-half berth.

2013-14 was hardly a success for Aston Villa. After assuming control of the midlands club prior to 2012-13, Scotland’s entry for World’s Most Charismatic Man Paul Lambert orchestrated something of a revolution-on-the-cheap by marginalizing the overpaid stars who had taken Villa to the brink of Champions League football and then almost to relegation. Little was heard of Alan Hutton, Darren Bent, Charles N’Zogbia and Jean Il Makoun, while young players from lower divisions like Matthew Lowton and Ashley Westwood came in as accoutrements to centerpiece Christian Benteke, who was acquired from Genk.

While some Villans – not least captain Ron Vlaar and central midfielder Fabian Delph – had solid seasons, many of the players who finished 2012-13 so promisingly regressed or struggled to impact games in 2013-14 as they had the previous season.

One of the more notable examples was Lowton, who managed only 23 matches for the term. The right-back, who scored the goal of the season against Stoke City the year before, was a defensive liability and spent vast stretches of the season either on the bench or unselected. While Andreas Weimann’s numbers according to the Individual Plus/Minus system compared quite favourably to his teammates, far more was expected from the Villa player to whom the Spiderman Principle most obviously applies.

A Plus/Minus Glossary

Plus/minus (+/-) = The amount of cumulative goals over the course of a season a club scores with a particular player on the field. For example, should Manchester United score 50 goals over a season with Ryan Giggs on the field while conceding 34, his Plus/minus score would be +16.

Scored/90 (S90) = The amount of goals scored per 90 minutes a player is on the field. This varies slightly from goals scored per game, as often football players don't play entire games. Scored/90 allows us to observe the rate at which a team scores goals while a certain player is on the field.

Using the same example as above, were Ryan Giggs to play every minute of every game (ie. 3420 for the season), his Scored/90 and Goals-per-game rates would be the same at 1.315. However, if Giggs played only 3000 minutes for the year over those 38 matches, his Scored/90 rate rises to 1.5.

Conc/90 (C90) = As for Scored/90, only tracking the rate at which a team concedes while a certain player is on the field. Any of these numbers adjusted for rate allows us to compare players within teams - if Giggs' Conc/90 is 1.0 and Nani's stands at 1.1, we can suggest that United are better off defensively on the wing with Giggs on the left wing.

Goal Difference (GD) = As calculated by leagues for years, Goal Difference (sometimes here you'll see it denoted as Team GD) is the season-long difference between goals scored and goals conceded. Adjusted slightly, it becomes GD/game, or how many goals on average a club scores or concedes than their oppositon per game. You (really should) know how it works.

+/- per 90 (PMP) = Adjusts Plus/Minus for rate, allowing us to judge a player by his compatriots. Using Ryan Giggs again, with a Scored/90 rate of 1.315 and Conc/90 rate of 1.0, his +/- per 90 stands at 0.315.

Impact Rating (IR) = Assesses the difference between a team's performance while a player is on the field, with their overall performance.

Again, using the above example involving Ryan Giggs, if United score 16 goals more than their opponents for the season when he plays, but have a goal difference of +7 for the season, his Impact Rating is his PMP compared to his team's GD/game. For this example, if Giggs played 1500 minutes over 30 games, his IR would rate at 0.324 - or Man U would be 0.324 goals per 90 minutes better than their opponenets when Ryan Giggs was on the pitch.

Individual plus/minus: Aston Villa, Cardiff City, Manchester United, Southampton & Tottenham Hotspur - 11th May 2014

Statistics valid before games on May 11^th.

Aston Villa

Player	G	Min	GS	GA	+/-	Scored/90	Conc/90	Team GD
Guzan	37	3330	39	58	-19	1.054	1.568	-19
Westwood	34	3007	34	53	-19	1.018	1.586	-19
Delph	33	2913	36	49	-13	1.112	1.514	-19
Vlaar	31	2715	35	44	-9	1.160	1.459	-19
Weimann	36	2564	34	43	-9	1.193	1.509	-19
Bacuna	34	2511	30	44	-14	1.075	1.577	-19
Baker	29	2384	25	41	-16	0.944	1.548	-19
Agbonlahor	29	2372	24	35	-11	0.911	1.328	-19
Clark	27	2206	30	38	-8	1.224	1.550	-19
Benteke	26	2170	26	38	-12	1.078	1.576	-19
El Ahmadi	30	2146	26	40	-14	1.090	1.678	-19
Lowton	22	1683	21	33	-12	1.123	1.765	-19
Luna	17	1381	15	18	-3	0.978	1.173	-19
Bertrand	15	1301	18	28	-10	1.245	1.937	-19
Albrighton	19	967	13	20	-7	1.210	1.861	-19
Kozak	14	781	10	9	1	1.152	1.037	-19
Tonev	17	596	6	8	-2	0.906	1.208	-19
Sylla	11	548	5	10	-5	0.821	1.642	-19
Holt	10	376	2	6	-4	0.479	1.436	-19
Bennett	5	244	0	2	-2	0.000	0.738	-19
Bowery	8	188	2	2	0	0.957	0.957	-19
Herd	2	171	1	4	-3	0.526	2.105	-19
Okore	3	167	1	3	-2	0.539	1.617	-19
Helenius	3	49	0	1	-1	0.000	1.837	-19
Robinson	3	39	0	3	-3	0.000	6.923	-19
Grealish	1	2.00	0	2	-2	0.000	90.000	-19