The Proof-Of-Concept Project at GKGrades.com

Due to recent events, I'm now provided the ability to bring my head above water from the pool of work and stress I was drowning in for a long time.  My free-time can now be partially filled with more Fox Soccer Channel, Gol TV, MLS Direct Kick, and UEFA Champions League.

In June, I began a proof-of-concept project where I formulated a system for which goalkeepers could be measured by a numerical rating that factors in many different applicable game statistics.  The intent is to revolutionize the way the goalkeeper position is graded.  It's my opinion that goals-against-average (GAA) and shutouts are the most deceiving positional statistics in all of sports.  The result is a numeric rating similar to that of the quarterback position in the National Football League.

Using the 2010 FIFA World Cup as the testing ground, I've got about 30 more games to analyze and log into the GKGrades intelligence database.  Many of the games have already been rated with published ratings on this website.   The final project deliverable will be a comprehensive look at goalkeepers and their performances throughout the entire World Cup.  I think many people will be surprised to see that the highest rated GK's might not be traditional powerhouse names like Iker Casillas or Gigi Buffon, but rather someone that plays for a national team or club that flies relatively low beneath the radar.

Goalkeeping is a position that requires intense focus and non-stop concentration. Don't get caught sleeping like the guys in this video...

Diego Binaglio of Switzerland Ranks As Best World Cup Keeper Through Game 1

The following table represents the GoalkeeperGrades.com GK Rating^TM for the first games in the 2010 World Cup. The GK Rating^TM, borrowed from the concept of a passer rating or quarterback rating in the NFL, is a cumulative measure of the total performance of a goalkeeper. GK RatingTM is an algorithm calculated using each goalkeepers's total touches, tests, goals allowed, PK goals allowed, routine saves, exceptional saves, PK Saves, errors, and critical errors. A GK Rating does not have a maximum or a minimum, but a rating of 100 or above is considered to be exceptional. A GK Rating below 80 is considered poor.

Download Complete Statistical Table (.pdf)

GK.Name	Squad	Opponent	GK.Rating
Diego.Binaglio	Switzerland	Spain	121.1
Vincent.Enyeama	Nigeria	Argentina	114.4
Jung.Sung.Ryong	South Korea	Greece	106.2
Tim.Howard	USA	England	106.0
Fernando.Muslera	Uruguay	France	105.8
Richard.Kingson	Ghana	Serbia	105.7
Maarten.Stekelenburg	Netherlands	Denmark	105.5
Samir.Handovic	Slovenia	Algeria	102.9
Itumeleng Khune	South Africa	Mexico	101.4
Claudio Bravo	Chile	Honduras	100.8
Sergio Romero	Argentina	Nigeria	100.7
Manuel Neuer	Germany	Australia	100.7
Eduardo	Portugal	Ivory Coast	100.6
Boubacar Barry	Ivory Coast	Portugal	100.5
Federico Marchetti	Italy	Paraguay	100.1
Thomas Sorensen	Denmark	Netherlands	98.4
Hugo Lloris	France	Uruguay	94.6
Noel Valladares	Honduras	Chile	93.3
Justo Pillar	Paraguay	Italy	92.2
Oscar Perez	Mexico	South Africa	90.5
Eiji Kawashima	Japan	Cameroon	90.0
Alexandros Tzorvas	Greece	South Korea	88.2
Ri Myong Guk	North Korea	Brazil	87.2
Hamidou Souleymanou	Cameroon	Japan	86.7
Jan Mucha	Slovakia	New Zealand	85.6
Mark Paston	New Zealand	Slovakia	79.5
Mark Schwarzer	Australia	Germany	77.6
Vladamir Stojkovic	Serbia	Ghana	75.6
Julio Cesar	Brazil	North Korea	75.4
Faouzi Chaouchi	Algeria	Slovenia	72.9
Robert Green	England	USA	68.1
Gianluigi Buffon	Italy	Paraguay	50.2
Iker Casillas	Spain	Switzerland	46.9
AVERAGE			91.7

Here are some statistical insights on goalkeeper performance after each nation's first match...

Highest Save %:
Diego Binaglio, Switzerland (1.000 - 11 saves)

Lowest Save %:
Gianluigi Buffon, Italy (0.500)

Highest Exceptional Save %:
Diego Binaglio, Switzerland (1.000 - 4 exceptional saves)

Most Involved GK:
Mark Paston, New Zealand (42 touches & tests)

Least Involved GK:
Julio Cesar, Brazil (13 touches & tests)*

Most Saves:
Ri Myong Guk, North Korea (12)

Highest Error Rate:
Eiji Kawashima, Japan (0.107)

Highest Critical Error Rate:
Iker Casillas, Spain (0.095)

* Italy's Buffon and Marchetti each logged 12 touches and tests, but only played 45 min each due to Buffon's injury.

If GAA was the primary statistic in use, Mark Schwarzer (Australia) would be last on the list with a GAA of 4.000. Anybody who watched this match would understand that this is further proof that GAA is not a proper measure of goalkeeper performance. Schwarzer's stat line of the match against Germany reads as:

GK Rating: 77.6
Touches &; Tests: 40
Goals Allowed: 4
Routine Saves: 6
Exceptional Saves: 2
Errors: 0
Critical Errors: 0

Your performance measure should be significantly affected when you let 4 goals get by, but it shouldn't necessarily rank you as the worst GK on the list. Schwarzer's GK Rating ^TM ranks him as 27th out of 33 goalkeepers who saw time in match 1.

The New Era of Goalkeeping Statistics

It is a wide-spread belief that the standard set of goalkeeping statistics are not a proper measure of goalkeeper performance. Looking at the stat sheet following a game can be very deceiving. The box score might read that a goalkeeper only had 2 saves which makes the reader believe that the GK was not very involved in the match... when in actuality that same GK may have had 30+ touches on the ball.

The traditional 'Shots on Goal' does not include every dangerous ball that is played into the goal area. The traditional 'Saves' only represents a limited subset of diffused chances that are directly related to the already skewed 'Shots on Goal' statistic.

Goals Against Average (GAA) is possibly the most useless statistic of them all. Great goalkeepers can play on bad teams and average goalkeepers can play on great teams. The GAA stat does not provide visibility into anything outside the number of times the ball ends up in the net. There is a place for this statistic at the team level, but should not be used as the primary measure of goalkeeper performance.

As Americans, we feel the need to track statistics for all aspects of sport to help measure team and individual performances as well as give sports talk radio a reason to exist. Because I have background in information systems, business intelligence, and goalkeeping, I feel compelled to make an attempt at the development of a proper quantitative and more comprehensive approach for measuring goalkeeper performance.

With that said, I will be using the World Cup as a testing ground to measure the following set of metrics and periodically reporting on the results.

• Touches: Number of times the GK physically touches the ball. A metric to measure the involvement of the GK in the game. Every touch is an opportunity to make an error. Includes goal kicks, back passes, set plays, saves, etc...

• Tests: Another metric to measure the involvement of the GK in the game. Contrary to a Touch, a Test is where the GK does not physically touch the ball. A test can include a goal allowed or a ball that hits the post. Also includes any situation where a GK is forced to make a play on the ball. Includes any situation where the GK directly or indirectly affects the result of a challenge withough touching the ball. Breakaways where the GK makes a play on the ball and the ball is shot off-target is a good example of a test.  

• Goals Allowed: Goals allowed during the course of normal play. Does not include goals converted via penalty kick.

• Penalty Kick Goals Allowed: A separate goals allowed metric to identify instances where the odds of scoring are heavily favored towards the offensive team.

• Penalty Kick Saves: Helps to identify GK's who specialize in this area of the game.

• Routine Saves: Includes all saves made where the general expectation is that these saves should be made by the GK during the normal course of play. Includes all situations where the GK diffuses a direct chance on goal or an indirect opportunity such as a dangerous cross or through ball. GK's should be given credit for all saves... not just the ones that are a result of shots on goal.

• Exceptional Saves: Great goalkeepers keep their squads in games by making a couple saves every game that could easily result in a goal for the other team. This subjective call by the statistician is a count the number of times a GK has made an incredible save. Penalty Kick Saves, Routine Saves, and Exceptional Saves are summed to create the Total Saves statistic.

• Exceptional Plays: Occasionally goalkeepers can be credited with an exceptional play that falls outside the traditional role of goalkeeper. This category counts all other exceptional plays that should be identified as part of goalkeeper performance. This would most likely include distribution the directly leads to a goal, moving forward to create a direct chance at goal, scoring, etc...

• Errors: Borrowed from baseball, an error is the act, in the judgment of the official scorer, of a goalkeeper faulting on a play in a manner that opens up an opportunity for the opposition to make a play, when such a play should have been prevented given ordinary effort by the goalkeeper.

• Critical Errors: Errors that lead to a goal allowed.

• Error Rate: A measure of Errors against the sum of Touches + Tests.

• Critical Error Rate: A measure of Critical Errors against the sum of Touches + Tests.

• Save Percentage: A measure of goals allowed during normal play (not including penalty goals) against against the sum of Routine Saves + Exceptional Saves. Save percentage has long been a staple statistic, but the GKGrades version is thought to be a better representation because saves include all situations where the GK diffuses a direct chance on goal or an indirect opportunity rather than just shots on goal.

• Save Percentage + PKs: A measure of goals allowed against against the sum of Routine Saves + Exceptional Saves + Penalty Saves. For purposes of this stat, 1 penalty kick goal is weighted at only 25% of a goal allowed in normal play.

• GK Rating^TM:  Borrowed from the concept of a passer rating or quarterback rating in the NFL.  A GK Rating^TM is a cumulative measure of the total performance of a goalkeeper. GK Rating^TM is an algorithm calculated using each goalkeepers's total touches, tests, goals allowed, PK goals allowed, routine saves, exceptional saves, PK Saves, errors, and critical errors. A GK Rating does not have a maximum or a minimum, but a rating of 100 or above is considered to be exceptional.  A GK Rating below 80 is considered poor.