.@TiagoEstv uses Opta data in this article for @StatsBomb, in which he assesses upcoming prospects in @ligaportugaltwitter.com/i/web/status/9… 11 Dec

OptaPro's data scientist @Worville will be speaking at the 'Getting your Data skills noticed - Data Science and Vis… twitter.com/i/web/status/9… 6 Dec

.@EuanDewar uses Opta data to dissect and discuss Atlético Madrid’s attack this season. bit.ly/1Hnj29V. Ailing. 4 Dec

.@jair1970 writes a 12-game Premier League shot analysis, focusing on teams’ performances in relation to previous s… twitter.com/i/web/status/9… 27 Nov

.@OptaPro will host a panel tomorrow at #SAC17 with @soccerquant, Ben Stevens (@CPFC head of performance analysis)… twitter.com/i/web/status/9… 23 Nov

.@EuanDewar and @EveryTeam_Mark co-author this data-driven article that explores Manchester City's impressive start… twitter.com/i/web/status/9… 23 Nov

The latest #ThreeAtTheBack episode outlines further information on @scout7football becoming part of OptaPro.… twitter.com/i/web/status/9… 17 Nov

Following the acquisition, the #ThreeAtTheBack pod discusses this in further detail, along with a conversation on c… twitter.com/i/web/status/9… 17 Nov

The OptaPro Blog

RSS Feed

Welcome to the OptaPro blog, featuring news and analysis from OptaPro's cutting-edge research team.

GUEST BLOG: An alternative analysis of Key Passes

In an article published over on Statsbomb last week I took a look at potential methods of obtaining as much useful information as possible from the Key Passes (including goal assists) metric.

To recap; I started off looking at the number of Key Passes per90 minutes and Goals per Key Pass before I introduced the concept of Expected Goals, or Expected Assists to be more precise.  The introduction of Expected Assists allows us to objectively assign a numerical value to each Key Pass.  This numerical value tells us the goal probability of the shot or header that each Key Pass set up, so no longer do we have to wonder whether a certain player’s Key Pass numbers are inflated due to a penchant for setting up low probability chances.

It’s fair to say that most people use Key Pass as a proxy for creativeness, and the purpose of this small series of articles is to provide some potential ideas as to how this level of creativity can be objectively measured across a series of players.  In my previous piece I’d already controlled the total Key Pass numbers for time spent on the pitch and the quality of the shots that the Key Passes created, but there was one other aspect that I wanted to control for.

Total Number of Potential Key Passes

The final aspect that I wanted to control for in compiling a list of creative players was the total potential Key Passes that the player in question could make.

A hypothetical player may have 3 Key Passes per 90 minutes and an Expected Assist (ExpA) figure per90 minutes of 0.45.  These figures are elite and suggest that the player in question is probably one of the more creative players in the league.  However, if our player was the main creative outlet for his team and virtually every attack went through him then perhaps his statistics may not be as impressive as they seemed on first viewing.
This is because, it could be argued, the player had so many attempts at setting up shots that sheer volume of numbers would ensure that he would eventually succeed with some of them.

There is no metric available that can tell us when a player should have made a Key Pass but didn’t, or just how many attempts at providing Key Passes were unsuccessful, but the folks at OptaPro have provided me with some data that can be used as a proxy.

Final Third Passes

OptaPro have provided me with the number of Final Third Pass (including cross) attempts made by the top 20 Key Pass players through the first eleven weeks of the 2012/13 EPL season – this is the same sample of players that I analysed in my previous piece.

I am fully aware that the total Final Third Pass attempts made by each player can never be thought of as the number of Key Passes that a player should make.  Likewise, there will be some Key Passes which are made that are not Final Third Passes, however, all I’m after is a measure that shows how “involved” a player is in the attacking part of the pitch, and I think that Final Third Passes is as good a metric as we have for this task.

In summary, the logic here is that a player that makes a large number of Final Third Passes will have more chance of creating Key Passes than one who makes relatively few passes.

Positional Challenges

We would also expect to see the players’ positions to have an impact on the total number of Final Third passes attempted by each individual.  A winger would have fewer touches or passes in the attacking third than someone involved in the engine room of the midfield.  This means that our metric cannot be used across different player categories, but instead it would be used to develop benchmark efficiencies for each position.

The following table has the number of Key Passes and total Final Third Passes played by our top 20 Key Pass owners.  I have simply divided the number of Key Passes into the number of Final Third Passes to show the number of Final Third Passes that are played for each Key Pass – the table is ranked by this measure.

(Click on table to expand)

Key Passes Table 1

The position of each player in the above table was decided by me.  Most of the players have played in more than one position this season, but I have stated the role that I typically see each of the players in.

We can see that even the most efficiently creative players take more than 8 Final Third Passes for each Key Pass they make, and our list ends at 14 Final Third Passes per Key Pass.

Even with just a list of 20 players (which is a very small sample) we can see positional patterns emerge.  Wingers and forwards tend to inhabit the top of our list and attacking midfielders fall towards the bottom of the rankings.

These findings will not come as a surprise, as wingers and forwards would be expected to be involved less in general build-up play than their teams’ creative playmakers would.  However, some of the positional exceptions are interesting, especially those of the Tottenham players.

Christian Eriksen is comfortably the most efficient of my attacking midfielders with a Key Pass every 9 Final Third Pass attempts.  Perhaps this suggests that Eriksen almost plays the role of a winger as opposed to the Attacking Midfield position that I have pigeon holed him in, or else he is just super-efficient at creating chances.  Either way, his ratio is in sharp contrast to the Dane’s Tottenham colleague, Andros Townsend.  Townsend is most definitely a winger, so it would therefore be disappointing for the England man to see that he takes more than 13 Final Third Pass attempts for each Key Pass that he makes.  This, combined with his love of speculative shooting (45 shots in the EPL with an average goal expectancy of just 3.6% per shot), suggests that Townsend has some improvements to make in terms of productive end product before he fulfils the potential that he possesses.

Combined with Expected Assists

The information in the above table was useful, but we can improve on this by replacing the number of Key Passes for each player with the Expected number of Assists (ExpA) earned by their passes.  These ExpA values have been developed by Constantinos Chappas and I, and they objectively measure the number of goals that our model estimates should have been scored by the shots the Key Passes created.  The inputs to our model include the location of the shot, the type of the shot and how the shot came about.

Final Third Passes per Assist

(Click on table to expand)

Key Passes Table 2

The above table gets us to the position of controlling for both the attacking involvement of the players as well as the quality of the chances that they have set up in respect of playing the final pass.

As with our previous table, there is a clear demarcation in terms of player positions.  The two forwards in the list of top 20 Key Pass makers occupy the top two positions in this table as they require just 60 Final Third Pass attempts before they would expect to set up a goal.

The three most efficient wingers in our sample are closely grouped with Morgan Amalfitano, Adam Johnson and Stewart Downing all requiring less than 70 attempted Final Third Pass attempts before they could be expected to set up a goal.

The quality of chances set up by Andros Townsend is not sufficient to prevent him being the lowest ranked winger in our table.  At 139 Final Third passes, he takes twice as many passes as the most efficient wingers in the EPL before we can expect him to create a goal.

Many Liverpool fans seem to wonder whether Steven Gerrard is still worthy of his place in the Liverpool starting team.  I can’t give a definitive answer to that question, but in terms of the creative side of his game he’s head and shoulders clear of the other midfielders in our table – that’s quite an accomplishment for a player of his “experience”.

David Silva

Special mention must go to the little Spanish Wizard, David Silva, as he emerges as the most creative of our Attacking Midfielders.  I tend to think of Silva, Mesut Ozil and Eden Hazard as playing the same role, that of a creative fulcrum in their respective teams.  Yet it is very noticeable just how few passes he needs to play before he is expected to create a goal compared to Ozil and Hazard.

Thus far in the 2013/14 EPL season, for every 100 passes attempted in the Final Third Silva can expect to see scoreboard operator troubled.  This compares very favourably with 166 passes for Hazard and 181 for the German Ozil.  These numbers reaffirm my personal belief that Silva is the most creative player in the Premier League.  Yes, even more creative than Mesut Ozil.

Summary

In this series of articles I have provided a few different ideas around how to objectively measure the attacking creativity of players.  Readers will note that I did not make reference to the number of goal assists that each of the top 20 players had this season.  That is because the actual number of goals scored is not important; provided our model is accurate we do not need to know what happened to the shots that were created.

Remember, our aim is to develop metrics that measures the creative capacity of players.  In doing this we are controlling for as many of the variables that are outside the control of the Key Pass maker as possible.  The conversion, or otherwise, of the subsequent shot is one such factor.

I am sure that some readers will argue with my choice of metric used (Final Third Passes) in this piece.  But even if you have misgivings about the details of the calculation, this piece is as much about illustrating that the right questions need to be asked of the top level statistics for us to obtain as much meaningful information as possible from them.

I believe that any scout or club that is in the market for a player needs to develop some benchmark models that control for as many factors as possible.  Statistics and analytics will never totally replace the subjective eye of a technical scout, but using objective data to narrow down a potential list of players will ensure that time spent actually eye balling prospective targets can be reduced.  I would also contend that using analytics in a club’s transfer activity will reduce the number of bad buys that will be made.

Written by Colin Trainor - @colinttrainor

Posted by Colin Trainor at 10:35

Related Links

Assessing key passes

Colin Trainor extracts greater meaning from the Key Passes metric.

Click here to view article

How much does Juan Mata to Chelsea?

A brief but interesting article from Paul Riley, looking at how effective Juan Mata is for Chelsea.

Click here to view article

Liverpool’s chance quality

Andrew Beasley experiments with a different way of effectively assessing chance quality.

Click here to view article

0 Comments

Post a comment