How to scout a striker?

Scouting strikers should not be that hard, right? Their prime responsibility is putting the ball in the back of the net, and goals are one of the few elements of football where traditional fans and nerdy analysts agree. A goal is a goal, counting goals cannot go wrong. Strikers who score a lot of goals are better than strikers that score less goals. Or not?

In our previous piece on scouting offensive talent, we’ve distinguished two elements that constitute a good striker.

  1. The striker has to get into good scoring positions, and accumulate good shots. This is best measured as Expected Goals (ExpG) per 90 minutes, with exclusion of penalties.
  2. The striker has to convert these chances into goals. This can be measured by comparing ExpG and actual non penalty goals.

The previous post on strikers illustrated how we can measure those two elements and judge strikers separately on both of these qualities. Today we will take it a step further and see what scouting implications come from it. We will show that sometimes it is better to buy a lower scoring striker, and which high scoring strikers to avoid. But first, I want you to meet someone.
 

Meet our striker!

He plays in a big league, for a good team, where he has taken 160 non penalty shots in the past season. On average, each shot was good for 0.152 ExpG, so over all shots together we could have expected 24.4 goals from him.

The thing is, our striker is pretty good, so instead of 24.4, he scored 43 non penalty goals for an over performance of 18.6 goals. We can stick an ugly acronym to it and say his non penalty goals above replacement (NPGAR) is 18.6.

NPGAR = Non Penalty Goals – Expected Non Penalty Goals

You’ve probably guessed by now that our striker is Lionel Messi. This season, Messi still plays for Barcelona, where he has taken 75 non penalty shots to date. On average the quality of the chances was comparable to last season, with an ExpG per shot of 0.149. Overall, we should expect 11.1 goals.

The thing is, Messi is suddenly not so excellent at finishing, and he has come up with 9 non penalty goals instead of 11. His NPGAR is now -2.14, which indicates that the average player, not even the average striker, would have scored two more goals with the type and number of shots that Messi has taken this season.

 

Analysis

A story about Messi is not analysis, it’s anecdote. And anecdotal evidence is no evidence. We could ‘prove’ that finishing does stick with a player by simply picking someone else that happened to follow an excellent finishing season with another excellent finishing season and fire that point home.

It makes more sense to repeat this work for all 479 players of the top-5 leagues who took at least 10 non penalty shots in the baseline 2012/13 season. We take separate looks at the creation of goal scoring chances (ExpG per 90) and at the conversion of chances into goals (Goals minus ExpG). Both parameters will be compared over one season and the next.

 

ExpG per 90

In the first graph we will look at the repeatability of non penalty Expected Goals per 90 minutes (ExpG NP per90). The horizontal axis shows ExpG NP per 90 for the first season, and the vertical axis shows the same for the next season.

ExpG90 correlationExcellent! It turns out that players with a high ExpG per 90 in one season, are also the players with a high ExpG per 90 in the next season. This is not too surprising, as several factors influencing ExpG per 90 will remain constant over time. Strikers will still be playing as strikers, and most players playing for top team will still be playing for top teams. More work needed here, but we’ll leave that for another post, as there is a far more interesting graph coming up.

 

NPGAR

The next graph shows the repeatability of non penalty goals above replacement (NPGAR). This represents the conversion of goal scoring chances into actual goals.

NPGAR correlationIt turns out that if you correct for the quality of goal scoring attempts, there is absolutely no connection between conversion in one season and the next. A high or low NPGAR in one season has zero relation with NPGAR in the next season.

Messi is the dot in the lower right hand corner, who had an unworldly 2012/13 season, with an NPGAR of +18.6, followed by the current season of -2.1.

 

Scouting

This is a shocking conclusion with huge implications for striker scouting. If a striker bases his goal scoring mainly on conversion, he has a good chance to fail in the next season. If a striker bases his goal scoring mainly on good underlying ExpG numbers, he has a good chance to persist his level of scoring.

Buying strikers who score their goals due to a high NPGAR is something you should always avoid.

We all know these famous examples of one season wonders, who got transferred for big money, only to disappoint at their new clubs. Usually, loads of soft factors like the higher level of competition, language issues, or playing style are used to explain the disappointing results, while the only thing going on is regression of NPGAR.

Regression does not always occur though, and you can see in the scatter plot that some players do indeed follow a season of high NPGAR with another season with high NPGAR. But just as many players do not, and just as many players with high NPGAR in the second season come off seasons with low NPGAR.

 

Finnbogason

We should use NPGAR as a red flag in striker scouting. A player like Alfred Finnbogason, currently the Eredivisie top scorer with 21 goals in 20 matches, is a nice example. We can put up several red flags.

First, 8 of his 21 goals are penalties. Second, his NPGAR is +2.68, indicating that he is nearly three non penalty goals above expectations. There is no ground at all to assume that he, or any other player, will outperform the ExpG model  next year. All in all, Finnbogason’s non penalty ExpG per 90 is 0.51, which is still a good number, but by no means near the present perception of a striker that scores 1.05 goals per 90.

For next season, 0.51 goals per 90 seems a reasonable estimate. The problem is, next season Finnbogason will not be playing at Heerenveen, as he will make the step up to a bigger league, where he won’t contribute the same number as in the Eredivisie. His true level should then be estimated somewhat lower than  0.51 goals per 90 minutes, and we will all start wondering what is going on with all these high scoring strikers who just don’t cut it outside the Eredivisie.

 

Exceptions

Inevitably, though, there will be players who seems to disprove the workings of NPGAR. We can assume that half of all players will have a positive NPGAR and half will have a negative NPGAR. A season later, one quarter of players will have two consecutive positive NPGAR seasons. One eighth will have three consecutive seasons where they outperform ExpG, and so on.

In this study among players from top-5 leagues with at least 10 shots, we find 479 players. With such a big group of players, there will inevitably be some players who consistently outperform ExpG to produce season after season of positive NPGAR. This is a misleading situation, as these players will be credited with finishing skills that are basically the product of an unrepeatable effort.

 

In the end

The message in striker scouting is quite clear. Familiarize yourself with the terms ExpG and NPGAR and these mistakes of flopping striker are generally avoidable. Stay away from strikers with high NPGAR and aim for those with high ExpG numbers, as the latter group will cut it next season, while the first group has every chance of falling back.

Probably, a negative NPGAR in a player with good underlying ExpG numbers is a sign of a bargain buy. The world will see a striker struggling to convert, and it takes some balls to buy him, but the numbers indicate that a return to scoring form is right around the corner.

24 thoughts on “How to scout a striker?

  1. Sam

    Awesome article. Just a little quibble. To the extent you’re trying to recall baseball stats like VORP with the NPGAR stat, I think you want to use NPGAA (where the “A” is “average”) rather than NPGAR, because a “replacement player” traditionally refers to someone you could sign for free off the street–i.e. someone very bad, but just barely good enough to play.

    Reply
    1. avfcus

      IIRC, “replacement player” as it is used in baseball stats (baseballprospectus.com) refers to an exactly league-average player, not someone free off the street. Context (league) is important too and statistics are adjusted accordingly.

      Reply
  2. Lars

    The non-repeatability of conversion rates is unbelievable. Very surprising result. This result alone makes lets me recommend this article.

    ExpG however does not only tell you something about a player but also about the team he is playing in and the player’s role. Maybe if one would only consider strikers that changed clubs between season 1 and 2 the result would be different. Also differenting between centre forwards and offensive wingers would be necessary, as the former will always have more shots and higher quality shots than the latter. Not sure if this is easily doable but it is a thought I wanted to share with you.

    Reply
    1. 11tegen11 Post author

      Thanks, Lars!

      I agree with you that isolating subgroups of players may be an interesting step. The problem then, as you also mention, is the classify which players play in which position, as there would be a lot of overlaps.
      Also, numbers for analysis will be a lot lower, with all sorts of disadvantages. Still, if repeatability is this clearly absent in the entire group, I’d expect it to be absent in subgroups of any decent proportion too.

      Reply
  3. Ralph

    Good stuff m8 this stuff has some serious implications 4 scouting glad to see smart bros like you doin ya thang congrats m8

    Reply
  4. luk

    Interesting piece. Thanks for sharing.

    I think what could make it even more interesting is if you tried to account for the number of shots in the NPG figure. Perhaps conversion rate is less repetitive if there are fewer observations in any of the seasons?

    Reply
  5. staty

    Great article, as usual.

    There are two questions I am interested in yet. Maybe you have already done research, on these questions, or take them as proposals for new articles.:

    1) How repeatable are goals (NPG) of one player? The answer to this question could illustrate the advantage of ExpG ahead of “effective” goals (or the opposite).

    2) How big is the difference between NPGAR and the convetntional chance conversion metric (goals per shot)? Are they well correlated?

    Reply
  6. Mark

    Great article, very interesting stuff

    For me the main question is whether conversion is indeed non repeatable or whether it is still too difficult to prove the skill of scoring goals. I imagine that even for the most prolific strikers the number of shots taken over a season, combined with the many factors which influence the probability of scoring, might be to small to isolate a goalscoring skill.

    I understand from your other articles that expected goals takes into account shot location, but also type of assist (cross e.d.) and header or shot, but that leaves many other factors which are difficult to model for such as first touch, defensive pressure, goalkeeper positioning and so on. I wonder whether these factors influence NPGAR enough to make it statistically non repeatable given the sample size.

    This wondering is fueled by my believe that some strikers are consistently better finishers than others and also because i remember an article (I believe on statbomb) of someone finding some repeatability in conversion, but only in much larger samples.

    However it does not take away the fact that it might be (currently) impossible for a strikers scout to accurately assess the skill of finishing. This would imply it is better to select strikers on different (measurable) qualities and creates opportunities for bargain deals (As the article suggests).

    Reply
    1. 11tegen11 Post author

      Hi Mark,

      I think that’s very well said.
      The current best way to look at it is indeed that we can’t rule out a difference in finishing skill between players. But, as you said, the impact of the player finishing skill on the outcome (goal / no goal) is very small compared to other factors involved. Therefore, with the small samples that the nature of a football season offers, it is best to assume that there is no difference in finishing skill.
      Beware that nearly all articles looking at conversion before have looked at goal per shot. That parameter is indeed repeatable to a certain extent, but that can be explained by shot quality. Certain players take low quality shots, while others take high quality shots, year in year out. So, some players will indeed convert better in terms of goals per shot, but that should not be interpreted as better finishing. It’s finishing of better chances.

      Thanks again for a quality comment!

      Reply
  7. The Woolster

    Hi Sander,

    I’ve come to the comments section as requested.

    In terms of the repeatability of conversion rates, I am much in agreement with Mark. Although these Expected Goal models perform very well when looking at large samples of all shots, when we start looking at individual players then the variables that we cannot yet measure can have a large effect, so we can never be really sure that a player is over or underperforming expectations, or the magnitude of that performance. Also the performance of the person shooting is not in isolation, as he has a goal keeper to beat. He may be hitting the top corners but the keepers he has faced have made saves that they would not usually do. There is a section on this in this peice by Sam Green, in his sample, Suarez would usually have scored 6 more goals but for goal keeper performance, which can would a large effect on his conversion metrics http://www.optasportspro.com/about/optapro-blog/posts/2012/blog-assessing-the-performance-of-premier-league-goalscorers.aspx

    On the issue of looking for strikers with high expected goals, I think this metric just has too much noise with regards to the tactics fo the team, instructions of what runs to make from the manager, and the ability of their team mates to pass the ball to them in good positions for us to link it to the individual player, so if we simply search for those players with a high ExpG, then I think we would get a number of false positives. As I pointed out, Soldado could be an obvious case. My guess is that he is better than his perofrmances at Spurs are showing, but my guess is that his ExpG is quite a bit lower than it was in Spain as Spurs do not create the same types of chances for him.

    Finally, I think I have come to realise through some of the stuff I have looked at with Clear Cut Chance conversion (unfortunately not written about it yet) and the work on penalties by James Grayson, that the easier the chance, then the less difference in skill we are likely to see. The players used in the sample, we have to assume, are the elite of their profession, they ‘should’ all be able to convert the majority of easy chances they have. I think the finishing skill is more likely to show up in the conversion of more difficult chances, but the conversion rates for these are so low that I am just not sure we will see enough shots from enough players to be able to confirm which ones are the most skilled. But if we could, then these are the ones that we should be scouting, as they are more likely, over the longer term, to finish the difficult and the easy chances.

    Having said all that, it was a very interesting and thought provoking article.

    Cheers,

    Ben

    Reply
    1. 11tegen11 Post author

      Hi Ben,

      Thanks for putting those thoughts out here.

      I think you, and other commenters, are absolutely correct that we cannot say that there is no skill in finishing.
      For now, given the limitation of the data we have, we cannot demonstrate any such skill. Which is still a huge leap in terms of its implications for scouting.
      It’s much like ‘absence of evidence’ versus ‘evidence of absence’.
      Until we have more detailed data on defensive pressure and goal keeper positioning, we’re probably best off by assuming equal finishing conditional on ExpG for all players.

      Your second point already moves to the next question we’ve called upon ourselves.
      If we can’t scout strikers by their finishing, how can we scout them? I’m convinced that it’s somewhere in the generation of ExpG, but we can’t simply say that strikers producing more ExpG are the better players, because of team effects. I’ll write on this topic in a follow-up article, once I feel I have enough data to start exploring how to separate player performance and team performance. Beforehand, this seems very challenging without any off ball information.

      Once more, thanks for your comment!

      Reply
  8. The Woolster

    I think one possibility to for seeing if their is a ‘skill’ in shooting will be to see if something like Colin Trainor’s ExpG/ExpG2 efficiency is repeatable, ie are players able to take better/more accurate shots (hit the corners) than average consistantly. This would also partly take out the issue of goal keeper saving ability.

    Reply
  9. James (@RowleyBushes)

    Great article!

    One possibility you leave open is that player-specific expected goals is repeatable season-on-season, but that player-specific goals above replacement is not.

    This would accord with many ways we think about strikers non-statistically. For instance, it’s often thought that a good goalscorer is someone who makes late runs into the box. The temptation is to get in a goalscoring position early, and it’s plausible there’s a genuine differentiation of skill between players rush in and players who are cool-headed enough to wait. A player with above-average expected goals may also be better at reading the game i.e. not running into a standard position early but finding space or waiting for the ball to break to him. It’s hard for me to know whether this is possible without knowing what’s in your EPG metric but we all know of cases where strikers have picked up goals by latching onto loose second and third balls.

    Reply
  10. Pingback: The Numbers That Show England’s World Cup Performance Was Much Better Than You Think |

  11. Pingback: By the Numbers: What to Make of Mattia Destro? - Chelsea Index

  12. Pingback: Examining Conversion Rates | Points

  13. Pingback: Målskytte och underliggande siffror på spelarnivå – Premier League | fotbollssiffror

  14. Pingback: Is Michiel Kramer de gedroomde spits van Feyenoord? | Leven met Feyenoord

  15. Pingback: By the Numbers: What to Make of Mattia Destro? - Chelsea Index - Chelsea Index

  16. Pingback: Een eerste expected goals (xG) model | persoonlijkefouten

  17. Pingback: Using Expected Goals and Simulations to Guess How Leigh Griffiths and the SPFL’s Top Strikers Will Perform – The Backpass Rule

  18. Pingback: What can we learn from attacking statistics? - All Sports News

Leave a Reply