Eredivisie Top Scorers Chart based on Win Points Added

Traditional top scorers charts rank players simply by counting the number of goals scored, thereby assuming that every single goal scored has equal value. This goes against both emotional feeling and rational thinking. A late winner against superior oppositions ranks higher on both these scales than a goal that extends a dominating home team’s lead from four to five goals.

 

Explaining Win Points Added

During the summer break, we’ve looked into this subject, aiming to design a score based on Win Points Added, estimating the expected points won from a particular match and the difference a goal made to that value. This leads to all goals being valued between 0 and 2, with 0 for the goal extending a lead in a practically already won match, as it doesn’t add anything in terms of points won from that match, and 2 for the goal that turns a level game into a victory just at the final whistle, thereby turning the one points from a draw into the three points from a win.

The value for expected points won was derived from four parameters: opposition strength, venue, timing of the goal and, most importantly, goal differential before the goal was scored. The first two parameters, opposition strength and venue were derived from the average bookmaker odds for the game, while a formula for how these odds develop during the match, as well as the impact of goal differential was derived from the former www.whatifodds.com website. More details on the exact build-up of this system can be found in the original post introducing it.

 

Bony versus Mertens

One player sticks out far beyond the rest. The goals scored by Vitesse’s Wilfried Bony, six in total, have contributed to 4.52 points won for his team. While the average value of all 243 goals scored in the first eight rounds of Eredivisie matches is 0.49 points, Bony’s goals added 0.75 points per goal on average. This represents the fact that two of his goals were among the highest scores possible: a 2-1 winner in the 89th minute against Vitesse and the opening goal in the 88th minute against Heerenveen. Bony scored four of his six goals while the game was level, goals that are appreciated in this system.

In contrast, the conventional top scorer of the Eredivisie, PSV’s Dries Mertens, averaged only 0.18 points per goal, hence scoring a total of 1.97 points with 11 goals. Like Bony, Mertens also scored four goals while the game was level, but these goals were scored rather early in matches where PSV already had favourable odds of winning: Excelsior and RKC at home and ADO away. Mertens’ four goal at home against Roda, scored at goal differentials of 1,2,3 and 5 are excellent examples of goals that contribute very little in terms of win points added, but significantly bias a conventional goal scoring chart.

 

In the end

Ranking top scorers in terms of Win Points Added presents the impact of the goals scored, rather than the absolute number. This seems a better representation of the impact that these goals had on the ranking of teams in the league table. Players scoring goals that win games for their teams will top these kind of weighted scoring charts, representing the superior emotional and rational value of the goals they scored.

 

A .pdf version of the Eredivisie League Table and Weighted Top Scorer Chart is available by clicking here.

6 thoughts on “Eredivisie Top Scorers Chart based on Win Points Added

  1. Thomas Boeschoten (@boeschoten)

    Very nicely done. Also good to explain the formula behind the ranking. Am I correct that, if Mertens would have scored less goals, and therefor PSV wins would not be so confincing, he would have been rated higher on average on the Win Points Added? And maybe even in total? It would suggest that it might be easier to be important in a team that has less convincing wins.

    Reply
    1. 11tegen11 Post author

      The ‘Win Points Added’ are presented as a cumulative score, not as an average value per goal scored. So every goal scored adds points to the total, although some goals add more points than others, depending on the factors explained. Scoring near meaningless goals wouldn’t diminish the Win Points Added. If Mertens scores a fifth goal in, say, a 5-1 win over Roda, this does not impact the value of his opening goal in the same match. Both goals are valued independently, based on the comparison between the expected points won by his team before and immediately after scoring the particular goal.

      Reply
  2. Vitessefan

    Nicely done but I would value not the time a goal is scored but rather if it turns out to be decisive. A decisive goal scored in the 35min is of more value in my opinion than a non-decisive one scored in the 88 min. Which now not is the case when i read that Bony’s goal against Heerenveen has the highest score possible.

    Reply
    1. 11tegen11 Post author

      In the system that you propose, events after scoring a goal influence the value attributed to it. If Heerenveen’s Gouweleeuw would have missed his chance instead of scoring the equalizer, Bony’s opening goal would have had less value.
      In the system as it works at the moment, Bony’s goal is valued according to the expected amount of points Vitesse would take from a home game against Heerenveen with an 88th minute one goal lead minus the points they would take from the same game with the scores level at the same point in the match.
      A goal scored in the 35th minute of a match is valued according to the shift in chances of obtaining three, one or zero points from that match, and not according to whether goals are scored afterwards or not.

      I think goals should not be valued according to events happening after scoring them.

      Reply
  3. Vitessefan

    That’s not really valid logic. You value a late goal higher only because is probably going to be decisive (something which is also a occurrence in the future). At least I can see no other reason to value a late goal different for any other than that.

    So why don’t substitute the probable event (late goal which is probable decisive) for the real event (=is it decisive)?

    Reply
  4. Alberto

    I agree with Vitessefan. For a good example look at the Newcastle-Arsenal 4-4 draw from last season.

    In the current system Arsenal’s 4th goal (the 0-4) would have been of little value, even though in the end it secured 1 point for the Gunners.

    Same way that Newcastle’s first goal (1-4) would haven been of little value, even though without it Newcastle would not have been able to equalise.

    Look also at last weekend’s Lecce-Milan game. After Lecce ran out to a 3-0 lead, Milan’s first goal after half time would probably have been of little value according to the current system. But without that goal, Milan would not have won the game 3-4 in the end.

    The final result of the game SHOULD account for some of the value assigned to a goal.

    Reply

Leave a Reply