The story of a football match in one picture

Football matches are not just football matches. They are stories, they have identities and each match is a unique experience. No two football matches are the same. Yet, it the end, results of football matches are not all that different from each other. We’ve only got three outcomes: a home win, a draw and an away win. And we’ve got a fairly limited set of score lines to go with that.

But one 2-0 home win may be vastly different from another 2-0 home win, and we all know that some 0-0’s are a genuine waste of precious time, while others are the outcome of an intriguing battle back and forth. And then you have score lines like a 1-2, which may have seen a comeback, or not.

You get the point: the final score line does not suffice.

 

Match report

Last weekend, Eredivisie Champions Ajax beat newly promoted side Cambuur 1-2.  If I were to tell you about that match, I could say that Ajax created next to nothing offensively, while Cambuur had their share of low quality chances, up until Ajax opened the score shortly before half-time with their first shot of the match.

In the second half, Ajax put on an abominable offensive show, but on the other hand, Cambuur could not put anything together either. Until the final fifteen minutes, that is. Cambuur managed to equalize with their best chance of the match. In the end, Ajax was lucky to get away with a dying seconds winning goal, as their total offensive output did not even match their underdog opponents.

Now, if I’d use this as a base for my match report, add in a little background info, some smart language, specific match details, player names and a tactical twist here and there, I’d say I have a decent match report.

Well, what if I did not say anything, and just presented this image?

ExpG stands for Expected Goals, where each attempt to score is assigned a probability of resulting in a goal, based on location, type of shot, assist details and some other factors.

ExpG plot Cambuur - Ajax

Please, take a moment and re-read the paragraph above with this image in mind, if needed, just click to enlarge. Indeed, it’s all in there: the opening goal being Ajax’ first shot of the game, the low quality of shots created, the long dull phase after the opening goal, Cambuur scoring with their best chance of the game, and the fact that Cambuur created more than Ajax.

I think this is as close to a box score as we can get in football, but please don’t hesitate if you’ve got any improvements in mind. A short summary of the ebb and flow of a football match. In the era of short chunks information, I think this will prove a useful visualization and I hope to use it a lot in the future.

Oh, and there is another advantage, as this picture tell you exactly how a match went. Not how another observer thought it went, no room for shaping up a match report to fit the narrative. Just data in, a little automated computing and graph out, that’s it…

So while we’re at it, here are the other Eredivisie matches of the past weekend, so you can just pick your team… Click the picture for a full size version.

ExpG plot Zwolle - HeerenveenExpG plot N.E.C. - Roda JC ExpG plot Twente - Go AheadExpG plot Den Haag - RKCExpG plot Feyenoord - GroningenExpG plot Heracles - AZExpG plot Vitesse - NAC BredaExpG plot FC Utrecht - PSV

Data comes from various OPTA based sources.

4 thoughts on “The story of a football match in one picture

  1. Tom

    Nice! So is there any correlation between the difference between ExpG and actual score at a given point in the game and a team’s likelihood to score? For example, in the 75th minute of a 0–0 game, Team A has an ExpG of 2 and Team B an ExpG of 0.5. Is there a greater likelihood for Team A to score? Would the answer to this question be different if it was already 2–0 for Team A?

    Reply
    1. 11tegen11 Post author

      In a broader sense, yes, ExpG and goal scoring are quite well correlated. However, the shorter the scope, the less correlation you will find. So, between ExpG at a certain point in the match, and goals scored later on, the correlation will be quite weak, as ‘random’ events play a big role. On the longer run, most notably ExpG after X games and points at the end of the season, the correlation will be quite strong. But it’s the same underlying link at work.

      And your second question is a different one, albeit a very intriguing one. This refers to the concept of Game States, which is very much still unexplored territory in football analysis. Depending on tactics applied, so with variance among different teams, yes, in general team that hold a lead have a better chance of scoring more goals, as their opponents need to take risks to get back in the game.

      Reply
  2. Tom

    Yeah, that second point is sort of what I was trying to get at with my first point, I just didn’t make it very clear. I was thinking more along the lines of whether a team with a ‘deficit’ of goals scored (i.e. if their ExpG suggests they ‘should’ have scored more goals than they actually have) would be more likely to score goals in the remaining part of the match, and how the game state would affect this.

    Thanks for explaining, and it’s really nice to see some posts appearing on your blog again.

    Reply
  3. Pingback: Statische Methoden zur Leistungsbewertung im Fußball: Expected Goals » abseits.at

Leave a Reply