Statistical trends in Goalkeeper Distribution

written by – Rangers Report    photo courtesy of –Reuters

This tweet from Sam Jackson this morning reminded me that I never actually published the results of my very first tracking project back in 2015.  (Note:  Jackson is easily the most insightful writer out there who applies statistical analysis to goalkeeping…go do a deep dive on his work).

I began dabbling in analytics in the summer of 2015 & wanted to find a doable entry point to tracking what occurs in a football match.  After watching Rangers goalkeepers consistently punt the ball up the pitch to only see “50/50 battles” lead to the other team gaining possession was a real point of frustration.  Why they did this in the lower leagues was a serious pet peeve of mine?  If you are clearly the team that should dictate play, why make it easy on the other team  to get the ball?

This inspired my initial dabbling at tracking matches.  I wanted to keep it simple & straight forward.  I would track how the keeper passed the ball & then record at what rate the team was able to establish possession in the attacking half following the goalie’s distribution.

I dubbed the rating Goalkeeper Distribution Rating (GDR) (I really struggle at coming up with creative names) & began releasing the results as part of my match reports that I published on the site.

Given Mark Warburton’s deliberate attempt to have Rangers play out of the back, well….he actually cited my work to justify the team’s approach.

“The stats I looked at were that Wes played out 87 per cent of the time yesterday and we got started in their half 77 per cent of the time from that. These are big numbers. Against St Mirren, it was 74 and 88 respectively. It lets us keep the ball in their half. If our keeper launches the ball and it’s a 50/ 50 battle, that’s not for us.”

For six months, I tracked goalkeeper distribution anytime I knew I had the opportunity to watch a full match.  The following results come not only from Rangers matches but also from the EPL, MLS, the Gold Cup, Copa America, other SPFL games & some other random one-offs.

I tracked a total of 2,880 goalkeeper distributions & the overall result was that 44% of them led to the team establishing possession in the attacking half.

Here are the stats broken down by the type of distribution, what percent of the total distributions it was & the GDR (the rate of how often the distribution led to possession in the attacking half)

  • The go-to route for goalies in this study was long balls – 60% of all distributions were your traditional long kick up the pitch.  The results suggest that this 50/50 battle was actually a 28%/72% battle.  The vast majority of time a team punted the ball up the pitch they were giving the ball right back to the opponent.  Obviously, if you are the inferior possession team in a match-up this can relieve pressure & move the flow of play further away from your goal.
  • Short passes were the next most popular route & the rate of getting the ball into the attacking half more than doubles.  Of course, not every team can pull this off.  Sometimes the goalie isn’t as comfortable playing balls out of the back.  For example, the team I consulted last season wanted their keeper to play more out of the back & it must not have been in his comfort zone because he still relied almost exclusively on long balls.  For other teams, the defenders just aren’t capable of playing out of the back once they see any amount of pressure.  Sometimes, managers haven’t implemented a tactical plan to help support playing out of the back.  It’s not for everyone, but the data suggests it can be a much better route than the long punt up the pitch.
  • Back to Sam Jackson’s tweet that spurred my revisiting these results.  A goalie that can distribute via his hands is creating a tangible advantage for his team to get the ball up the pitch.  Of the 268 rolled balls tracked, 184 (69%) led to possession being established in the attacking half.  That rate is pretty consistent with the 821 short passes that were played with the goalie’s foot.
  • Goalkeeper coaches are likely already working with their keepers on playing short balls with their feet but I wonder how many are taking the time drilling their keepers to throw balls to teammates.  Think about it, of the 2,880 distributions tracked only 76 (!!!) were thrown balls & nearly every time it led to possession in the attacking half (79%).  It’s a small sample but it’s worth exploring more.
Derek Gaston

While in the midst of collecting this data, I reached out to Derek Gaston (Morton’s goalkeeper since 2012) for his insight on the decision making behind how to distribute the ball & to what extent a manager dictates how he wants the ball played.

Gaston’s response was, “I’ve had coaches in the past who have only wanted me to try& play short passes.  I’ve also had coaches who have only wanted me to kick the ball long.”

“But most of the time it is left to me to make the decision on where I kick the ball.”

“Sometimes if the opposition press high up the pitch then you have no option but to kick the ball long. Most of the time it’s just about assessing the situation at the time and deciding what method is more beneficial to the team.”

That last point would dictate how I would alter this kind of study if I ever did it again.  It would add so much more meaning if you included the game state (the score of the game) to the results along with whether the other team was pressing the defense/goalie or not as he made his distribution.


One thought on “Statistical trends in Goalkeeper Distribution

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s