Goalkeeper Mishmash

So one of the comments on my previous post about goalkeepers talked about how the relegated teams (Wolves, Bolton and Blackburn) had the worst keepers. So I wondered how they would have done had they had better goalies. I’ve still not figured out how to correlate a goalie’s distribution success to goals scored and so I’ll simply stick to shot stopping criteria.

I use the ratio of big chances to goals in each game to figure out how a different goalkeeper would have reacted. So if I have a goalie with a 90% shot-stopping ability and the opposing team has 10 big chances in the game, then I concede 1 goal. However, if my goalie has a 50% stopping ability I let in 5.

Based on the shot-stopping success ratio of each goalkeeper and the number of big chances faced by each team in each game, I have estimated the number of goals the team would have let in in each game. Comparing this against goals scored, I have come up with a hypothetical points tally for the season.

I know I abuse excel graphics a lot but I couldn’t think of any non-excel method to present the data here. I paired each goalie who played at least 1000 minutes during the season with each team and estimated how many points the team would have raked up.

Goalie Mishmash

Some pertinent observations.

1. The teams on whom the quality of goalie had the most impact are Arsenal, Blackburn, Wigan and Wolves. This goes to show how much Arsenal have to credit Sczsesny for their ability to reach the Champions’ League.

2. Everton is the team where the maximum and minimum possible points due to change in goalie is minimum (4, opposed to 14 for Arsenal). Shows that they have a pretty compact and tight defence, and what stops them from a top four slot is the quality of attack.

3. Due to the low number of big chances that occur in each game and due to rounding of goals conceded, you see some kind of a discontinuity in scores as you go down the list, as well as lots of ties. There is no mistake in the data or the calculations.

4. Manchester United has a much lower “goalkeeper impact” than Manchester City. With a lesser goalie than Joe Hart, it is unlikely City would have won the title.

5. Since we use overall averages of a goalie’s shot stopping ability, these simulations show different numbers for “real” goalie-team pairs than what the teams actually achieved.

6. The difference in maximum and minimum possible points as a function of a goalkeeper is a good indication of the overall quality of a team’s defense. The table below ranks the teams as per quality of defense.


7. While Blackburn and Wolves both had poor defence, part of Bolton’s relegation blame can be attributed to the quality (or otherwise) of their goalkeepers (Adam Bogdan and Juusi Jaaskaleinen). Which makes it even more surprising that West Ham (upon re-entry to the Premier League) sold Robert Green (to QPR, where he warms the bench) and recruited Jaaskaleinen in his place.

8. Last season, Liverpool had a pretty good defence (especially their first-choice back four of Johnson-Skrtel-Agger-Enrique). Their attacking ability (and especially their finishing – same story this season) let them down badly.

2 thoughts on “Goalkeeper Mishmash”

  1. Good stuff da.

    Is there an already existing field called “big chances”? Or is it just shots on target? (sorry for so many questions, but it is very easy to ask questions about someone else’s model)

    You can also get some idea of error in the model by comparing your predicted goals with the actual number of goals.

    1) Villa would have been relegated with almost every keeper and if Robinson hadn’t played for Blackburn. And QPR should have survived comfortably. But then, the last day would not have been the fun that it was.
    2) Last season, Hart was indeed responsible for keeping City in the game in many matches. One save against a Carroll header stands out clearly in my memory.
    3) Two things about Jaaskelainen – he was out for a long time last season and he used to work for Sam Allardyce before.

    1. 0. There are separate columns for “big chances faced” and “big chances”.
      1. Yeah Robinson seems to have become extremely cuppax. Can’t believe he once kept for Spurs and England. And Villa defence has been cup for a long time now. Basically ever since they got that mancity irish guy – forget the name – dunne.
      2. i remember that too. heartbreak happened.
      3. yeah the work with allardyce is what got him to whu for sure

Put Comment