Coach was Right: Aim for the Far Corner

When I was a youth player a recurring theme from my coaches was to aim for the far corner when a shot opportunity arose. So as a player who typically played on the right side of the field, most of my shots looked something like this:

rightSideToOppositeCorner

As I grew older I came to suspect that aiming for the far corner was really just a coaching ploy to prevent the player from thinking too much about picking out their shot location. However, it turns out that there is quite a significant difference in shot results depending on which side the shooter picks out. By analyzing more than 200,000 shots (throwing out the ones blocked by defenders or goalkeepers) from European league play the following near and far corner shot results were produced:

Shot Results When Shooting At The Near or Far Goal Post
Aiming Near Corner – Goal % Aiming Far Corner – Goal %
16.6% 18.3%

So what causes this disparity in results? I suspect the main reason has do with the preceding play’s momentum. As the ball swings out towards the wing, there is a subtle shift of the goal keeper towards the play. This means when the shot comes back across the goal the goalkeeper is initially moving in the wrong direction like this:

goalKeeperMomentum

This suggests that there are opportunities for goal keepers to improve their results by not overshifting towards the play. It also suggests that young players should listen to their coaches.

Press Your Luck: The Surprising Importance of an Early Yellow Card

Rafael da Silva was awarded the very first yellow card of the match in the 18th minute of Manchester United’s last game in the 2010 UEFA Champions league quarterfinal against Bayern Munich. With Manchester United winning 2-0 in the early stages of the game and leading on aggregate there was little concern from the home team for the early yellow. The two teams traded goals just before half and then headed to the locker rooms with the score at 3-1.

The tide began turning early in the 50th minute when Rafeal was shown a red card for pulling on Frank Ribery’s jersey. A goal in the 74th minute by Robben tied the two teams on aggregate and eventually pushed Bayern Munich through on away goals against the shorthanded squad.

The importance of Rafeal’s early yellow card becomes more apparent when you take a look at the following table:

Average Time to Straight Red Card 14739 Minutes
Average Time To Second Yellow Card 1021 Minutes

A player with a yellow card is about 14(!) times more likely to get sent off in a match as any other non-yellow carded player on the field. In the case of a player receiving a yellow card before or during the 20th minute of a match, the data shows that the player has a 6% chance of receiving a red card sometime during the remainder of the match. The players that do receive red cards play on average about 42 additional minutes before receiving their marching orders. So what does this mean in terms of winning or losing you might be wondering?

In order to figure this out, we need to know the value of a red card over the course of a game. There has been a lot published on the value of red cards so I won’t recount it here. For this example, we’ll assume that playing a man short is worth about 1 goal over the course of a full 90 minute match between teams of equivalent strength. In the case of Rafeal’s early yellow we would expect him to miss about a third of the game 6% of the time. This means Manchester United took 0.02 goal differential hit in the aforementioned Champions League match.

To be fair, there are significant differences in card issuance rates between leagues. For example, La Liga saw about 5.2 yellow cards issued per match over the last seven years, while the relatively peaceful English Championship saw only 2.95. Therefore you would need to make league adjustments, but in the end the conclusion is the same. The early yellow card is not nearly as benign as most fans and coaches assume.

The Case Against Ball Possession

One of the little explored areas of the modern soccer analytics movement is centered around how playing styles impact performance. When you watch a possession oriented team like FC Barcelona play against a counter attacking squad like AC Milan in last years UEFA Champions League matchup the advantages and disadvantages of each playing style become apparent.

In the first leg of their knockout round matchup, the away team Barcelona possessed the ball for 72% of the match, but still walked away losers 0-2. In the second leg of the knockout round, Barcelona again dictated most of the action possessing the ball 66% of the time but this time managed a victory 4-0 to advance to the next round.

So the question remains what effect does a quick counter attack have on a team’s ability to put the ball in the net? Looking at thousands of games from top flight European teams, a model was created to examine how quickly a shot is taken after the ball first crosses the midfield line and what impact that had on the outcome of the shot. So for example, if a team recouped possession in their own half and then crossed midfield a stopwatch is started to see how long they take from crossing the midfield line to taking their first shot. From this information the following results were produced:

Soccer Counter Attack Shot Expectation

% of Shots that are Goals 10.81% 10.83% 10.02% 8.83% 8.48% 8.52% 7.80% 9.28% 6.91%
Seconds after crossing midfield 0-4 5-9 10-14 15-19 20-24 25-29 30-34 35-39 40-44

As you can see from the results above the faster that a team gets a shot off, the higher the probability that the shot will result in a goal. You start to appreciate how different the results are for counter attacking teams once you look at how a long lasting ball possession impacts shot results. A team that held the ball for more than 40 seconds decreases their results by about 40% over a quick attacking team.

The reason for this difference has to do with the defensive positioning in response to a counter attack. When the ball is moved towards the goal with pace the defense is often caught out of position. The longer the team in possession gives the defense to get settled in front of their goal, the worse their results become.

So the next time you see your favorite team with a mere 28% of the possession like AC Milan against Barcelona don’t get discouraged. A team that counter attacks quickly is far more likely to see an individual shot wind up in the net.

Super Sub or Lousy Team?

There is a growing body of evidence that game states in soccer are extremely important (but often overlooked) in most player analysis. As teams gain the lead or lose the lead they tend to make tactical adjustments in how they play the rest of the match.

The table below details how MLS teams in 2013 have performed when leading or trailing by a certain number of goals. Each MLS game was broken into segments with a certain score differential and all the shots by both teams were counted. The percentage is the number of shots that a team takes depending on the goal difference in the game:

Team Name Trailing By 2 Trailing By 1 Tied Leading By 1 Leading By 2
Kansas City 67% 75% 65% 54% 59%
Portland Timbers 89% 58% 50% 41% 35%
Houston 56% 57% 53% 37% 38%
Montreal Impact 55% 44% 51% 42% 37%
Toronto FC 52% 46% 44% 26% 18%
DC United 62% 49% 42% 43% 0%
Seattle Sounders FC 24% 54% 50% 43% 48%
Philadelphia Union 48% 65% 53% 49% 32%
Chicago 65% 55% 49% 48% 48%
Real Salt Lake 67% 61% 51% 51% 53%
San Jose 57% 50% 56% 34% 20%
Vancouver Whitecaps 44% 60% 44% 44% 42%
Chivas USA 36% 44% 36% 29% 18%
New York Red Bulls 68% 47% 50% 47% 39%
Columbus 70% 66% 48% 51% 34%
New England 56% 65% 42% 44% 37%
Colorado 83% 70% 49% 45% 37%
FC Dallas 55% 50% 50% 42% 36%
Los Angeles 71% 55% 68% 61% 50%
League Averages 58% 55% 50% 45% 42%

It is probably not suprising to most MLS fans that Kansas City and Los Angeles are extremely dominant when the score is tied. They both take almost 2 out of every 3 shots when the game score is tied.

The worst teams so far in 2013 also make a notable appearance in the table. As Chivas and Toronto take the lead, their ability to create shots diminishes substantially. Once either of those two teams takes the lead by just a single goal they only account for 1 out of 4 shots taken in the ensuing play.

The contextual analysis of soccer games is only in its infancy. You could imagine a new generation of player stats which take into account what the score is when that statistical event happened. For example, lets say you are substitute striker for one of the best teams in the world. Whenever you enter the game the score is very likely to be a goal or two in your favor. This means that tactical shifts by your team are likely to limit the number of shots that you get to take, making your overall effectiveness look worse than it actually is. On the other hand if you played for a below average team, your shot rate would improve the later in the match you made an appearance.

How To Take A Corner Kick Properly

Most fans have no trouble differentiating a horrible corner kick from an OK one, but what sort of kick placement makes a really good corner kick?  The average corner kick only scores about 2.7% of the time, so are there ways to optimize our play to improve upon those scoring chances?

The heatmap below shows the goal rate in relation to the place where the ball is first contacted after the corner kick. This chart includes corner kicks that began on both the left and right sides of the goal by mirroring the field across itself so that all the corner kicks originate from the top of the diagram. The percentages are created by taking the number of goals that occur 15 seconds or less following the initial kick divided by the number of kicks overall to that location. The data was compiled using roughly 8000 games from the top flight leagues of Europe.

A heatmap showing the success rate of corner kicks. All corner kicks are taken from the top corner on the image.

A heatmap showing the success rate of corner kicks. All corner kicks are taken from the top corner on the image.

22.5% or less effective than average corner kick
between 7.5% and 22.5% less effective than average corner kick
between -7.5% and 7.5% as effective as average corner kick
between 7.5% and 22.5% more effective than average corner kick
between 22.5% and 37.5% more effective than average corner kick
37.5% or more effective than average corner kick

Looking at the diagram you can notice several interesting facts:

  1. Getting the ball past the “first defender” is of crucial importance
  2. Keeping the ball away from the area the goalkeeper inhabits is generally a good idea.
  3. A corner kick that lands past the back post is occasionally successful but difficult to execute.
  4. Receiving the ball just slightly before the midway point (towards the kicker) on the field is better than precisely in the middle of the field.

The really interesting conclusion to draw from this is that corner kicks can be made much more effective by highly skilled corner kick takers.  A very skilled kicker should be able to consistently play the ball into the magenta area above resulting in corner kicks that score 40% more often than average.  Using an average team in the English Premier League with about 5 corner kicks per game this works out to 2 additional goals per season.

Of course any sort of analysis like this invites new questions:

  • Does a corner kick taken as an inswinger or outswinger perform better?
  • What are the optimal locations on goal to aim a shot following a corner kick?
  • How do particular player attributes like height affect corner kick performance?

On these pages we’ll investigate these problems and many more. Welcome to the Soccer Perfected Blog!