The following is a blog post from John De Witt Performance Analysis - Evidence Based Conclusions. De Witt heavily involved with the sport he analyzes, as he is currently a performance coach for the Houston Dynamo Academy and a current member of the NSCAA Convention Committee. Additionally, De Witt has earned his NSCAA Premier diploma, US Soccer A license and USYS National Youth diploma and has presented at several past NSCAA Conventions. De Witt is a senior biomechanist with the NASA Johnson Space Center's Exercise Physiology and Countermeasures Group and a member of the NSCAA Convention Committee.

English soccer has been stereotyped as a game that features long passing. This label may be unfair or outdated as the game has progressed with time. Since Opta provides long passing statistics, I thought it would be interesting to look at long passing more in depth to determine if any trends could be identified that related to success in 2011-12.

MUFC & Opta provides frequency counts of successful and unsuccessful long passes for each player in each team. I want to state up front that I am not sure what defines a long pass as I could not find the definition of long pass in Opta’s event definitions. I assume that the pass must cover a minimum distance, but I am not sure that that distance is – if anyone can clarify please post a comment and I will update this blog accordingly. For now, however, we will assume a long pass is one that travels some distance longer than a short pass. I think a reasonable way to think of this is as a pass that bypasses some teammates.

Number of long passes

In the Premiership in 2011-12, there were 23,014 successful long passes, 18,710 unsuccessful long passes and 41,724 total passes. With a total of 380 matches, this equates to 60.6, 49.2, and 109.8 long passes per match. Overall passing accuracy was approximately 55%.

There were no correlations between league finish and successful passes or pass attempts, but there was a correlation between unsuccessful passes and league finish. The negative correlation between unsuccessful passes and league finish indicates that teams that finished stronger tended to have less unsuccessful long passes than teams finishing lower in the table.  Note that in the chart, the slopes of the trend-lines are opposite than the correlations. This is simply because finishing higher in the table corresponds to a lower numerical position. I felt that plotting the data in order of finish makes sense, as does reversing the correlation coefficients.

The larger correlation between finish and unsuccessful passes as compare to successful passes is probably best explained by the increase variability in successful passes from team to team. In general, however, as teams finish better in the league, they make more successful long passes and less unsuccessful passes than teams finishing lower in the table. However, the lack of correlation between passing attempts and league position indicates that teams that finished higher or lower in the table were not more dependent on a long passing attack than teams on the other side of the table.

Passing Success Percentage

Understanding the absolute quantity gives us some quantitative data regarding how often teams use the long pass in the attack. We can also look at the numbers in terms of percentages to give us an idea of how proficient teams are when using the long pass.

Long passing completion percentage was significantly correlated to league finish (r=0.70). This finding coupled with the absolute data suggest that while teams that finish higher in the league table do not necessarily attempt more long passes than teams lower in the table, they are more proficient at the tactic.

Number of long passes as % of all passes

At the start of the blog, I mentioned that the English game is stereotyped as a long passing game. The numbers don’t necessarily agree. Overall, there were 41,724 long pass attempts and 351,867 total pass attempts of all passes, corresponding to an overall average of 11.5 ± 2.6% of all passes being long passes. In other words, approximately 1 in 10 passes in 2011-12 were characterized as long passes.

Correlation analysis reveals a strong negative correlation between league finish and long pass attempts as a percentage of all passes (r=0.74). This finding indicates that as teams finished deeper in the table, they tended to rely more upon the long passing tactic than teams finishing higher in the table. However, the overall amount of long passes was still very small relative to all passes. Stoke City lead the league with 18% of their passes being long passes.

Player analyses

It is useful to look at teams methods of play to gain a better understanding of team tactics and patterns of play. It is also useful to look at the passing statistics at an individual level to determine standards of play in order to assess player performance.

The following tables list the top 10 players in the Premiership with regards to long passing performance in 2011-12. I excluded goalkeepers from the analysis as their long pass attempts are typically much larger than field players and I wanted to focus on the use of long passes during open play.

The first table lists the top 10 players in terms of long passes completed in 2011-12:

The second lists the top 10 in terms of passing completion rate. I limited the analysis to those attempting at least 100 passes throughout the season:

Examining these two tables gives us a few indicators of the top long passing performers in the Premiership. Top players are completing 175 or more long passes through the season. Completion rate for the top 10 individuals are on the order of 79% or higher.  Overall completion rate throughout the league for all players was 60%, but when limited to those with over 100 attempts, the league average completion rate was 62 ± 13%. Therefore, top performers in the Premiership (top 33%) are completing 75% or more of their long passing attempts. Taken as an average per match, the top players are completing 4.6-8.5 long passes per match.

Conclusions

There are several more ways that we can look at the long passing data, which may be useful in future analyses. Looking at the data as a whole, however, gives us some indication of how teams tend to play over an entire season.

The results are interesting and can be interpreted in a variety of manners. I feel that based on the data, the following conclusions can be inferred:

1. Teams in the Premiership tend to attempt about 110 long passes per match, but this equates to roughly 10-15% of all passes in the match.

2. More successful teams complete a greater percentage of their long passes than unsuccessful teams, but do not attempt more long passes. In fact, as a percentage of all passes, more successful teams attempt less longer passes than less successful teams.

3. The top players in the Premiership are very proficient at the technical and tactical aspects of long passing, completing 175 or more long passes with a 79% or higher completion rate.

Based on these results, I am not sure if the long-ball reputation of the Premiership by some persons is substantiated by the data. One thing that I can say is that there are plenty of statistical analyses that show that goals are typically scored after a few rather than many successive passes. These results suggest that it is most probable that the few passes occur due to a regain of possession near the goal rather than from a long pass after a regain of possession deep in one’s defensive half because of the relatively few attempts of long passes as a percentage of all passes.

NSCAA members log in to comment. Not a member? Learn more today.