Author Names: Liam Dixon (BSc (Hons) Sport Studies) and Dr Laurence Protheroe
A broader measure of pass frequency during build-up play which leads to chances created within a game helps to provide a greater insight into producing goals and inevitably success in football. Previous studies have produced contrasting arguments about whether direct or possession styles of build-up play are most effective in the creation of shots. Coupling style of build-up play with the areas of the pitch which produce the greatest amount of chances allows coaches and analysts to provide accurate and objective feedback to the teams or individuals. The purpose of this study was to provide a comparative analysis of chances created following pass frequency during possession, and the area from which key passes were played to determine which type of build-up play is more effective against predetermined, categorised successful and unsuccessful teams in the English Premier League; showing whether different tactics, if any, are more effective against differing levels of opposition. 36 Full games from the 2014/15 English Premier League on Sky Sports and BT Sport live games were analysed. Teams were categorised as a top 6 or bottom 10 team based on the final standings of the 2013/14 season. The build-up play which leads to a chance created was analysed using a hand notational method focusing on pass frequency per possession leading to a chance (either on or off target) and area of the pitch in which the final pass was played. Statistical analysis using a Kruskal-Wallis test showed no significant difference in passes per possession and chances created between each type of game (X29= 7.586, p=0.576) as both successful and unsuccessful teams adopted a direct style of play to create the most chances, with up to 62% and 78% of chances being created from passing sequences of 4 or less in top 6 and bottom 10 teams respectively. Chi-square tests were used to show a significant association with key pass locations and chances created with regards to both overall frequency of chances for top 6 and bottom 10 teams (X25= 13.637, n= 522, p= 0.018) and key pass location and chances created when playing against varying levels of opposition (X29= 21.11, n= 522, p= 0.012). The results suggest an association in creating chances from specific areas, namely the attacking 1/3 of the pitch, with a combined total of just under 50% of chances created from attacking wide areas for both teams as well as around 30% from central attacking 1/3 areas; suggesting a slight preference towards using wide areas in terms of creating chances. This is most effectively done through a direct build-up to enter the attacking 1/3 area as fast as possible regardless of the level of opposition.
Match analysis has been present for over 40 years within football, with various techniques and technological advances emerging as time progressed (Reep and Benjamin, 1968); from the use of pencil and paper as a hand notation technique to the introduction of computer analysis software such as Prozone or Dartfish to allow for more complex and detailed analysis. Coaches and analysts can use the data collected for their own team or individual’s sake and to feedback the strengths or areas to improve on from previous performances. It can also provide information on future opposition tactics, including ways to exploit their weaknesses through video analysis (Carling, Williams and Reilly, 2005). Detailed quantitative analyses can also enhance performance through the improvement of performer feedback (Franks, 1997); Clemente et al (2012) indicate that the information fed back to the coach must be relevant and important, with a well-designed system using relevant key performance indicators (KPIs) to supply accurate and reliable information to the coach (Hughes and Bartlett, 2002).
Notation can be used to create a permanent record of predetermined events through analysis of either team or individual performances (James, 2006), with the need for such techniques being reinforced by the findings of Franks and Miller (1986). Creating a reliable notation system between a coach and analyst can be of great benefit, helping to provide accurate and objective feedback to those involved (James, 2006). Mere observation of performance or recollection techniques can lead to unreliable and subjective findings. Research by Franks and Miller (1986) suggest that high level international coaches could only recall around 30% of incidents during a match when using observation alone. Observation can also be affected by psychological factors such as the halo effect (Nisbett and Wilson, 1977); a coach will have a positive, biased opinion of his own players when observing, thus altering perceptions of performance.
One main area of analysis in football is the focus on goals and match results (Reep and Benjamin, 1968) which are evidentially linked – essentially goals need to be scored to win a match, yet the difficulty comes in identifying and recognising the performance that led to these results. Notational analysis helps to record events and numerous KPIs such as pass frequency in possession, key pass locations in the lead up to a chance and the actual type of chance created by an individual or team (Carling, Williams and Reilly, 2005). These play a key role in understanding how goals are scored or games are won.
Previous research has helped to create an understanding of which areas of a football match will bring about success, from different approaches to play, a direct or a possession style, or goal scoring patterns by looking into the build-up play leading to a goal. Reep and Benjamin (1968) were pioneers in match analysis through the use of notational systems and KPIs to identify findings through hand notation to compare pass frequency and goals scored. Their findings suggested a direct style of play should be adopted for coaches to provide success. This research was a stepping stone for future research, with various studies agreeing with the suggestion (Hughes and Franks, 2005; Bate, 1988; Pollard et al., 1988). Not all would agree with these findings however, promoting deeper and differing techniques in match analysis (Collet, 2012; Lago-Peñas and Dellal, 2010); concluding a possession style of play is more effective.
More recent research introduced the concept of categorising teams in order to determine whether a certain type of play is more effective. Collet (2012) and James (2006) considered successful and unsuccessful teams during their research; results found possession lengths of 3 to 7 passes seemed more likely to produce goals, especially for successful teams (James, 2006). Statistical analysis was integrated into findings to give more depth, with Hughes and Franks (2005) normalising data to improve clarity as to which style of play is most suitable for creating chances and goals in relation to build-up play. Literature has provided various suggestions with regards to types of build-up play and its effectiveness in creating chances within a game. Taking this into consideration, analysis providing the number of passes per possession is required as there has been very little study in terms of the actual location of key passes/assists which lead to the chances created. Providing this information could help to create more accurate tactics along with a view to developing play by utilising the results of the study.
Taking into account the previous research, it is clear that there is still a debate surrounding the type of strategy which needs to be deployed with regards to creating more chances for both successful and unsuccessful teams. This research looks to achieve the following aim and objectives:
- To provide a comparison of chances created following pass frequency during possession and key pass location between successful and unsuccessful teams, through analysis of games within the English Premier League.
- Determine the differences between successful and unsuccessful teams in pass frequency during the build-up to a chance created (whether this is a shot on or off target).
- Determine which type of build-up play is most effective; either direct or possession style, against both successful and unsuccessful teams with regards to chances created.
- Determine the area of the pitch the key pass was played, allowing a comparison of whether central or wide, if any, is more effective against successful and unsuccessful teams.
2.1 Sample Size
36 Full games from the 2014/15 English Premier League on Sky Sports and BT Sport live games were analysed. They were also recorded on a Virgin TiVo box allowing for extra analysis via replaying required areas of each game. This aided in greater accuracy and reliability when determining key positions and passing frequencies during the games.
Teams were categorised as a top 6 or bottom 10 team based on the final standings of the 2013/14 season – the 3 promoted teams replaced the 3 relegated teams from the division. These categories were decided due to the unpredictable and frequently changing nature of league positions during the season. Data for both teams in each fixture was recorded, with eight top 6 vs top 6, twenty top 6 vs bottom 10 and eight bottom 10 vs bottom 10 matches being observed. This helped to quantitatively assess determinants of different styles of play for both sets of teams against both groups of opposition (Field, 2009). Using just the bottom 6 teams didn’t provide enough games to produce a valid sample size so the use of top 6 and bottom 10 teams allowed for a greater range of games to be analysed whilst keeping a difference in standings and classifications of both groups, as the top teams had finished in the European cup qualification places, whilst still retaining a difference in ability between top and bottom teams.
2.2 Data Collection
The build-up play which leads to a chance created was analysed using a hand notational method and presented in Microsoft excel, focusing on the performance indicators; pass frequency per possession leading to a chance (either on or off target) and the area of the pitch in which the final pass was played (Figure 1). The data collected followed protocols of leading sports data company Opta Stats, using data from open play and did not include set pieces such as corners or free kicks as this can be used as a separate form of analysis during a game – this is due to defenders and attackers both setting up to deal with these separately (Opta Stats, 2013).
Figure 1: Grid to determine the key pass and assist location for chances created during a game.
Definitions of key performance indicators (KPIs) were key to the research and allowed for a consistent and reliable analysis within each game (Hughes and Bartlett, 2002). Performance indicators such as passing frequency and chances created were defined and followed a universal set of terms used in previous research by Reep and Benjamin (1968); Pollard et al. (1988); Hughes and Franks (2005) and James (2006). Within this research, loss of possession was defined as attackers losing the ball through an interception, meaningful touch or tackle by the defensive team or a foul/stoppage in play or a shot at goal. Frequency of passing was defined as a series of successful passes by the attacking team until there is a loss of possession or chance created, but once possession is regained it becomes the start of a new possession. If there are no passes in the possession this means the first attempted pass has been unsuccessful for various reasons or the attacker regained possession and shot without any preceding passes. Finally, a chance created was defined as receiving a key pass leading to an attempt on goal, this can either be on (scoring a goal or a shot within the goal posts) or off (the area outside and including the goal posts and crossbar) target including deflected shots which still head towards goal after hitting a defender, but excluding a shot which doesn’t reach the goal due to being blocked by a defender, with a key pass being the pass that sets up the attack; leading to a chance and an assist being the pass which set up a goal.
Reliability of the data followed protocol from Lago-Peñas and Dellal (2010) using observation through an intra-observer test, analysing a randomly selected match from the sample twice within a 1 week period. Reliability was tested using the intraclass correlation coefficient through SPSS; pass frequency with shots on and off target were compared and results produced an intraclass correlation value of 0.960 and 0.911 respectively – showing the data to be very reliable (Hinton, 2004). Alongside this, pass location with shots on and off target were both compared, both with correlational values of 1.000 for both types of chances – showing that all data is reliable (Hinton, 2004).
Figure 1 shows the separate areas on the football pitch from which key passes and assist were played. The number of areas was chosen via a pilot test to be 9 to include central and wide areas, with data of key pass location categorised into areas shown in Figure 1. This gave an insight into which areas give a more accurate and higher quality set of results and considering both central and wide areas was applicable to this study (Lancaster et al, 2004).
2.3 Statistical Analysis
Methods of statistical analysis followed that of Hughes and Franks (2005) to measure data; Frequency of passing (passes per possession) was measured as 0v1v2v3…8v>8 as this allowed a more accurate data representation, but overall style of play comparison was grouped as direct: 0-4 passes, or possession: 5+ passes as used in Reep and Benjamin (1968).
Chances created with pass frequency per possession were compared along with the total number of times (1/2/3 etc.) a chance occurred over the duration of the study. Comparative graphs were also used to look at which style of play was used most, the number of chances created and key pass locations when playing against similar level opposition (Hughes and Franks, 2005). A comparative line graph of chances was also used with both passes per possession and key pass location to illustrate the differences between the top 6 and bottom 10 teams when playing against opposite levels of opposition (top 6 vs Bottom 10 or vice versa) (Hughes and Franks, 2005). Finally, pie charts were used to show the overall percentage of key pass locations for top 6 and bottom 10 teams.
Statistical analysis was carried out using the Statistical Package for the Social Sciences (SPSS). For all statistical tests, the significance level was set at p<0.05. Research was correlational as there is no influence on the variables or what happens and used a repeated measure design (Field, 2009) – all teams were measured on the same performance indicators each game. These data sets were nominal – recording the frequency with which each category occurred, so non parametric tests were used to compare and analyse for differences and relationships (Coolidge, 2006). A Kruskal-Wallis test was used to show the differences between pass frequency and chances created for both top 6 and bottom 10 teams. This was used not only due to being a non-parametric test but also due to it allowing for varying sample sizes to be analysed among 3 or more groups (Field, 2009). A Pearsons Chi-Squared test was used to determine whether there was any association between location of key pass and type of chance created and the key pass location between the top 6 and bottom 10 sides when playing against each other and the opposing type of team (Field, 2009).
3.1 Direct and Possession Styles of Play
Results show that it is clear that no matter which opposition a team is against (Figure 2), the most effective way to create chances is through a direct style of play (0-4 passes) compared to a possession style of play (4 or more passes), with 62% of chances for top 6 teams playing against similar counterparts, 75% for bottom 10 and 61% and 78% for top 6 and bottom 10 playing against opposing level teams respectively.
Figure 2: Percentage of overall chances created from direct and possession styles of play.
A considerably larger percentage in both situations for bottom 10 teams produced a greater amount of chances, with 13% more chances when against another bottom 10 team when compared to top 6 vs top 6 games and an increase of 17% when playing against more successful opposition.
3.2 Effect of Passes per Possession on Chances Created
A Kruskal-Wallis test showed that there were no significant differences in passes per possession and chances created between each type of game (X29= 7.586, p=0.576). Median values (Table 1) also help to demonstrate that there is no statistical difference as median values only varied between 2 and 3 between all 8 groups. This indicates that the null hypothesis can be accepted as the medians of all groups are equal and there were no statistical differences between each of the groups at p>0.05.
Table 1: Frequency and median value of chances created during each number of passes in a possession.
|Passes per Possession||Chances Created (n)||Median Value|
Irrespective of league position most chances were made from passing sequences of 0-4 (Figure 3). Results show 99 out of 132 chances were created from passing sequences of 4 or less for bottom 10 teams and 90 out of 145 for top 6 teams with the only exception coming from passing sequences of 7 creating 13 shots for these more successful teams.
Figure 3: A comparison of frequency of chances created and passes per possession when playing against same level opposition.
From the descriptive data, successful teams were evidently more efficient, especially after 1 or 2 passes, with more shots on target (15 and 16 respectively) compared to values of 8 and 9 shots off target for these passing frequencies. Overall, bottom 10 teams created less overall shots on target (57) when compared to top 6 teams (78) but had more shots off target (75 compared to 67) meaning top 6 teams were more prolific with their chances created when playing against teams of similar levels of ability.
Top 6 teams had a considerable amount of shots on target when playing against bottom 10 teams (Figure 4). They showed an increase in chances from 25-31 chances after passing sequences of 0-3 with a decline in the amount after 4-5 passes or more, as well as having a vastly greater number of overall shots (200) compared to bottom 10 teams (130) with 75 and 40 shots on target for the successful and unsuccessful teams respectively when comparing descriptive data.
Figure 4: A comparison of frequency of chances created and passes per possession when playing different level opposition.
Although bottom 10 teams playing against a top 6 team produced less shots, the majority of chances (96 out of 130) also came from passing sequences of 0-4 per possession. With the exception of a passing sequence of 3, the trend follows that of top 6 teams in which the number of shots on target decreases after 4 or more passes per possession; suggesting a more direct style of play can lead to more chances created.
3.3 Key Pass Location Results
A Pearson Chi-Square (X2) showed there is a relationship between the overall key pass location for both top 6 and bottom 10 teams (X25= 13.637, n= 522, p= 0.018) – the defensive 1/3 was removed from the test as they did not give sufficient data as very few passes were played from this area. The null hypothesis can be rejected as p<0.05, showing there is a relationship in key pass locations for successful and unsuccessful teams.
The statistical findings also allowed for comparison of actual and expected counts as shown in Figure 5. Similar variances for both variables were evident, with bottom 10 teams creating around 9 more chances than the expected value in the attacking wide right. Top 6 teams produced considerably more key passes compared to bottom 10 teams within the attacking 1/3; 249 key passes leading to chances compared to 169 for bottom 10 teams. They also exceeded expected values in areas, with between 8-10 more chances in the wide attacking left and central attacking areas.
Figure 5: Pitch map to show actual and expected values for chances created in each area.
Evidently, bottom 10 teams attempt more key passes from central areas in the middle 1/3 of the pitch (Figure 6), successfully playing 16% of key passes that led to a chance compared to the 12% of top 6 teams (Figure 7). Any areas producing 3 or less chances were excluded from the results. Both teams look to play the majority of key passes from attacking 1/3 with 26% and 35% coming from the central attacking areas for bottom and top teams respectively, and a combined figure of 48% and 49% for both sets of teams coming from the wide areas.
Figure 6: Percentage of key pass locations percentages for bottom 10 teams.
Figure 7: Percentage of key pass locations percentages for top 6 teams.
3.4 Relationship Between Key Pass Location and Chances Created
A Pearson’s Chi-Square (X2) showed significant statistical associations between key pass locations and chances created (X29= 21.11, n= 522, p= 0.012). The middle 1/3 (areas 4, 5 and 6) were grouped together and the defensive 1/3 areas were removed from the test as they did not give sufficient data as very few passes were played from this area, with the main focus on the attacking areas of the pitch being reported individually. The null hypothesis can be rejected as p<0.05, thus illustrating that there is an association between key pass location and chances for both top 6 and bottom 10 teams.
Top 6 and bottom 10 teams produced most shots from the attacking 1/3 of the pitch when playing a team of the same level (Figure 8).
Figure 8: A comparison of key pass location with frequency of chances created when playing against same level opposition.
The top 6 teams had an overall greater number of shots (129) when compared to bottom 10 teams (110). From wide attacking areas, 61 chances were created by top 6 teams, with raw data showing 29/61 of chances created from attacking wide areas (areas 7 and 9) were on target, but the central attacking area (area 8) produced 20 shots on target compared to just 21 off target. Using the attacking wide areas produced more shots for less successful teams compared to using central attacking areas (52 chances compared to 33).
Most chances created were from the attacking 1/3 of the pitch: 149 chances for top 6 teams and 84 for bottom 10. Top 6 teams produced more than double the amount of shots on target when playing a bottom 10 team compared to the other way around (Figure 9), 55 shots and 27 shots correspondingly.
Figure 9: A comparison of key pass location with frequency of chances created when playing different level opposition.
A comparison of Figures 8 and 9 show a large increase in the amount of chances created when playing less successful teams instead of more successful; increasing from 129 shots to 175. However, there was no change in the total amount of chances created against either opposition for the bottom 10 teams, with values staying at 110 and 114 chances.
4.1 Discussion of Results
The aim of this study was to provide a comparison of chances created following pass frequency during possession between successful and unsuccessful teams. Alongside this, the study aimed to find whether a direct or possession style of play was most effective when playing against both categories of teams and determine if key pass locations in specific areas of the pitch were influential in creating chances on goal. Results showed that there was no statistically significant difference in passing frequency and chances created for both the top 6 and bottom 10 teams involved but descriptive data shows a direct style (0-4 passes) is most effective in creating chances for both successful and unsuccessful teams. A significant association when comparing key pass locations leading to chances between top 6 and bottom 10 teams was found, alongside a significant association when looking at chances from key pass locations for top 6 and bottom 10 teams, suggesting a relationship in creating chances from specific areas, namely the attacking 1/3 of the pitch.
4.2 Build-up Play Leading to Chances Created
An analysis of pass frequency in the build up to a chance created supported the claims of Reep and Benjamin (1968) that a more direct style of play is most effective. Results show that over 60% of top 6 team chances come from passing sequences of 0-4, with over 75% for bottom 10 teams, regardless of the opposition they were facing.
No significant statistical differences were apparent when comparing the differences in passes per possession and chances created between each type of game. This disagrees with the findings of Hughes and Franks (2005) and Jones et al. (2004) that successful teams typically had longer possessions before a chance was created. Although top 6 teams playing against a bottom 10 team produced 36 chances from passing sequences of 7 or more this was not the main finding of the research, as the majority of chances arose from passing sequences of 4 or less (90 chances created). This greater passing sequence could be due to various factors influencing performance such as being in a winning position and holding out for the win as suggested by Lago (2009). Shots on target are one of the main indicators of success within football and are best at discriminating between successful and unsuccessful teams (Lago et al. 2010). Descriptive data from the study reinforces these findings, showing top 6 teams almost doubled the amount of shots on target when compared to bottom 10 teams, 75 and 40 shots on target respectively.
The work of Pratas et al (2012) is supported by the results that a more direct style of play is adopted by unsuccessful teams when playing a higher level opposition; 96 out of a total of 130 chances playing against a top 6 team coming from passing sequences of 0-4 per possession. Although this data shows less chances are created when playing against top 6 teams, this could be due to a lack of ability in producing shots or because of factors such as higher levels of tactical organisation or defensive ability of the opposing team as shown in Lago (2009). With both categories of teams preferring a more direct approach, top 6 teams were more creative, especially when facing unsuccessful teams, producing 31 more shots more compared to playing against similar level opponents.
4.3 Key Pass Location and Chances Created
Ball possession is identified as a key factor in improved performance, especially for successful teams (Lago and Martin, 2007), however results from the study have disagreed with this when comparing successful and unsuccessful teams as both adopt a direct style of play of passes of 0-4 passes. This shows key pass location has proven to be more influencial in creating chances. Even though passes from the middle 1/3 of the pitch can create chances, the work of Olsen and Larsen (1997) and Wright et al. (2011) are supported in the results, the most efficient style of play is to enter the attacking 1/3 as quickly as possible to create the most amount of chances. Coupled with the findings of passing sequences of 4 or less, it is evident that this study suggests that this is the most effectual way to create chances when playing against both same and different levels of opposition. A significant association was found between key pass locations for both top 6 and bottom 10 teams. This was evident with both teams creating most chances from the attacking 1/3 of the pitch, with successful teams exceeding expected values for chances created in the middle attacking 1/3 area. This further reinforces the supported findings from Olsen and Larsen (1997) and Wright et al. (2011).
Although there is a lack of previous literature that take into consideration key pass locations on the pitch leading to chances created, the findings of Konstadinidou and Tsigilis (2005) were confirmed such that crosses from wide areas can lead to shots on goal. A combined total of just under 50% of chances created were from passes in the attacking wide areas for both teams as well as 26% from central attacking 1/3 areas for bottom 10 teams and 35% for top 6. This suggests a slight favourability to using wide areas in terms of creating chances. When looking at a more in depth analysis, there was a significant association in location of key passes regardless of the levels of opposition they were facing.
The use of grid templates can help to give visual aids to key pass locations, but limitations still remain within the study. This is firstly with regards to the type of pass played as many different types of pass can produce varying results such as through balls, lofted passes or short passes and the areas in which passing sequences were mostly utilised before the key pass was played. Secondly, formations and match situations can affect these results (Lago, 2009) with a formation involving no wingers potentially suggesting a team is set up to play through the middle of the pitch instead of using wide areas for example. Lastly, it may be effective to enter the attacking 1/3 as quickly as possible, but as suggested by Olsen and Larsen (1997), this could also have a negative effect with regards to levels of ability. There is a potential for an increase in the amount of times a team loses possession with these tactics, especially if the players do not have the ability due to the quick nature of the attacks.
4.4 Tactical Approach
Chances created are believed to be a key indicator of success in modern day football, so creating more chances will lead to more success (Casamichana and Castellano, 2013). This is reinforced in the results as top 6, or more successful, teams created considerably more chances compared to bottom 10 teams. Further supporting this claim is the work of Hughes and Barlett (2002), who found a relationship between creating chances on goal and scoring goals.
Results of the study agree with previous findings of Reep and Benjamin (1968), Bate (1988) and Pollard et al (1988) that a direct style of play produces the greatest amount of shots for both successful and unsuccessful teams, helping to conclude that entering the attacking 1/3 of the pitch as quickly as possible can help to produce more shots on goal. This could suggest adopting a direct style of play can also lead to more goals scored if a greater number of chances are created from this style of play. One limitation that can alter the results is set plays such as corners and direct freekicks. This study excluded set plays from the data as the main focus was from open plays in matches (Jinshan et al, 1993) which could make for a different tactical approach if teams show a weakness to set plays.
An issue arising with the classification of successful and unsuccessful teams is the ever changing nature of football including playing styles (Yiannakos and Armatas, 2006). The need for stability is key in categorising teams, but the difficulty comes in the reliability of the constant changing of the current league tables, a bottom 10 team from last season could be within the top 6 rankings of the current season table. Results of the current study were validated by previous research categorising levels of success and integrating this into studies (e.g. Collet, 2012; Lago-Ballesteros and Lago-Peñas, 2010; Hughes and Franks, 2005) identifying that bottom 10 teams produced considerably less shots when playing against a top 6 team (130 shots from 20 games) than compared to playing against a similar bottom 10 team (130 shots from 8 games). This suggests that different tactical implementations could be used when playing higher levels of opposition to improve performance. Even though this has not been measured in the study, consideration must still be given to the ability levels of defenders and attackers which can have an effect on performance (Collet, 2012; Pratas et al., 2012; Hughes and Franks, 2005) as well as other factors which can produce limitations to playing styles and thus chances created, such as match location and current match status (Lago, 2009).
Another problem comes with the sample sizes chosen and what sample size will allow for a typical performance. Reep and Benjamin (1968) used 578 matches across different leagues, whilst research by Hughes and Franks (2005) used international tournaments with a combined total of 116 matches from 24 and 32 teams being sufficient to produce valid results. Collet (2012) used 5 elite leagues from different countries, but this produced contradictory results due to differing levels of opposition and tactical approaches within each country. Hughes et al (2001) suggest the need to build a performance profile with research by Larsen et al (2000) using just 4 sample sizes compared to the much larger quantity used by research from Reep and Benjamin (1968) for example. With the current study using 36 games there is no clarity in what sample size will produce the most reliable results, but greater quantities of data within studies could help stabilise findings (O’Donoghue, 2005).
4.5 Future Research
Future research should aim to provide greater focus on the types of key passes and the location of the shots using a grid template. This allows for a greater analysis as to whether teams are limited to shots inside/outside the penalty box resultant of a through ball or cross for example as there is very little research into the locations of shots and which are the most effective to integrate into match tactics (Konstadinidou and Tsigilis, 2005; Jinshan, 1993). Using a different grid template focusing on narrower wide areas and a specific area designated to the penalty box can also produce more accurate data with regards to pass location and with the introduction of shot location on top. Analysing these greater quantities of data could help to validate and stabilise the data with regards to build-up play and chances created (O’Donoghue, 2005).
5.0 Conclusion and Application
In summary, the comparative analysis of pass frequency with chances created may reveal any differences between successful and unsuccessful teams but suggestions of the effectiveness of a direct style of play is evident throughout in order to create shots, which are considered to be an important factor of success in football today. This study has analysed build-up play statistics and key pass locations leading to chances created in successful and unsuccessful teams in the English Premier League, showing what tactics are used when facing differing levels of opposition and whether different approaches are deployed with regards to the aforementioned teams. It is evident within the study that most successful attacking play (in terms of chances created) is developed from entering the attacking 1/3 in as few passes as possible, usually between 0-4 passes, with a slight preference to using wide areas being most effective in creating shots on goal. Coaches could benefit from using the results of the study with respects to designing training sessions or tactics to give information about what style of play and the requirements of the attacking play are best in order to produce success. In this sense, game tactics based on direct styles of passing and entering the attacking 1/3 as quickly as possible seem to have more chance of success in the modern game, however future research should consider the types of passes being played from these key areas and type/location of resulting shots which could help identify whether any are more effective.
Bate, R. (1988) Science and Football: Proceedings of the First World Congress of Science and Football [online]. Liverpool: Routledge. [Accessed 08 October 2014].
Carling, C., Williams, A. and Reilly, T. (2005) Handbook of Soccer Match Analysis: a Systematic Approach to Improving Performance. New York, NY: Routledge.
Casamichana, D. and Castellano, J. (2013) Differences between winning, drawing and losing teams in the 2010 World Cup. In Science and Football VII: The Proceedings of the Seventh World Congress on Science and Football, p. 211. New York, NY: Routledge
Clemente, F., Couceiro, M., Martins, F.M.L. and Mendes, R. (2012) Team’s Performance on FIFA U17 World Cup 2011: Study based on Notational Analysis. Journal of Physical Education and Sport [online]. 12 (1), pp. 13-17. [Accessed 27 January 2015].
Collet, C. (2012) The possession game? A comparative analysis of ball retention and team success in European and international football, 2007–2010. Journal of Sports Sciences [online]., pp. 1-14. [Accessed 08 October 2014].
Coolidge, F.L. (2006) Statistics: A gentle introduction. 2nd ed. Thousand Oaks, CA: Sage.
Field, A. (2009) Discovering Statistics Using Spss. 3rd ed. London, UK: Sage.
Franks, I.M. and Miller, G. (1986) Eyewitness testimony in sport. Journal of Sport Behaviour [online]. 9 (1), pp. 38-45. [Accessed 26 January 2015].
Franks, I.M. (1997). Use of feedback by coaches and players. In: Reilly, T., Bangsbo, J. and Hughes, M. eds., Science and football III London: E. and F. Spon.
Hinton, P.R. (2004) Statistics Explained. 2nd ed. New York, NY: Routledge.
Hughes, M. and Bartlett, R. (2002) The use of performance indicators in performance analysis. Journal of Sports Sciences [online]. 20 (1), pp. 739-754. [Accessed 27 January 2015].
Hughes, M., Evans, S. and Wells, J. (2001) Establishing normative profiles in performance analysis. International Journal of Performance Analysis in Sport[online]. 1 (1), pp. 1-26. [Accessed 03 March 2015].
Hughes, M. and Franks, I. (2005) Analysis of passing sequences, shots and goals in soccer. Journal of Sports Sciences [online]. 23 (5), pp. 509-514. [Accessed 08 October 2014].
James, N. (2006) The role of notational analysis in soccer coaching. International Journal of Sports Science and Coaching [online]. 1 (2), pp. 185-198. [Accessed 08 October 2014].
Jinshan, X., Xiaoke, C., Yamanaka, K., and Matsumoto, M. (1993). Analysis of the goals in the 14th World Cup. In: Science and football II. London: E. and F. Spon. pp.203-205.
Jones, P., James, N. and Mellalieu, S. (2004) Possession as a performance indicator in soccer. International Journal of Performance Analysis in Sport [online]. 4 (1), pp. 98-102. [Accessed 08 February 2015].
Konstadinidou, X. and Tsigilis, N. (2005) Offensive playing profiles of football teams from the 1999 women’s world cup finals. International Journal of Performance Analysis in Sport [online]. 5 (1), pp. 61-67. [Accessed 10 February 2015].
Lago, C. (2009) The influence of match location, quality of opposition, and match status on possession strategies in professional association football. Journal of Sports Sciences [online]. 27 (13), pp. 1463-1469. [Accessed 08 February 2015].
Lago, C. and Martin, R. (2007) Determinants of possession of the ball in soccer. Journal of Sports Sciences [online]. 25 (9), pp. 969-974. [Accessed 08 October 2014].
Lago-Ballesteros, J. and Lago-Peñas, C. (2010) Performance in team sports: identifying the keys to success in soccer. Journal of Human Kinetics [online]. 25, pp. 85-91. [Accessed 08 February 2015].
Lago-Peñas, C. and Dellal, A. (2010) Ball possession strategies in elite soccer according to the evolution of the match-score: the influence of situational variables. Journal of Human Kinetics [online]. 25, pp. 93-100. [Accessed 08 October 2014].
Lancaster, G.A., Dodd, S. and Williamson, P.R. (2004) Design and analysis of pilot studies: recommendations for good practice. Journal of Evaluation in Clinical Practice [online]. 10 (2), pp. 307-312. [Accessed 05 February 2015].
Larsen, O., Zoglowek, H. and Rafoss, K. (2000) An Analysis of Team Performance For the Norwegian Women Soccer Team in the Olympics in Atlanta 1996. In: Hughes, M. and Franks, I. eds., Notational analysis of sport [online]. 2nd ed. London: Taylor and Francis. [Accessed 03 March 2015].
O’Donoghue, P. (2005) Normative profiles of sports performance. International Journal of Performance Analysis in Sport [online]. 5 (1), pp. 104-119. [Accessed 03 March 2015].
Olsen, E. and Larsen, O. (1997). Use of match analysis by coaches. In: Reilly, T., Bangsbo, J. and Hughes, M. eds., Science and football III London: E. and F. Spon. pp. 209-220.
Opta Stats (2013) Opta’s Event Definitions. Available from: http://www.optasports.com/news-area/blog-optas-event-definitions.aspx [Accessed 03 February 2015].
Pratas, J., Volossovitch, A. and Ferreira, A.P. (2012) The effect of situational variables on teams’ performance in offensive sequences ending in a shot on goal: a case study. The Open Sports Sciences Journal [online]. 5 (1), pp. 193-199. [Accessed 08 February 2015].
Pollard, R., Reep, C. and Hartley, S. (1988) The Quantative Comparison of playing styles in soccer [online]. New York, NY: Routledge. [Accessed 08 October 2014].
Reep, C. and Benjamin, B. (1968) Skill and chance in association football. Journal of the Royal Statistical Society [online]. 131 (4), pp. 581-585. [Accessed 08 October 2014].
Yiannakos, A. and Armatas, V. (2006) Evaluation of the goal scoring patterns in European championship in Portugal 2004. International Journal of Performance Analysis in Sport [online]. 6 (1), pp. 178-188. [Accessed 08 February 2015].