Rugby Union is a high-intensity intermittent sport that involves numerous phases of open play, separated by contact situations. The initial phase of play typically starts from a set-piece (e.g. scrum, lineout or kick-off), where the defensive team arranges itself in a horizontal line called the defensive line. The arrangement of the defensive line and the movement of this line during play is often in response to the movement and alignment of the attacking line (Hendricks et al., 2013). The ability of the attacking team to break through the defensive line is a key indicator of success as it creates opportunities to score tries (Bracewell, 2003; Wheeler et al., 2010; Diedrick and van Rooyen, 2011).

A study by Wheeler et al. (2010) identified three effective attacking strategies to break through the attacking line - tackle breaks, offloads and line breaks. A coordinated set of skills need to be executed effectively to achieve these attacking strategies. The set of skills associated with tackle breaks and offloads have been documented (Wheeler and Sayers, 2009; Wheeler et al., 2010). Similarly, effective defensive skills and tackle techniques have also been identified (Hendricks et al., 2014). However, not much is known about the skills associated with line breaks.

In an analysis of 58 Six Nations Rugby games between 2003 and 2007, Ortega et al. (2009) observed that successful teams break the line more frequently than unsuccessful teams. This suggests that line breaks are a key performance indicator of successful teams. Logically, the next step would be to describe the events, technical skills and situation in which line breaks occur. This will assist coaches in designing attacking drills to improve player’s skills and increase their chances of breaking the line in matches. Analysing the technical skill requirements for a line break, in addition to the events leading up to the line break and the situation in which the line break occurs provides a better understanding of the conditions for successful match performances (Mackenzie and Cushion 2013). Therefore, the primary aim of this study was to analyse the technical skills and match situations that led to line breaks in Rugby Union. A secondary aim was to determine which technical skills and situations influenced the outcome of the line break.

All matches (n = 125) of the 2013 Super Rugby competition were analysed. The 2013 Super Rugby competition was a professional men’s Rugby Union competition played between 15 franchises from Australia, New Zealand and South Africa. The footage was obtained from publically broadcasted matches that were recorded and saved onto a database. The study was ethically approved by the University of Cape Town Faculty of Health Sciences Human Research Ethics Committee (HREC REF: 559/2014). The recorded matches were analysed using Sports Code Elite Version 6.5.1 on an Apple iMac (Apple, USA) displayed at eye level. The analysis software allowed for the control of time lapse during the recorded match and the recording and saving of each event into a database. The analyst was able to pause, rewind and watch the footage in slow motion. The highest frequency at which the analyst could slow down the footage was 25 frames per second (25 Hz).

Line breaks were identified using the following criteria. Note, all of the criteria described below had to be fulfilled for an event to be classified as a line break.

The player in possession of the ball (i.e., ball-carrier) needed to visibly move through a straight line between two defenders in the defensive line or between a defender and the touchline, regardless of the shape and a defender and the touchline, regardless of the shape andmovement of the defensive line, and without being physically contacted by a defender (Gilmore, 2006; Wheeler et al., 2010; Hendricks et al., 2013).

The defensive line needed to consist of at least two players moving in unison. After an attacker penetrated the defensive line, at least one defender from the defensive line was required to turn toward the ball-carrier, in an attempt to chase the ball-carrier. This criterion allowed for the distinction between half breaks and complete line breaks (Burt et al., 2013).

The ball-carrier needed to receive the ball from a teammate before he breached the line. This criterion excluded interceptions and offloads from the analyses. When the starting phase was a turnover or loose ball, the line break could not occur in that phase of play, as it was deemed the defensive line would not have sufficient time to form.

A total of 362 line breaks were identified, with an average of 3 (SD = 2) line breaks per match. Thereafter, each line break was coded for match, situational and technical descriptors that preceded the line break. The analysis of the line break began at the first set-piece, or change in possession, preceding the line break and ended if possession of the ball was lost, a break in open play after the line break occurred, when the ball went into touch, an infringement occurred or if a try was scored.

The purpose of the descriptors was to adequately describe the technical skills which led to a line break. The descriptors needed to possess objective metric properties and a valid means of interpretation (O’Donoghue, 2010). The list of descriptors used to code the line breaks was created through the use of previous peer-reviewed published literature (Hendricks et al., 2014; Wheeler et al., 2010) and by identifying and describing characteristics required specifically for this study. The list was then reviewed by a panel of coaches, sports scientists and rugby administrators to assess the validity and relevance of the performance indicators, and the clarity of their definitions.

In line with the purpose of the study, the variables were divided into five categories: match situation, pre-line break characteristics, final phase characteristics, skills and post line break outcomes (Table 1 - Table 2 - Table 3). The match situation descriptors described factors related to the teams playing in the match. The pre-line break characteristics described the events that occurred in the phases leading up to the line break, and final phase characteristics, the events in the phase that the line break occurred. Post line break outcomes described the results of the line break, in terms of possession and tries scored. The list of descriptors selected for this study was inclusive to account for any factor which may be relevant. However, for the purpose of this paper, only the variables which were found to be influential will be discussed.

For intra-coder reliability, five randomly selected matches were coded twice using the variables and definitions described above. The coding of the same match was separated by a week (Wheeler et al., 2010). Kappa statistics (κ) were used to test the intra-reliability of the coder for each group of variables, each category and the overall agreement between all the variables between each set of matches combined. The overall agreement between the two sets of matches was κ = 0.81. Kappa values between 0.81 and 0.99 represent an almost perfect agreement between repeated measures (O’Donoghue, 2010).

The purpose of the statistical analyses was to determine the skills and match situations associated with line breaks. To achieve this, the first level of analyses was descriptive. The purpose of this was to describe and compare the frequency of occurrence. The second level of analyses tested the association between the skills and match situations variables that led to line-breaks using the Pearson’s Chi-squared test and Cramer’s V test. A Pearson’s Chi-squared P value of less than 0.05 was regarded as a significant association between the variables (Brace et al. 2003). Cramer’s V is a measure of the strength of the association between two variables, giving a value of 0 to 1. A Cramer’s V value that was less than 0.20 was observed as having a weak association, between 0.20 and 0.49 as a moderate association, and values above 0.49 as a strong association (Acock, 2012).

The purpose of the final analyses was to determine the likelihood of skills and situational variables on post line break outcomes. Relative risk ratio (RRR), the ratio of the probability of an outcome occurring when a characteristic was observed (versus the non-observed characteristic) was determined using multinomial logistic regression. All statistics were done using STATA 11.1.

The top five teams in the log, group 1, made 39% of the line breaks over the course of the season, with 17% of those line breaks against the bottom five teams (Figure 1).

Thirty-nine percent of line breaks resulted in a try being scored within two phases of the line break. Furthermore, possession was retained in 40% of line breaks where no try was scored.

The most common types of passes received by the line breaker were a short lateral pass (20%), long lateral pass (16%), or a skip pass (17%), and one-third of all line breaks were made when there were 2 passes in the final phase. The highest percentage of line breaks occurred when the ball-carrier ran straight (48%) or diagonal (44%), and at a fast speed (54%). Sixty-three percent of line breaks occurred when there was no change in running angle by the ball-carrier, no deception (70%) and no evasive movements by the ball-carrier (87%). The field position where the line breaks occurred is represented in Figure 2.

The attacking line shape with the highest frequency was deep and wide (30%). When the attacking line was deep and wide, 33% of line breaks occurred when 2 passes were made in the final phase [x² (42) = 71.875, p = 0.003, Cramer’s v = 0.2080, moderate].

The majority of line breaks occurred when the ratio of attackers to defenders was an overlap of attackers, one man overlap (34%) and multiple overlap (30%). Line breaks occurred more frequently when the defensive line was distant (65%), and when players in the defensive line shifted laterally (25%) at a moderate running speed (50%).

The distance of defence was significantly associated with the number of passes, type of pass, running speed, and running angle of the ball carrier. Pick and go’s (25%), 1 pass (25%) or 2 passes (36%) in the final phase resulted in line breaks when the defensive distance was close at ball reception. Similarly, one out of three line breaks occurred when 2 passes were made when the defence was distant [x²(12) = 30.348, p = 0.002, Cramer’s v = 0.2085, moderate].

Passing the ball laterally (short or long) resulted in 42% of line breaks when the defensive line was distant at ball reception. At a moderate distance, short lateral passes (19%) and skip passes (15%) attributed to the highest percentage of line breaks [x²(26) = 50.899, p = 0.036, Cramer’s v = 0.2484, moderate].

Line breaks occurred when the ball-carrier was running fast [61%, x²(4) = 25.784, p = 0.000, Cramer’s v = 0.1922, weak], diagonal [50%, x²(6) = 15.463, p = 0.017, Cramer’s v = 0.1488, weak] and straightening [31%, x²(8) = 17.765, p = 0.023, Cramer’s v = 0.1595, weak] at a defensive line that was distant.

Faster defensive line speeds resulted in more line breaks [x²(12) = 61.703, p < 0.001, Cramer’s v = 0.3026, moderate], especially when the ball-carrier was running fast (63%). Thirty-three percent of line breaks occurred when the speed of the defensive line was fast, and 35% when the speed was moderate, and two passes were made [x²(12) = 61.703, p < 0.001, Cramer’s v=0.3026, moderate].

The use of deception was found to be positively associated with the speed of defence [x²(10) = 21.814, p = 0.016, Cramer’s v = 0.1802, weak]. When the speed of the defensive line was fast, 19% of line breaks occurred using a decoy runner. When the speed of the defensive line was slow, 23% of line breaks occurred using a dummy pass. When the defensive line shifted laterally, 31% of line breaks occurred using 2 passes [x²(42) = 99.948, p = 0.000, Cramer’s v = 0.2271, moderate].

Line breaks were achieved most frequently by outside backs (60%) and inside backs (27%). Inside backs broke the line through pick and go’s (29%) or long lateral passes (24%), and outside backs broke the line upon receiving a skip pass (21%) or a short lateral pass (20%) [x²(39) = 86.502, p = 0.000, Cramer’s v = 0.2822, moderate]. Further, backs are accelerating [x²(6) = 14.877, p = 0.021, Cramer’s v = 0.1433, weak] and running faster when receiving the ball [x²(6) = 17.753, p = 0.007, Cramer’s v = 0.1566, weak] and attacking the defensive line diagonally (57%, inside backs) or straight (52%, outside backs) when breaking the line [x²(9) = 29.204, p = 0.001, Cramer’s v = 0.1640, weak]. For inside backs, when the inside defender was a tight forward line breaks occurred more frequently [x²(12) = 78.329, p < 0.001, Cramer’s v = 0.2686, moderate].

The likelihood of a line break resulting in a try was significantly increased when the final pass was a flat pass (RRR 29.78, 95% CI 1.39-635.60, p < 0.05) compared to no pass being made. A central initial field position significantly reduced the chance of scoring a try, relative to the right side of the field (RRR 0.43, 95% CI 0.19-0.98, p < 0.05). Field position zones 2 (RRR 0.05, 95% CI 0.09-0.29, p < 0.05) and 3 (RRR 0.04, 95% CI 0.04-0.34, p < 0.05) significantly increased the likelihood of retaining possession relative to zone 1. Further, line breaks that occurred in zone 3 were 94% more likely to result in the attacking team retaining possession post line break (RRR 0.06, 95% CI 0.01-0.58, p < 0.05).

This study showed that line breaks were associated with overall success in the Super Rugby competition and scoring tries. The technical attacking skills identified for line breaks were wide and deep positioning, receiving the ball at speed within 3 passes from the breakdown, and using passes that promote quick facilitation of the ball to outside backs. Furthermore, more line breaks were observed when these attacking technical skills were executed when the defensive line was further away and approaching quickly. In which case, the addition of a decoy or dummy runner proved useful to breach the defensive line untouched. Also, the skills associated with line breaks differed between playing positions.

The top five teams in the competition (based on final log position) created the most line breaks, compared to the middle five or bottom five teams. In contrast, Wheeler et al. (2010) reported little variation between a team’s overall success and line breaks. Note, there was a large difference between the sample size of the two studies, Wheeler et al. (2010) analysed seven matches compared to 125 in this study. Line breaks were also associated with tries scored in matches. Throughout the season, there was an average of 3 line breaks per match, with 39% of line breaks resulting in a try. This averages to one try per match resulting from a line break. This finding is in accordance with previous studies describing the effectiveness of line breaks in promoting try scoring opportunities in Rugby Union (Diedrick and van Rooyen, 2011; Wheeler et al., 2010). In 66% of the cases where a line break did not result in a try, possession was retained by the attacking team in the subsequent phase. Therefore, from an attacking territorial point of view, line breaks are also useful in moving the ball closer to the opposition’s try-line while still maintaining possession.

Ball-carrier speed as a key technical requirement for line breaks is not surprising. In an emerging environment like rugby, defenders (or a defender and the touchline) will inevitably close a wide gap in the defensive line to reduce opportunities to break the line. However, if the ball-carrier is able to identify the gap soon enough and accelerate and sprint through the gap faster than the defenders are able to close it, the ball-carrier will succeed in breaking the line (Correia et al. 2011). What this study adds though is that in addition to ball-carrier speed, movement speed of the ball between attackers creates line break opportunities. Quick facilitation of the ball to outside backs using either short, long or skip passes creates a high tempo attack which may afford defenders less time to organise themselves which results in more gaps between defenders. However, if defenders have organised themselves, adapting the tactics to add a decoy runner or dummy runner based on the speed and distance of the defence promotes line breaks. From a coaching perspective, coaches should emphasise the movement speed of the ball between attackers when training passing, and manipulate the speed and distance of the defenders.

The closer the initial, final phase or line break field position was to the try line, the more likely the team was to lose possession of the ball in the phase following the line break. Diedrick and van Rooyen (2011) reported little influence of the vertical field position on the post line break outcome. However, Diedrick and van Rooyen (2011) did not specify if contact was an exclusion criterion when identifying line breaks and the study included interceptions. Also, tackle breaks and offloads may influence the number of support players around the ball-carrier post line break increasing the likelihood of a team maintaining possession.

A limitation of the study was the absence of comparison phase outcome for line breaks. The lack of a comparator meant more sophisticated probability statistics could not be applied. Seemingly, tackle contact may be an appropriate comparator. However, all tackle contact events are not necessarily an unsuccessful line break, therefore tackle contact would have been invalid. Furthermore, the main purpose of this study was to describe non-contact skills that created line breaks, since effective contact have been described before (Wheeler & Sayers 2009; Wheeler et al. 2010). Lastly, the footage used in this study was obtained from publically broadcasted matches, which meant that the view of the match during analysis was restricted by the camera view. This was particularly apparent with the analysis of the attack-defence configuration.

Line breaks are associated with overall team success and try scoring opportunities. Awareness of the defenders line speed and depth, fast running speed when receiving the ball and quick passing between attackers to the outside backs creates line break opportunities. During training, coaches should emphasise the movement speed of the ball between attackers and manipulate the speed and distance of the defenders.