Collision count in rugby union: A comparison of micro-technology and video analysis methods

Assessment of contact involvements and scrums in international rugby union match-play using video analysis and microsensor technology methods

This study sought to assess the validity of contact involvement (CI) detection using microsensor technology (MST, Catapult Vector) within the context of a Tier One national rugby union (RU) squad, consisting of 44 players. Sensitivity of MST units to detect CI and scrums was assessed in eight test matches, by comparison with match data obtained by video analysis. This paper is the first to assess the sensitivity of MST to the full range of skilled CI which occur in RU, including evaluating "non-performance" collisions, such as incidental collisions or foul play. Sensitivity to tackles made (52.9-84.9%) and ruck hits (53.3-87.2%) was lower than previous research, although ball carries (71.9-93.5%) showed broadly similar sensitivity to established results. The sensitivity of the MST to detect scrums was substantially lower than previous findings, with large positional variation evident (51.4-91.5%). Further refinement of MST software should be considered in order to facilitate valid monitoring of RU performance and injury risk. An additional finding was that video analysis generally demonstrated satisfactory intrarater reliability. This result supports the use of video analysis as a reliable method of assessing RU performance, including CI.

The Relationship Between Subjective Wellness and External Training Load in Elite English Premier League Goalkeepers and a Comparison With Outfield Soccer Players

PURPOSE

To investigate the relationship between training load and subjective wellness in English Premier League goalkeepers (GKs) and examine potential positional differences in subjective wellness.

METHODS

A total of 34 players (GK = 7, outfield = 27) completed a daily subjective wellness questionnaire assessing sleep quality, sleep hours, fatigue, mood, soreness, and total wellness over two and a half seasons. Ten-Hertz GPS devices were worn during training to calculate previous-day and 7-day total distance, player load, total dives, total dive load, average time to feet, and high, medium, and low jumps.

RESULTS

All previous 7-day training loads were associated with all wellness markers (r = .073 to .278, P < .05). However, associations between previous 7-day dive load and mood, average time to feet, and both sleep quality and quantity, and between low jumps and sleep quality, were not significant. For previous-day metrics, total distance was associated with all wellness markers (r = .097 to .165, P < .05). In addition, player load and high jump were associated with fatigue, soreness, and wellness (r = .096 to .189, P < .05). Total dives and soreness were also related (r = .098, P < .05), and relationships were evident between average time to feet, medium jumps, and all wellness markers excluding sleep quality (r = .114 to .185, P < .05). No positional differences in subjective wellness occurred (P > .05).

CONCLUSION

Some GK GPS variables are associated with subjective wellness, which could inform training-load prescription to maximize recovery and performance. In addition, GKs are no more vulnerable to poorer subjective wellness when compared with outfield players.

Prototype Machine Learning Algorithms from Wearable Technology to Detect Tennis Stroke and Movement Actions

This study evaluated the accuracy of tennis-specific stroke and movement event detection algorithms from a cervically mounted wearable sensor containing a triaxial accelerometer, gyroscope and magnetometer. Stroke and movement data from up to eight high-performance tennis players were captured in match-play and movement drills. Prototype algorithms classified stroke (i.e., forehand, backhand, serve) and movement (i.e., "Alert", "Dynamic", "Running", "Low Intensity") events. Manual coding evaluated stroke actions in three classes (i.e., forehand, backhand and serve), with additional descriptors of spin (e.g., slice). Movement data was classified according to the specific locomotion performed (e.g., lateral shuffling). The algorithm output for strokes were analysed against manual coding via absolute (n) and relative (%) error rates. Coded movements were grouped according to their frequency within the algorithm's four movement classifications. Highest stroke accuracy was evident for serves (98%), followed by groundstrokes (94%). Backhand slice events showed 74% accuracy, while volleys remained mostly undetected (41-44%). Tennis-specific footwork patterns were predominantly grouped as "Dynamic" (63% of total events), alongside successful linear "Running" classifications (74% of running events). Concurrent stroke and movement data from wearable sensors allows detailed and long-term monitoring of tennis training for coaches and players. Improvements in movement classification sensitivity using tennis-specific language appear warranted.

An Analysis of Repeated High Intensity Efforts (RHIE) across Different Competition Levels in Elite Rugby Union

The current investigation aimed to understand the differing positional demands across two elite rugby union competitions, with special reference to high-intensity effort (HIE) and repeated high-intensity effort (RHIE) activity. Four hundred and forty-one (n = 441) individual game files from thirty-five competitive games from the European Rugby Champions Cup (tier 1; n = 8) and PRO12 League (tier 2; n = 24) were analysed. Players' locomotor profiles were recorded using wearable global positioning system microtechnology (10 Hz Catapult S5, Catapult Innovations, Australia). Locomotor activities were classified as running (≥4.4 m∙s-1), high-speed running (≥5.5 m∙s-1), accelerations (≥2 m∙s-2) and decelerations (≤-2 m∙s-2). Data was gathered on collisions (≥4 g-1), high-intensity efforts (HIE), repeated high-intensity efforts (RHIE), average number of efforts within a RHIE bout (n) and maximal number of efforts within a RHIE bout (n). Overall locomotor differences between competitions were trivial to small in nature, with tier 1 competition associated with a larger number of RHIE bouts (6.5 ± 1.4 vs. 5.7 ± 1.5, effect size, ES = 0.55) and efforts per bout (3.0 ± 1.1 vs. 2.4 ± 1.2, ES = 0.52). Collisions comprised a greater proportion of total HIE for forwards within tier 1 competition compared to tier 2 competition. The hooker (mean difference: 4 [-10 to 14]; ES = 0.30, small), lock (mean difference: 5 [-12 to 23]; ES = 0.36, small) and backrow (mean difference: 8 [-10 to 15]; ES = 0.54, small) positions engaged in more collisions during tier 1 competition compared to tier 2 competition. These findings can be used by athletic performance staff to design game-specific drills and recovery strategies during different competition weeks to ensure players are appropriately prepared for the differing demands of elite rugby competition.

Validating an algorithm from a trunk-mounted wearable sensor for detecting stroke events in tennis

This study analysed the accuracy of a prototype algorithm for tennis stroke detection from wearable technology. Strokes from junior-elite tennis players over 10 matches were analysed. Players wore a GPS unit containing an accelerometer, gyroscope and magnetometer. Manufacturer-developed algorithms determined stoke type and count (forehands, backhands, serves and other). Matches were video recorded to manually code ball contacts and shadow swing events for forehands, backhands and serves and further by stroke classifications (i.e., drive, volley, slice, end-range). Comparisons between algorithm and coding were analysed via ANOVA and Bland-Altman plots at the match-level and error rates for specific stroke-types. No significant differences existed for stroke count between the algorithm and manual coding (p > 0.05). Significant (p < 0.0001) overestimation of "Other" strokes were observed from the algorithm, with no difference in groundstrokes and serves (p > 0.05). Serves had the highest accuracy of all stroke types (≥98%). Forehand and backhand "drives" were the most accurate (>86%), with volleys mostly undetected (58-60%) and slices and end-range strokes likely misclassified (49-51%). The prototype algorithm accurately quantifies serves and forehand and backhand "drives" and serves. However, underestimations of shadow swings and overestimations of "other" strokes suggests strokes with reduced trunk rotation have poorer detection accuracy.

A game for all shapes and sizes? Changes in anthropometric and performance measures of elite professional rugby union players 1999–2018

Background

Rugby union player size has increased since the game turned professional in 1995. Changes in physical and performance capability over this period have yet to be fully described.

Hypothesis

Increases in player momentum would result from changes in body mass.

Methods

Within-player rates of change in anthropometric and kinetic variables with season played were sampled in three successively studied professional rugby union club cohorts playing at the highest level of European competition between 1999 and 2019. Data comprised 910 seasons of observation for 291 elite male players. Most players had 2, 3 or 4 seasons of observation. Mixed-effects modelling distinguished changes independent of position played, club and international status.

Results

With each season played, player body mass, fat-free mass and maximum speed increased significantly, while per cent fat decreased. The mean maximal velocity of a rugby player in 1999 was 8.2 (±0.18) m/s, which in 2019 had risen to 9.1 (±0.10) m/s. Player’s momentum in 2019 was 14% more than those playing in 1999. In the Front Five, momentum increased in this period by >25%, mainly driven by greater running speed, disproving our hypothesis.

Conclusions

The momentum of players, particularly forwards, increased markedly over 20 seasons of professional rugby. The resulting forces generated in collisions are thus significantly greater, although these may be mitigated by better player conditioning. Proactive regulation to address player safety may be required to address the changing nature of anthropometric measures and physical performance, minimising injury rates and potential long-term sequelae.

Quantifying Collision Frequency and Intensity in Rugby Union and Rugby Sevens: A Systematic Review

BACKGROUND

Collisions in rugby union and sevens have a high injury incidence and burden, and are also associated with player and team performance. Understanding the frequency and intensity of these collisions is therefore important for coaches and practitioners to adequately prepare players for competition. The aim of this review is to synthesise the current literature to provide a summary of the collision frequencies and intensities for rugby union and rugby sevens based on video-based analysis and microtechnology.

METHODS

A systematic search using key words was done on four different databases from 1 January 1990 to 1 September 2021 (PubMed, Scopus, SPORTDiscus and Web of Science).

RESULTS

Seventy-three studies were included in the final review, with fifty-eight studies focusing on rugby union, while fifteen studies explored rugby sevens. Of the included studies, four focused on training-three in rugby union and one in sevens, two focused on both training and match-play in rugby union and one in rugby sevens, while the remaining sixty-six studies explored collisions from match-play. The studies included, provincial, national, international, professional, experienced, novice and collegiate players. Most of the studies used video-based analysis (n = 37) to quantify collisions. In rugby union, on average a total of 22.0 (19.0-25.0) scrums, 116.2 (62.7-169.7) rucks, and 156.1 (121.2-191.0) tackles occur per match. In sevens, on average 1.8 (1.7-2.0) scrums, 4.8 (0-11.8) rucks and 14.1 (0-32.8) tackles occur per match.

CONCLUSIONS

This review showed more studies quantified collisions in matches compared to training. To ensure athletes are adequately prepared for match collision loads, training should be prescribed to meet the match demands. Per minute, rugby sevens players perform more tackles and ball carries into contact than rugby union players and forwards experienced more impacts and tackles than backs. Forwards also perform more very heavy impacts and severe impacts than backs in rugby union. To improve the relationship between matches and training, integrating both video-based analysis and microtechnology is recommended. The frequency and intensity of collisions in training and matches may lead to adaptations for a "collision-fit" player and lend itself to general training principles such as periodisation for optimum collision adaptation. Trial Registration PROSPERO registration number: CRD42020191112.

The Use of Global Positioning and Accelerometer Systems in Age-Grade and Senior Rugby Union: A Systematic Review

Background

Global positioning systems (GPS) imbedded with accelerometer systems (AS) are used in rugby union (RU) to collect information on absolute and relative distances, distances in different speed zones, high-speed running (HSR) distances, repeated high-intensity efforts (RHIE) and collisions and impacts. This information can be used to monitor match play which can then be used to plan training sessions. The objective of this review was to conduct a systematic review of studies which have reported the use of GPS and AS.

Methods

A systematic review of the use of GPS and AS in both age-grade and senior rugby was conducted. The authors systematically searched electronic databases from January 2010 until March 2020. Keywords included rugby union, GPS, global position* and microtechnology.

Results

A total of 51 studies met the eligibility criteria and were included in this review. There was a total of 34 studies utilising GPS and AS in senior RU players (mean ± SD; age 26.2 ± 1.9 years; height 185.7 ± 2.6 cm; mass 101.3 ± 4.2 kg) and 17 studies in age-grade RU players (mean ± SD; age 17.6 ± 1.5 years; height 182.1 ± 3.3 cm; mass 87.1 ± 8.6 kg). The results of this review highlighted that there are differences between backs and forwards and within these positions in these groups during both match play and training sessions. The backs covered greater total absolute, relative and HSR distance compared to forwards. Forwards are involved in more collisions and impacts than backs. When investigating the most intense periods of match play, studies in this review highlighted that the demands during these periods outweigh the average demands of the game. It was proposed that a rolling average over different time epochs is the best way to assess this and ensure that the most intense periods of play are assessed and monitored.

Conclusions

The information highlighted in this review can be used to help coaches assess performances in match play, allow them to plan appropriate training sessions and monitor training load.

Wales Anaerobic Test: Reliability and Fitness Profiles of International Rugby Union Players

Beard, A, Ashby, J, Chambers, R, Millet, GP, and Brocherie, F. Wales Anaerobic Test (WAT): Reliability and fitness profiles of international rugby union players. J Strength Cond Res XX(X): 000-000, 2019-To provide strength and conditioning coaches a practical and evidence-based test for repeated-sprint ability (RSA) in rugby union players, this study assessed the relative and absolute test-retest reliability of the Wales Anaerobic Test (WAT) and its position-specific association with other fitness performance indices. Thirty-four players (forwards: n = 19; backs: n = 15) of the Welsh rugby union male senior national team performed the WAT (10 × 50-m distance, 25-30 seconds of passive recovery) twice within 4 days. Time for each repetition was recorded, with the best (WATBest) and total time (WATTT) retained for analysis. Relative (intraclass correlation coefficient [ICC]) and absolute (SEM) reliability of the WAT indices were quantified. Furthermore, association (Pearson's product-moment correlations and stepwise backward elimination procedure) with other fitness performance indices (10- and 40-m sprinting times, 30-15 intermittent fitness test [30-15IFT] and the Yo-Yo intermittent recovery test level 2 [YYIR2]) was investigated. Pooled values revealed "moderate" to "high" ICCs for WATBest (ICC = 0.89, p = 0.626) and WATTT (ICC = 0.95, p = 0.342). Good test sensitivity was reported for forwards and backs' WATTT (p > 0.101). Both WATBest and WATTT correlated with 10-m and 40-m sprinting times (r > 0.69, p < 0.001) as well as with 30-15IFT (r < -0.77, p < 0.001) and YYIR2 (r < -0.68, p < 0.001) for pooled values. The WAT proved to be a reliable and sensitive test to assess the rugby union specific RSA-related fitness of international players.

Validity of a Microsensor-Based Algorithm for Detecting Scrum Events in Rugby Union

PURPOSE

Commercially available microtechnology devices containing accelerometers, gyroscopes, magnetometers, and global positioning technology have been widely used to quantify the demands of rugby union. This study investigated whether data derived from wearable microsensors can be used to develop an algorithm that automatically detects scrum events in rugby union training and match play.

METHODS

Data were collected from 30 elite rugby players wearing a Catapult OptimEye S5 (Catapult Sports, Melbourne, Australia) microtechnology device during a series of competitive matches (n = 46) and training sessions (n = 51). A total of 97 files were required to "train" an algorithm to automatically detect scrum events using random forest machine learning. A further 310 files from training (n = 167) and match-play (n = 143) sessions were used to validate the algorithm's performance.

RESULTS

Across all positions (front row, second row, and back row), the algorithm demonstrated good sensitivity (91%) and specificity (91%) for training and match-play events when the confidence level of the random forest was set to 50%. Generally, the algorithm had better accuracy for match-play events (93.6%) than for training events (87.6%).

CONCLUSIONS

The scrum algorithm was able to accurately detect scrum events for front-row, second-row, and back-row positions. However, for optimal results, practitioners are advised to use the recommended confidence level for each position to limit false positives. Scrum algorithm detection was better with scrums involving ≥5 players and is therefore unlikely to be suitable for scrums involving 3 players (eg, rugby sevens). Additional contact- and collision-detection algorithms are required to fully quantify rugby union demands.

Automatic detection of one-on-one tackles and ruck events using microtechnology in rugby union

OBJECTIVES

To automate the detection of ruck and tackle events in rugby union using a specifically-designed algorithm based on microsensor data.

DESIGN

Cross-sectional study.

METHODS

Elite rugby union players wore microtechnology devices (Catapult, S5) during match-play. Ruck (n=125) and tackle (n=125) event data was synchronised with video footage compiled from international rugby union match-play ruck and tackle events. A specifically-designed algorithm to detect ruck and tackle events was developed using a random forest classification model. This algorithm was then validated using 8 additional international match-play datasets and video footage, with each ruck and tackle manually coded and verified if the event was correctly identified by the algorithm.

RESULTS

The classification algorithm's results indicated that all rucks and tackles were correctly identified during match-play when 79.4±9.2% and 81.0±9.3% of the random forest decision trees agreed with the video-based determination of these events. Sub-group analyses of backs and forwards yielded similar optimal confidence percentages of 79.7% and 79.1% respectively for rucks. Sub-analysis revealed backs (85.3±7.2%) produced a higher algorithm cut-off for tackles than forwards (77.7±12.2%).

CONCLUSIONS

The specifically-designed algorithm was able to detect rucks and tackles for all positions involved. For optimal results, it is recommended that practitioners use the recommended cut-off (80%) to limit false positives for match-play and training. Although this algorithm provides an improved insight into the number and type of collisions in which rugby players engage, this algorithm does not provide impact forces of these events.

PlayerLoad Variables: Sensitive to Changes in Direction and Not Related to Collision Workloads in Rugby League Match Play

PURPOSE

To determine (1) how change-of-direction (COD) workloads influence PlayerLoad (PL) variables when controlling total distance covered and (2) relationships among collision workloads and PL variables during rugby league match play.

METHODS

Participants completed 3 protocols (crossover design) consisting of 10 repetitions of a 60-m effort in 15 s. The difference between protocols was the COD demands required to complete 1 repetition: no COD (straight line), 1° × 180° COD, or 3° × 180° COD. During rugby league matches, relationships among collision workloads, triaxial vector-magnitude PlayerLoad (PL_VM), anteroposterior + mediolateral PL (PL_2D), and PL_VM accumulated at locomotor velocities below 2 m·s^-1 (ie, PL_SLOW) were examined using Pearson correlations (r) with coefficients of determination (R²).

RESULTS

Comparing 3° × 180° COD to straight-line drills, PL_VM·min^-1 (d = 1.50 ± 0.49, large, likelihood = 100%, almost certainly), PL_2D·min^-1 (d = 1.38 ± 0.53, large, likelihood = 100%, almost certainly), and PL_SLOW·min^-1 (d = 1.69 ± 0.40, large, likelihood = 100%, almost certainly) were greater. Collisions per minute demonstrated a distinct (ie, R² < .50) relationship from PL_VM·min^-1 (R² = .30, r = .55) and PL_2D·min^-1 (R² = .37, r = .61). Total distance per minute demonstrated a very large relationship with PL_VM·min^-1 (R² = .62, r = .79) and PL_2D·min^-1 (R² = .57, r = .76).

CONCLUSIONS

PL variables demonstrate (1) large increases as COD demands intensify, (2) separate relationships from collision workloads, and (3) moderate to very large relationships with total distance during match play. PL variables should be used with caution to measure collision workloads in team sport.

The Use of Microtechnology to Monitor Collision Performance in Professional Rugby Union

Purpose: To determine if microtechnology-derived collision loads discriminate between collision performance and compare the physical and analytical components of collision performance between positional groups. Methods: Thirty-seven professional male rugby union players participated in this study. Collision events from 11 competitive matches were coded using specific tackle and carry classifications based on the ball-carrier’s collision outcome. Collisions were automatically detected using 10 Hz microtechnology units. Collision events were identified, coded (as tackle or carry), and timestamped at the collision contact point using game analysis software. Attacking and defensive performances of 1609 collision events were analyzed. Results: Collision loads were significantly greater during dominant compared with neutral and passive collisions (P < .001; effect size [ES] = 0.53 and 0.80, respectively), tackles (P < .0001; ES = 0.60 and 0.56, respectively), and carries (P < .001; ES = 0.48 and 0.79, respectively). Overall, forwards reported a greater number and frequency of collisions but lower loads per collision and velocities at collision point than did backs. Microtechnology devices can also accurately, sensitively, and specifically identify collision events (93.3%, 93.8%, and 92.8%, respectively). Conclusion: Microtechnology is a valid means of discriminating between tackle and carry performance. Thus, microtechnology-derived collision load data can be utilized to track and monitor collision events in training and games.

The what and how of video analysis research in rugby union: a critical review

BACKGROUND

Video analysis is a common tool used in rugby union research to describe match performance. Studies using video analysis range from broad statistical studies of commercial databases to in-depth case-studies of specific match events. The range of types of studies using video analysis in rugby union, and how different studies apply the methodology, can make it difficult to compare the results of studies and translate the findings to a real-world setting. In attempt to consolidate the information on video analysis in rugby, a critical review of the literature was performed.

MAIN BODY

Ninety-two studies were identified. The studies were categorised based on the outcome of the study and the type of research question, sub-categorised as 'what' and 'how' studies. Each study was reviewed using a number of questions related to the application of video analysis in research. There was a large range in the sample sizes of the studies reviewed, with some of the studies being under-powered. Concerns were raised of the generalisability of some of the samples. One hundred percent of 'how' studies included at least one contextual variables in their analyses, with 86% of 'how' studies including two or more contextual variables. These findings show that the majority of studies describing how events occur in matches attempted to provide context to their findings. The majority of studies (93%) provided practical applications for their findings.

CONCLUSION

The review raised concerns about the usefulness of the some of the findings to coaches and practitioners. To facilitate the transfer and adoption of research findings into practice, the authors recommend that the results of 'what' studies inform the research questions of 'how' studies, and the findings of 'how' studies provide the practical applications for coaches and practitioners.

The worst case scenario: Locomotor and collision demands of the longest periods of gameplay in professional rugby union

A number of studies have used global positioning systems (GPS) to report on positional differences in the physical game demands of rugby union both on an average and singular bout basis. However, the ability of these studies to report quantitative data is limited by a lack of validation of certain aspects of measurement by GPS micro-technology. Furthermore no study has analyzed the positional physical demands of the longest bouts of ball-in-play time in rugby union. The aim of the present study is to compare the demands of the single longest period of ball-in-play, termed “worst case scenario” (WCS) between positional groups, which have previously been reported to have distinguishable game demands. The results of this study indicate that WCS periods follow a similar sporadic pattern as average demands but are played at a far higher pace than previously reported for average game demands with average meters per minute of 116.8 m. The positional differences in running and collision activity previously reported are perpetuated within WCS periods. Backs covered greater total distances than forwards (318 m vs 289 m), carried out more high-speed running (11.1 m·min^-1vs 5.5 m·min^-1) and achieved higher maximum velocities (MaxVel). Outside Backs achieved the highest MaxVel values (6.84 m·sec^-1). Tight Five and Back Row forwards underwent significantly more collisions than Inside Back and Outside Backs (0.73 & 0.89 collisions·min^-1vs 0.28 & 0.41 collisions·min^-1 respectively). The results of the present study provide information on the positional physical requirements of performance in prolonged periods involving multiple high intensity bursts of effort. Although the current state of GPS micro-technology as a measurement tool does not permit reporting of collision intensity or acceleration data, the combined use of video and GPS provides valuable information to the practitioner. This can be used to match and replicate game demands in training.