Technical performance on ATP top level, future level and Swedish youth national level male tennis tournaments : Notational analysis of point characteristics in three different tournaments on three different performance levels

(1)

Technical performance on ATP top

level, future level and Swedish youth

national level male tennis tournaments

- Notational analysis of point characteristics in

three different tournaments on three different

performance levels

Frej Hallgren

GYMNASTIK- OCH IDROTTSHÖGSKOLAN

Master Thesis Sport Science 2016:24

Supervisor: Johnny Nilsson

Examiner: Mikael Mattsson

(2)

Abstract

Aim and research questions

To investigate technical performance in three different tennis competitions (ATP Masters AM, Falu Future, FF & Swedish youth national championships, YNC) by collecting data of point characteristics.

Are there any differences or similarities between the competitions analyzed concerning type of shots or shot combinations used, from which hitting zone on the tennis court the shots or shot combinations are hit and the placement of the different shots when scoring points? Are there any differences or similarities between the competitions analyzed concerning number of valid shots over the net in a rally?

Are there any differences or similarities between the competitions analyzed concerning number of errors (forced and unforced) and winning shots committed in matches?

Method

The sample consisted of a total of 24 matches with 40 different players from three different tournaments which were analyzed using notational analysis software (Dartfish, version 8, Switzerland). Total number of points analyzed were 3154 (AM, n = 968, FF, n = 1068, YNC, n = 1118). Data were compiled in Excel (2013) and descriptive analyses were performed in IBM SPSS Statistics 24. Statistical analyses looking for overall significant differences between the groups were made using Chi square cross tab test. Due to the number of statistical tests that were performed for each domain in the post hoc test, an adjusted significance level of p < 0.001 was used to reduce the risk of Type 1 error.

Results

Significant differences were observed between groups for serve placement, shot used after hitting a serve, type of 2nd last and last shot used, hitting zone and placement by the point winner on last shots. Rallies of longer duration were significantly more frequent in the AM & FF groups compared to the YNC group. Concerning serve outcome, serve return, return placement, shot after serve placement, shot combinations, length on 2nd last and last shot, unforced, forced errors and winners no statistical differences were observed between groups.

Conclusion

This study indicates that higher demands are placed on placement accuracy in the ATP masters and Falu Future tournaments, specifically for the serve, but also for groundstrokes compared to the Swedish youth national championships tournament. This knowledge can be used to identify technical skills and physiological abilities that are important to practise in order to improve performance in tennis on different levels.

(3)

Tables and Figures

Table 1………..13 Table 2………..22 Table 3………..38 Figure 1……….12 Figure 2………..………...14 Figure 3……….15 Figure 4……….17 Figure 5……….18 Figure 6……….19 Figure 7……….24 Figure 8……….25 Figure 9……….26 Figure 10………...27 Figure 11………...27 Figure 12………...28 Figure 13………...29 Figure 14………...30 Figure 15………...31 Figure 16………...32 Figure 17………...32 Figure 18………...33 Figure 19………...34 Figure 20………...35 Figure 21………...35 Figure 22………...36 Figure 23………...37 Figure 24………...37 Figure 25……….……..39 Figure 26………...39 Figure 27……….………...……….…….40

(5)

1

Definitions and abbreviations of relevant concepts

The relevant concepts in this study were stipulated by the author of this study. See descriptions below:

- ATP top level (AM group) = top one to eight on the ATP ranking (Nov 2015). - Future level (FF group) = top 250 to 1250 on the ATP ranking (Sep 2015).

- Swedish youth national level (YNC group) = top one to 300 on the Swedish (SWE) male national ranking (Jan 2016).

- Technical performance = the performance of technical skills in tennis on different competitive levels consisting of different point characteristics. A major factor to sustain the technical skill level in high performance tennis is the level of physiological abilities that the players have developed.

- Technical skills in tennis = Serve (S), return of serve (R), forehand topspin (FH), backhand topspin (BH), slice (SL), volley (VO), smash (SM), drop shot (DS), lob shot (LB= a shot that passes over the head of the player attacking the net) and passing shots (PS= a shot that passes on the left or right hand side of the player attacking the net). - Physiological demands = the demands placed on physiological abilities that are

needed to perform technical skills in tennis on different performance levels. Examples of tennis specific physiological abilities are flexibility, muscle strength, coordination, endurance, speed, agility and power etc.

- Point = a rally starting with a serve and finishing with a winner, forced error or unforced error.

- Point characteristics = describe a rally in a tennis match taking into consideration the following parameters:

o Number of valid shots over the net in a rally.

o Type of shots used and from which location (hitting zone) on the tennis court the shots are hit.

o Placement of different shots in the tennis court. o Shot combinations used to score points.

o How the player moves (between hitting zones) on the court.

(6)

2 - Last shots in a rally = the 2nd last and last shot used by the point winner in a rally. - Shot combination = type of shots used on the last two shots by the point winner in a

rally of four or more valid shots over the net and the location on the tennis court from which the shots are hit and the placement of the shots (cross CR, midcourt MC or down the line DTL. For description see page number 20).

- Valid shot = Shots where the tennis ball passes the net and bounces inside or on the lines of the tennis singles court or in the service box when hitting a serve.

- Unforced error net = shot where the tennis ball is hit in the net under small time pressure.

- Unforced error out = shot where the tennis ball bounces outside the tennis singles court under small time pressure.

- Forced error net = shot where the tennis ball is hit in the net under large time pressure.

- Forced error out = shot where the tennis ball bounces outside the tennis singles court under large time pressure.

- Small time pressure = when a player is able to prepare the shot properly in good balance while moving to reach the location where the ball bounces after the opponent´s shot.

- Large time pressure = when a player hits the shot without time to prepare the shot properly and/or while running fast to reach the location where the ball bounces after the opponent´s shot.

- Winner = a serve ace or a last valid shot that the opponent doesn´t touch with his racket in a rally.

- Placement accuracy= Shots placed closer to the lines on the singles tennis court.

Acknowledgements

A special thanks to my supervisor Johnny Nilsson for guiding me through the process and making me believe that it was possible to accomplish this complex study. I would also like to thank all my colleagues at University of Dalarna for supporting me, the teachers at GIH for your valuable advice and my mother for helping me with the language. At last a special thanks to Falu Tennis club for making the data collection process possible.

(7)

3

1 Introduction

Notational analysis is a fast growing research discipline in tennis. One reason for that is all the new technology that can provide coaches and researchers with data from tournament match play (Hughes & Franks 2010). Studies of tennis up to this date have emphasized movement characteristics on court, rally time, number of shots played in a rally, serve speed and placement, percentage of 1st and 2nd serves points won, aces, differences between surfaces and gender etcetera (e.g. Johnson & McHugh 2006; Gillet et al. 2009; O´Donoghue &

Ballantyne 2004; Over & O´Donoghue 2008; O´Donoghue & Ingram 2001 Unierzyski & Wieczorek 2004). Often statistics from Grand Slam tournaments are used in the studies that the International Tennis Federation officially collects. One problem with this statistics is that it usually looks at single parameters that give a limited picture of the technical performance in tennis on different performance levels.

An important factor that may limit the technical skills of a player is the level of physiological abilities (e.g. speed, endurance, flexibility, strength). Therefore an important factor when doing notational analysis of technical performance in tennis is that it also should consider looking at the physiological demands. To my knowledge no studies have yet been

investigating technical performance on different performance levels (ATP top level, Future level and Swedish youth national level) in relation to the physiological demands in tennis. To increase the knowledge in this area more research is needed.

1.1 General description of the game of tennis

Tennis is a sport consisting of short duration interval work with a work to rest ratio ranging from 1:2 to 1:5. Players have 25 seconds to prepare for the serve after a rally, 90 seconds of rest between games in a match and 120 seconds rest between sets in a match (Kovacs 2007; Kovacs 2006; Murias et al. 2007; Sánchez-Alcaraz Martínez 2015). The performance indicators of tennis can be divided into technical and tactical skills as well as physiological and psychological abilities (Hornery et al. 2007). Technical skills involve all the tennis

strokes (e.g. serve, forehand, backhand, volley) where racket velocity, ball spin and placement accuracy in the tennis court are crucial performance indicators (e.g. Elliot, Reid & Crespo 2003; Kovacs 2007; Roetert, Ellenbecker & Reid 2009; Roetert, Kovacs, Knudson & Groppel 2009). Tactical skills involve decision making, perceptual elements and strategic thinking (Nielson & Mcpherson 2001). In addition tennis is a physiologically highly demanding sport

(8)

4 (Mendez-Villanueva et al. 2007) where abilities like flexibility, muscle strength, endurance, speed, agility and power are typical performance indicators (Chandler 2000; Girard & Millet 2009; Kovacs 2007; Smekal et al. 2001). Different surface characteristics also demand different physiological abilities since the distance run during a match, heartrate and lactate levels are significantly higher when playing on clay courts in comparison with playing on hard courts (Murias et al. 2007). Psychological abilities involve things like positive thinking, stress management and mental skills like relaxation strategies.

To score points in a tennis match one of the players must either hit a winning shot or provoke the opponent to make an error (unforced or forced). The deciding shot can be of different types (e.g. serve, return of serve, forehand, backhand, volley, smash or drop shot), spin (topspin or slice) and placement (DTL, CR or MC). To score points players use different strategies depending on strength and weaknesses in their technical skills and physiological abilities (Larsson 2001). Usually players try to use a strategy to put themselves in a good position for the last shot in a rally by forcing the opponent to hit a shot in a demanding position on the court under time pressure. In a study of top players (ATP ranking 1-50 in the world) results show that, on hard courts, players need to hit approximately 45% of all strokes under time pressure whereas, on clay courts, the corresponding figure is 29% (Pieper, Exler & Weber 2007). This indicates that technical performance of tennis players on top level requires highly developed technical skills and physiological abilities to be able to produce explosive well-coordinated movements while hitting shots frequently under time pressure.

1.2 Physiological demands and technical performance in tennis

Magill & Anderson (2015) classify tennis as a serial open skilled sport since tennis requires a battle between two (singles) or four (doubles) players in an open skill environment where technical skills are repeated many times. Typical for serial open skilled sports like tennis is the unpredictable nature of the game. Thus many different strategies can be used to be successful in competitive tennis on different levels. Therefore it’s hard to create a general physiological demand profile that applies to all tennis players in different matches (Kovacs 2007). Chandler (2000) argues in his review study that 70% of the metabolism during effective playing time in tennis matches is alactic anaerobic, 20% lactic anaerobic and 10% aerobic. That is supported in another review study by Kovacs (2007) taking into consideration that the majority of points in a tennis match in elite tennis last shorter than 10 seconds.

(9)

5 In extreme situations a tennis match can last up to six hours. The final in Australian Open between Rafael Nadal (RN) and Novak Djokovic (ND) in 2012 lasted five hours and 53 minutes. Unpublished Hawkeye data presented by Reid & Duffield (2014) show that the match consisted of 369 points and RN ran 6219 meters compared to ND 6625 meters during the match. The players covered approximately 10% more ground when losing points in comparison when winning points and reached maximum running speed of 20 km/h. During the match 40% of the rallies were longer than eight shots over the net and each player completed over 1100 groundstrokes at average velocities of more than 95 km/h (Reid & Duffield 2014). This is an extreme example but to compete on a top level a player needs to be prepared for this type of work load, even after playing seven matches with a total match time of over 12 hours within a period of two weeks. In a review by Sánchez-Alcaraz Martínez (2015) authors conclude that the average three set match lasts between 60-90 minutes during which the effective playing time is about 15-30% of the total time. When comparing the average match to the match between Nadal and Djokovic it´s possible to understand the huge difference in demands placed on technical skills and physiological abilities between different matches.

In the following Chapters knowledge from research done as to physiological demands and technical performance in tennis will be presented. To be able to identify and discuss technical performance in tennis we also need to look at the physiological demands. The reason for that is that performance of technical skills is dependent on the players level of physiological abilities. These factors will highly influence a tennis player´s performance level.

1.3 Research on physiological demands in tennis

VO2 max for tennis players on male professional level has been shown to vary between 44 to 69 ml/kg/min (e.g. Fernandez, Mendez-Villanueva & Pluim 2006; Kovacs 2006; Kovacs 2007; Smekel et al. 2001; Torres-Luque et al. 2011) and match duration can vary between one to five hours (Kovacs 2007). Even though matches in tennis are of long duration the effective playing time normally varies between 15-30% of total match time (Christmass et al. 1998; Mendez-Villanueva et al. 2007; Reid & Duffield 2014; Sánchez-Alcaraz Martínez 2015). Fernandez, Mendez-Villanueva & Pluim (2006) found that effective playing time varies dependent on surface, gender, ball type, strategies, tactics, performance level and game style of players.

(10)

6 Several studies have been measuring the intensity of match play by using percentage of VO2 max as an indicator for intensity. Data from these studies show that players in general have a mean value of around 50% of VO2 max during match play (Murias 2007; Smekel et al., 2001; Torres-Luque et al. 2011). In a review study mean values range between 45-70% of VO2 max (Fernandez, Mendez-Villanueva & Pluim 2006). In another review by Reid & Duffield (2014) mean VO2 ranges between 60-70% of max. The highest mean value reported for a single game in a match was 76.7% of VO2 max. When two defensive players faced each other mean VO2 was higher compared to when at least one of the two players was an offensive player (Smekel et al. 2001). Mean heartrate (HR) seems to vary between 60-80% of maximum HR during match play (Fernandez, Mendez-Villanueva & Pluim 2006; Reid & Duffield 2014; Torres-Luque et al. 2011) with one study reporting mean HR of 86.1% of maximum HR (Christmass et al. 1998). When the player is serving mean values of both HR and VO2 max are higher compared to when the player is returning serves (Torres-Luque et al. 2011). No differences in mean values of VO2 were observed when comparing clay and hard court matches but the fluctuations between highest and lowest VO2 values during a match were more pronounced on hard courts than on clay courts (Murias et al. 2007). One study analyzed the time players (advanced to elite level) spent in different aerobic intensity zones during simulated match play. Result showed that players spend more than 75% of the time under match play conditions in the low intensity zone and less than 25% in moderate to high intensity zones. Aerobic intensity during match play ranged from 12-83% of VO2 max. The conclusion from the authors was that players with higher aerobic fitness could play at lower aerobic intensities during a match (Baiget et al. 2015).

Lactate levels during match play range between 2-3 to 4-5 mMol/L but can at some point of a match reach as high as 8 mMol/L. (e.g. Christmass et al. 1998; Kovacs 2007;

Mendez-Villanueva et al. 2007; Reid & Duffield 2014; Torres-Luque et al. 2011). Higher values of lactate have been reported for clay compared to hard court matches (Murias et al. 2007; Torres-Luque et al. 2011). Fernandez, Mendez-Villanueva & Pluim (2006) found that lactate levels were higher in service games than when returning the serve with maximal values of 8.6 mMol/L for professional players during match conditions. A conclusion from the study was that tennis is predominantly an anaerobic alactic sport with periods of recovery demanding a good aerobic condition to recover from fatigue (Fernandez, Mendez-Villanueva & Pluim 2006).

(11)

7 In a case study of a four set match between two Davis Cup players, ranked 82 and 120 in the world at the time, results show that rally length and number of strokes per rally decreased during the match. On the other hand HR and perception of effort increased during the match (Gomes et al. 2011). The decrease in rally length, HR and perception of effort may indicate that even highly trained players suffer from fatigue during longer matches. Kovacs (2006) mentions in his review study that hitting accuracy can be reduced by as much as 81% when a player is fatigued. That could be an explanation why rally length decreases in longer matches.

Another important physiological ability for tennis players is to be able to accelerate and decelerate over short distances and change direction moving around the court. In a study of junior players it was found that running activity that involved high acceleration/deceleration was three times as frequent as was high running velocity (Hoppe et al. 2014).

Reid & Schneiker (2008) and the review study by Kovacs (2009) found that players need to run between 2.5 and as much as 12 meters before hitting a shot and repeat this procedure up to 1000 times in a match. Fernandez, Mendez-Villanueva & Pluim (2006) report that players run on average 8-12 meters per rally and change direction on average four times in a rally while moving around the court. These data vary considerably depending on surface, ball type, strategy, tactics, game style and gender. Kovacs (2009) for example reports that directional changes when running on court in a rally can be as many as 15. In French Open 80% of all strokes were played with less than 2.5 meters of position change on the tennis court and 5% required more than 4.5 meter of position change. Murias et al. (2007) found that distances run both per point and per match were significantly higher on clay than on hard court. In the review by Kovacs (2009) it´s reported that players move 70% from side to side of the court, 20% forward and 8% backward. Under medium time pressure the error rate is higher on the backhand side but under large time pressure players commit more errors on the forehand side. A player runs in general longer distances during large time pressure before hitting a shot on hard courts (3.83m) compared to clay courts (3.65m). (Pieper, Exler & Weber 2007)

Another important aspect that affects the physiological demands on a tennis player is rally time and number of shots hit in a rally. Many studies have been reporting data for these variables. (e.g. Christmass et al. 1998; Kovacs 2006: Fernandez, Mendez-Villanueva & Pluim 2006; Mendez-Villanueva et al., 2007; Reid & Duffield 2014; Torres-Luque et al. 2011). These studies show that mean rally time in tennis matches varies between 6-10 s and mean

(12)

8 number of strokes per rally varies between 3-6 strokes for both players. Fernandez, Mendez-Villanueva & Pluim (2006) found that each player hits on average 2.5 to 3 strokes during a rally depending on surface, ball type, strategy, tactics, game style and gender. Average

number of hits per rally in French Open for males was 4.5 and for females 5.8. In Wimbledon the numbers were 2.6 for males and 3.2 for females (Torres-Luque et al. 2011). Mendez-Villanueva et.al (2007) found that in matches between professional players on clay courts 56% of all rallies lasted between 1-6 seconds, 18% 6-9 seconds and 26% 10 seconds or more. Kovacs (2006) found that during a high level collegiate tournament average rally time was 4.8 s when the player in control was an attacking player. Among whole court players rally time ranged between 6-11 s and for baseline players rallies lasted on average 15 s.

1.4 Research on technical performance in tennis

Longitudinal data from the ATP official website between 1991 and 2008 show that the percentage of points won on 1st and 2nd serves are crucial performance indicators in the men´s game. To be able to convert and save break points was also seen as an indicator of success. Players who served more than 4 aces per match and hit less double faults were more likely to win matches. (Ma et al. 2013) The serve accounts for 45 to 60% of the total strokes produced in service games in the Grand Slam tournaments. The second most used shot in a match is the return of serve, the third most used the topspin forehand and the fourth the backhand topspin (Johnson & McHugh 2005). Hallgren & Hjelm (2009) found that more than 50% of the rallies in Grand Slam tournaments were finished in less than three shots in both men´s and women´s games. Another study by Brown, O´Donughue (2008) found that 25 to 40% of all points for both men and women were service points (rallies ending after one or two shots). The highest number of service points was found in Wimbledon and the lowest number in French Open for both genders. Data from O´Donoghue & Ingram (2001) showed that baseline rallies in French Open accounted for 51% of the total number of rallies. In Australian Open the number was 46%, US Open, 35% and Wimbledon, 19%.

The tennis serve has been studied comparing different age groups (junior and senior

professional players), parameters (speed, placement, spin, 1st and 2nd serve) genders (male and female) and surfaces (clay, grass and hard court) etc. (e.g. Gillet et al. 2009; Hizan, Whipp & Reid 2011; O´Donughue & Ballentyne 2004; O´Donughue & Brown 2008; Unierzyzki & Wieczorek 2004). Higher serve speed is positively correlated with winning more points in

(13)

9 both men´s and women´s games even though correlations were not very strong (O´Donughue & Ballentyne 2004). Male senior professionals serve more aces, make less double faults and win significantly more points on the 1st serve compared to females and junior players of both genders. For percentage of points won on the 2nd serve no differences were observed between groups (Hizan, Whipp & Reid 2011). Placement of the serve has also been shown to vary on different surfaces (Unierzyzki & Wieczorek 2004) and in a study of serving strategies on clay courts the topspin serve was frequently used on the 2nd serve. It was also found that the serve down the T-line was the most effective to score points (Gillet et al. 2009). The importance of serve has also been related to the number of shots played in a rally. The results show that in the men´s game the server is more likely to win the point on the 1st serve as long as the rally is finished within four shots or less. On the 2nd serve the advantage for the server was lost if the rally lasted more than two shots. In the women´s game the advantage for the server on the 1st serve was lost in rallies of more than two shots and on the 2nd serve women had no advantage (O´Donughue & Brown 2008). Some studies on serve and return statistics have been

comparing professional senior players with junior players. Hizan, Whipp & Reid (2011) and Stare, Zibrat & Filipcic (2015) found that when the serve returners manage to return a 2nd serve, professional players are more successful to win points than juniors, but on the 1st_serve it´s the opposite. Stare, Zibrat & Filipcic (2015) also found that juniors hit more backhand returns in the net on the 1st_{serve than senior male professionals.}

These data from different studies give us an indication of how important the serve and the return of serve are for success in modern tennis. Research also shows that in the men´s game the trend in the last ten years has been towards rallies of longer duration in the Grand Slam tournaments on all surfaces (Over, O´Donughue 2008). Barbaros-Tudor, Zecic & Matkovic (2011) present data indicating lower speed on both 1st and 2nd serve in 2011 Grand Slam tournaments were compared to year 2010. Another interesting conclusion from the authors (ibid.) is that on slow surfaces, where safe game strategies dominates, the trend goes towards a more aggressive way of playing with a bit more risks taken. That should be compared to faster surfaces, where risk game strategies dominate, the trend goes towards a bit more safe game strategies.

(14)

10 Compared to all the research that’s been done on the serve and return of serve very few

studies have been done looking at point characteristics in different court positions where different shots and combinations of shots can be used to score points. Some research has been done investigating time spent by players in different areas of the court in relation to winning or losing a game. Studies show that the player who wins a game spends more time in the offensive zone, hits more winners and forces the opponent to commit errors more frequently than game losers. Players who win the game also move longer distances on the court but at a slower speed than game losers (Martínez-Gallego et al. 2013a; Martínez-Gallego et al. 2013b). A case study by Nowak & Panfil (2012) of two matches in 2007 US Open and 2008 Australian Open semifinals between Federer and Djokovic is an example of research that’s been looking at shot combinations. In this study the type of shot used, placement of shot and positions in the court where the shot was hit were analyzed. Results showed that players use forehand topspin more than backhand topspin to score points. Differences could be observed from which zone of the court the two players preferred to score points. 37% of the points in the match were scored using risky shots (shots close to line and net and with high speed) strategies. One study of Stare, Zibrat & Filipcic (2015) compares point characteristics between senior professionals and junior players. They found no differences in errors (forced and unforced) and rates of the topspin shots between groups. As to the slice shot the study showed that senior professional players used it more frequently than juniors and made fewer errors.

Hughes & Franks (2010) define performance indicators as parameters or combinations of parameters that will decide whether a performance was successful or not. One way to identify performance indicators in tennis is to collect data of point characteristics from different

technical skills in tournaments on different performance levels. To get more knowledge in this area technical performance for different technical skills in three different tournaments on three different performance levels will be investigated in this study. The tournaments investigated were ATP Masters (AM) year 2015, Falu Future (FF) year 2015 and Youth national championships of Sweden (YNC) year 2016.

(15)

11

2 Aim and research questions

To investigate technical performance in three different tennis competitions (ATP Masters AM, Falu Future, FF & Swedish youth national championships, YNC) by collecting data of point characteristics.

Are there any differences or similarities between the competitions analyzed concerning type of shots or shot combinations used, from which hitting zone on the tennis court the shots or shot combinations are hit and the placement of the different shots when scoring points?

Are there any differences or similarities between the competitions analyzed concerning number of valid shots over the net in a rally?

Are there any differences or similarities between the competitions concerning number of errors (forced and unforced) and winning shots committed in matches?

(16)

12

3 Method

In the following method chapter the sample, procedure of data collection, reliability, statistical analysis and ethical considerations will be described.

3.1 Sample

The total sample consisted of 24 three set matches with 40 different players (see Table 1 on the next page). Eight matches from ATP masters, England, London (2015), eight matches from Falu Future, Sweden, Falun (2015) and eight matches from Youth national

championships of Sweden, Falun (2016). All matches were played indoors on hardcourt surface. ATP Masters (AM) matches were selected from top ranked players (ranking 1-8 in the world on the ATP ranking) who qualified for the masters. In that tournament the eight players are divided into two groups with four players in each group. In each group all players face each other in one match. From the total of twelve group play matches in both groups eight matches were selected for the study with the condition that all players had to participate in two matches (see Figure 1). The other two tournaments consisted of 32 players in the main elimination draw. In Falu Future (FF) all matches from the 2nd round were selected (see table 1). In Youth National Championships of Sweden (YNC) four matches from the 1st round, three matches from the 2nd round and one match from the semifinal were selected (see table 1). As to ranking, match results and number of points analyzed in each match see table 1. The matches from AM were ordered from the website http://www.tennisondvds.com/index.html. The matches from FF and YNC were recorded with a stationary web camera (AXIS 1114, Sweden) by the author of this master thesis. The reason for differences in samples between tournaments were limited access to recorded matches.

(17)

13

Table 1. The sample of all 24 matches analyzed in the AM, FF and YNC tournaments with results. Number (n)

of points analyzed in each match and the contribution of points analyzed in relation to total points analyzed in the study (%) from each match is also presented.

Group Match Rank player one Rank player two Round Result Points analyzed in a match (n)

Contribution from each match (%)

AM 1 ATP (1) ATP (8) Group 6-1, 6-1 92 3% AM 2 ATP (2) ATP (7) Group 6-4, 6-4 110 3% AM 3 ATP (3) ATP (1) Group 7-5, 6-2 117 4% AM 4 ATP (5) ATP (2) Group 6-4, 6-1 103 3% AM 5 ATP (4) ATP (7) Group 7-5, 6-2 133 4% AM 6 ATP (5) ATP (4) Group 6-3, 6-2 122 4% AM 7 ATP (8) ATP (6) Group 7-5, 3-6, 6-3 193 6% AM 8 ATP (3) ATP (6) Group 6-4, 6-2 98 3%

Total 8 4 4 968 31%

Group Match Rank player one Rank player two Round Result Points analyzed in a match (n) Contribution from each match (%) FF 9 ATP (676) ATP (895) 2nd 7-5, 6-0 115 4% FF 10 ATP (542) ATP (1207) 2nd 6-4, 1-6, 7-5 185 6% FF 11 ATP (290) ATP (709) 2nd 4-6, 6-0, 6-2 157 5% FF 12 ATP (740) ATP (670) 2nd 6-2, 6-2 113 4% FF 13 ATP (370) ATP (890) 2nd 7-5, 6-4 123 4% FF 14 ATP (844) ATP (734) 2nd 6-1, 7-5 140 4% FF 15 ATP (654) ATP (1004) 2nd 6-4, 6-4 125 4% FF 16 ATP (423) ATP (1002) 2nd 6-2, 6-1 110 3% Total 8 8 8 1068 34%

Group Match Rank player one Rank player two Round Result Points analyzed in a match (n)

Contribution from each match (%)

YNC 17 SWE (93) SWE (271) 2nd 6-1, 6-3 99 3% YNC 18 SWE (154) SWE (223) 1st 6-2, 6-2 104 3% YNC 19 SWE (112) SWE (182) 2nd 7-6, 3-6, 6-1 186 6% YNC 20 SWE (79) SWE (235) 2nd 7-6, 6-3 150 5% YNC 21 SWE (78) SWE (229) 1st 6-2, 6-4 134 4% YNC 22 SWE (91) SWE (22) semi 6-7, 7-6, 6-4 223 7% YNC 23 SWE (248) SWE (260) 1st 6-0, 6-3 91 3% YNC 24 SWE (270) SWE (243) 1st 6-4, 7-6 131 4%

Total 8 8 8 1118 35%

(18)

14

3.2 Procedure

All the recorded videos from the 24 selected matches were analyzed in notational analysis software (Dartfish, version 8, Switzerland). Matches were analyzed in the following order; AM, followed by FF, followed by YNC. That order was repeated until all 24 matches were analyzed. Data from parameter 1-11 (see Figure 3) were collected using Dartfish 8 Tagging function. This program is useful for this type of observational study where data about behavior is collected because:

1. Video sequences can be watched several times in slow motion.

2. Using the Analyzer function scale coordinates the can be drawn to judge location of events in different stages of a tennis match.

3. Data can by registered for each event in Dartfish 8 software and exported to Excel.

Before collecting data with Dartfish 8 software lines were drawn in the video clips like in Figures 5 and 6 (see Chapters 3.2.2 and 3.2.3) in the Analyzer function. This enabled the researcher to assess hitting zone and shot placement bounce zone. To avoid parallax error the distance measuring function in the Analyzer was used (see Figure 2). By knowing the length of the lines on the tennis court it was possible to calculate correctly the distances in the video clips and draw scale coordinates on the screen as shown in Figure 2.

Figure 2. Lines that were drawn in the video clips for the recorded matches in the analyzer function. See Figures

(19)

15 3.2.1 Data collection

In Figure 3 the tagging panel created by the author of this master thesis is presented.

Figure 3. Tagging panel for data collection in Dartfish 8 software.

When collecting data each point in a match was observed by the author. All points in tennis start with a 1st or 2nd serve either by player one or player two. That was registered by pressing a continuous event button created in the tagging panel and described as parameter 1 in Figure 3.

(20)

16 Then data for both players from each specific point in all 24 matches were registered in the following order by the author by pressing keyword buttons created in the tagging panel shown in Figure 3:

- Parameter 2. Registered where the ball bounced from a serve in the service box.

- Parameter 3. Registered the outcome of a serve. Serves that were returned back into

the court by the opponent were registered as valid serves (in). Serves that generated points directly by forcing the opponent to miss the return of serve were registered as a serve winner. Serves that were valid and didn´t let the returner touch the ball with his racket were registered as an ace. Serves that were not valid were registered as net, out or double fault when a 2nd serve was hit in the net or out.

- Parameter 4. Registered return shot (forehand or backhand) as in or missed and

placement in court of the return as midcourt (MC), down the line (DTL) or cross (CR).

- Parameter 5. Registered the type of shot that the serving player hit after his serve and

its placement (MC, DTL or CR).

- Parameter 6. Registered the number of valid shots in a rally.

- Parameter 7. Registered the type of 2nd_{last shot that was hit by the winning player in} a rally. Winning player means the player who won the specific rally analyzed.

- Parameter 8. Registered the hitting zone and different types of placement of the 2nd last shot by the winning player in a rally. See definitions Figures 5 and 6.

- Parameter 9. Registered the type of last shot that was hit by the winning player in a rally.

- Parameter 10. Registered the hitting zone and different types of placement of the last shot by the winning player in a rally.

- Parameter 11. Registered in what way the point was scored. With a winning shot or by provoking the opponent to commit an unforced or forced error either in the net or outside the singles tennis court.

- To finish data registration for a point. Continuous event button needs to be pressed again. Data for that point are then registered as a separate event in Dartfish 8 software.

(21)

17 An example of data registered for two different types of rallies is shown in Figure 4 below.

Figure 4. Examples of data registered in Dartfish 8 software for two different points played.

As we can see in Figure 4 above a short rally with a winning serve ace (only one valid shot) requires less data to be registered since no return of serve, shot after serve and 2nd last and last shot in the rally is produced. In rallies of four and more valid shots data from all 11

parameters are registered.

To develop the method to assess hitting zone and shot placement bounce zone (see 3.2.2 and 3.2.3) the methods performed in earlier studies by Gillet et al. (2009), Hallgren & Hjelm (2009) and Nowak & Panfil (2012) were used. The author also completed two pilot analyses of two sets in two different matches. The purpose of that was to decide where to draw the lines to divide the court into zones.

(22)

18 3.2.2 Hitting zones

Figure 5. Hitting zones on 2nd_{last and last shot. Zone A to K on each side of the court. A= hitting zone A, B=}

hitting zone B, C= hitting zone C, D= hitting zone D, E= hitting zone E, F= hitting zone F, G= hitting zone G, H= hitting zone H, I= hitting zone I, J= hitting zone J & K= hitting zone K.

In Figure 5 the hitting zones defined with letters from A-K are shown. Hitting zone is defined as the position in the court from which the player strikes the tennis ball with his racket to hit a shot. For example when a player hits a shot located in hitting zone I that shot is registered as a shot produced in that specific hitting zone. In the study hitting zone for the 2nd last and last shot in the rally were registered.

(23)

19 3.2.3 Shot placement bounce zones

Figure 6. Bounce zones in the tennis singles court. Bounce zone 6s was only used to register serve placement.

1= bounce zone 1, 2= bounce zone 2, 3= bounce zone 3, 4= bounce zone 4, 5= bounce zone 5, 5s= bounce zone 5s, 6= bounce zone 6 & 6s= bounce zone 6s.

In Figure 6 bounce zones are shown. In the list below the different types of shot placements used in the tagging panel to register shot placement are defined.

- Serve mid = 1st or 2nd serve where the tennis ball bounced in bounce zone 6. - Serve T-line = 1st and 2nd serve where the tennis ball bounced in bounce zone 6s. - Serve wide = 1st_{and 2}nd_{serve where the tennis ball bounced in bounce zone 5s.}

(24)

20 - Midcourt (MC) short (S) = shots where the tennis ball passed the net and bounced in

bounce zone 6 or 6s.

- Midcourt (MC) medium (M) = shots where the tennis ball passed the net and bounced in bounce zone 4.

- Midcourt (MC) long (L) = shots where the tennis ball passed the net and bounced in bounce zone 3.

- Cross (CR) = shots that were hit from either left to left or right to right side of the tennis court where the tennis ball passed the net and bounced in bounce zone 1 (long), 2 (medium) or 5, 5s (short).

- Down the line (DTL) = shots that were hit from either left to right or right to left side of the tennis court where the tennis ball passed the net and bounced in bounce zone 1 (long), 2 (medium) or 5, 5s (short).

3.2.4 Assessment schedule

An assessment schedule was created to assess the hitting zone, shot placement, winners and errors (forced or unforced) following the criteria listed below:

1. Criteria to assess hitting zone.

a. Shots, in which the majority of the player´s body limbs were located within the zone at ball contact with the racket, were registered as hits from that zone. b. When it was difficult to decide in which zone the majority of the body limbs

were located at ball contact with the racket, the zone, in which the racket was located at ball contact, was used to decide hitting zone.

c. The slow motion and playback functions in Dartfish 8 software were used when needed to ensure correct assessment.

(25)

21 2. Criteria to assess shot placement bounce zone.

a. The definitions of shot placements presented in chapter 3.2.3 were used to assess the placement of serves, serve returns, shot after serve by serving player, 2nd last shot and last shot.

b. When a serve bounced on a line that divides the service box into bounce zones, the bounce zone closest to singles line or midline was registered as placement. c. When a return of serve, shot after serve by serving player, 2nd last or last shot

in a rally bounced on a line that divides the tennis singles court into bounce zones, the bounce zone closest to singles line or baseline was registered as placement.

d. The slow motion and playback functions in Dartfish (8) were used when needed to ensure correct assessment.

3. Criteria to assess winners and errors.

a) Winners were registered if the tennis ball was not touched by the racket of the opponent before bouncing twice.

b) Unforced errors were registered when the researcher considered the player to be under small time pressure when hitting the shot.

c) Forced errors were registered when the researcher considered the player to be under large time pressure when hitting the shot.

d) The slow motion and playback functions in Dartfish (8) were used when needed to ensure correct assessment.

All the assessments were done by the author of this report who has been working as a tennis coach for 20 years and competed in tennis on a high national level. The time to analyze one match varied between 5 to 10 hours.

(26)

22

3.3 Reliability

Table 2. Data from the intra reliability test of match one. Total number of assessments made for each parameter

and the assessment deviance between test and retest in (n) and (%).

Parameter Total (n) Deviant (n) Percent (%)

Rally 127 2 2%

Serve placement 88 4 5%

Serve outcome 127 0 0%

Return 79 6 8%

Return placement 61 4 7%

Shot after serve 54 0 0%

Placement shot after serve 54 5 9%

Number of shots 92 0 0%

2nd last shot 48 0 0%

Hitting zone 2nd last 48 4 8%

2nd last shot placement 47 6 13%

Length 2nd last 47 9 19%

Last shot 61 5 8%

Hitting zone last shot 62 8 13%

Last shot placement 59 5 8%

Length last shot 59 5 8%

Point outcome 92 7 8%

Sum total (± sd) 1205 (±25.3) 70 (±2.8) 6.8% (±5.2)

In Table 2 we can see the result from the intra reliability test of the data collection method. The first match (match one in Table 1) was analyzed twice on separate occasions.

Statistical correlation was done for assessments made at test and retest for all parameters analyzed. Correlation was very strong for all parameters with a mean value of r = 0.97 (±0.04) for all 17 parameters seen in Table 2. A period of two and a half months passed between test and retest to avoid memory bias. The validity of this study will be discussed in chapter 5.2.

(27)

23

3.4 Statistical analysis

By means of the database created in Dartfish 8 software, calculations of the distribution in percent of point characteristics were done for all tennis matches. Data were compiled in Excel (2013) and descriptive analyses were performed in IBM SPSS Statistics 24. First, Chi square crosstabs test was used to evaluate overall effects (p-values), i.e. it examined whether the three-way Chi squared model was significant between the groups on the overall level (AM vs. FF vs. YNC). If significant (p<0.05), chi square crosstabs post-hoc tests were used to analyze specific effects (p-values) between the groups (AM vs. FF, AM vs. YNC, and FF vs. YNC) for each parameter analyzed. Due to the number of statistical tests that were performed for each domain, an adjusted significance level of p <0.001 was used to reduce the risk of Type 1 error. That decision was made in the light of criticism of Bonferroni and other corrections when examining data that includes hypotheses (Perneger 1998).

3.5 Ethical considerations

Gratton & Jones (2010) argue that certain research designs can be questioned for their social and moral acceptability, especially experimental studies with children involved. This study includes data from officially recorded matches from ATP Masters and recorded matches that were officially broadcasted on the web (FF & YNC). All participants that were observed in the three tournaments were older than 18 years and no experiments were accomplished in the study. Based on these facts ethical considerations are not a major problem in this study. One thing that could be an ethical issue is the confidentiality of the players analyzed in the three tournaments. To avoid that unauthorized people had access to data the author stored all data on a compatible hard disk in a safe place. In that way personal information that could be coupled to the data was only available to the author of this master thesis.

(28)

24

4 Results

In the following chapters the results from the analysis of point characteristics for the ATP Masters (AM), Falu Future (FF) and Swedish youth national championship (YNC) will be presented. To see more data from the results watch appendix 1-10.

4.1 Serve characteristics

1st Serve

Group Overall effect Specific effects (p-value)

Wide Mid T-line

YNC vs. AM <0.001 0.001 <0.001 <0.001

YNC vs. FF <0.001 0.243 <0.001 0.001

AM vs. FF <0.001 <0.05 <0.001 0.077

Figure 7. Distribution of serve placement on 1st_{serve for the AM, FF and YNC groups. Inset table shows the}

level of significance for overall effect and specific effects between group parameter distributions.

The AM group placed significantly (p<0.001) more 1st serves to Wide and T-line area and less serves to Mid in the service box than the YNC group. The AM group also placed significantly (p<0.001) less serves to Mid than the FF group. In addition the FF group placed significantly (p<0.001) more serves to T-line and less serves to Mid than the YNC group.

(29)

25

2nd Serve

Wide Mid T-line

YNC vs. AM <0.001 0.576 0.005 <0.001

YNC vs. FF 0.433

AM vs. FF 0.006 0.349 0.085 <0.001

Figure 8. Distribution of serve placement on 2nd_{serve for the AM, FF and YNC groups. Inset table shows the}

level of significance for overall effect and specific effects between group parameter distributions.

The AM group placed significantly (p<0.001) more 2nd serves to T-line than the FF and YNC group. Other differences in parameter distribution between groups can be seen in Figure 8 but none of them reached the adjusted level of significance used in this study.

(30)

26

Figure 9. Distribution of serve outcome (both 1st_{and 2}nd_{serve) for the AM, FF and YNC groups.}

The serve outcome in Figure 9 shows that the AM group, when missing the serve, had slightly more serves that missed (out) the service box than the FF and YNC groups. The AM and YNC groups also had slightly less frequency of serves that were in and returned back in court by the opponent than the FF group. Serves that generated a missed return by the opponent (serve winner) occurred more frequently in the YNC group compared to the AM and FF groups. Concerning missed serves in the net, double faults and ace no obvious differences were observed between groups. None of the differences in parameter distribution between groups were statistically significant at p<0.001.

(31)

27

4.2 Return characteristics

Figure 10. Distribution of return of serve in (over the net and inside the singles lines) and missed for the AM, FF

and YNC groups. FH in= Forehand return in, BH in= Backhand return in, FH missed= Forehand return missed & BH missed= Backhand return missed.

Figure 10 shows that all three groups were hitting more backhand returns back into court than forehand returns. It can also be seen that the YNC group has a higher rate of missed returns both on the forehand and backhand side compared to the AM and FF groups. None of the differences in parameter distribution between groups were statistically significant at p<0.001.

Figure 11. Distribution of placement of returns (FH & BH) for the three groups AM, FF and YNC. DTL= Down

the line, CR= Cross & MC= Midcourt.

Considering the placement of the returns it can be seen in Figure 11 that all three groups returned the majority of the serves to MC. The AM and FF groups placed more returns CR than DTL compared to the YNC group that placed equal number of returns CR and DTL. None of the differences in parameter distribution between groups were statistically significant at p<0.001.

(32)

28

4.3 Shot after serve characteristics

Shot after serve

FH BH SL VO

YNC vs. AM <0.001 0.072 0.012 <0.001 <0.001

YNC vs. FF <0.001 0.363 <0.001 <0.001 <0.001

AM vs. FF 0.537

Figure 12. Distribution of shots used directly after hitting a serve in a rally for the AM, FF and YNC groups.

FH= Forehand, BH= backhand, SL= Slice and VO= Volley. Inset table shows the level of significance for overall effect and specific effects between group parameter distributions.

As can be seen in Figure 12 all three groups used the FH clearly more frequently than the BH directly after hitting the serve in a rally. The YNC group used SL and VO

significantly (p<0.001) more frequently directly after serve compared to the AM and FF groups. In addition the FF group used BH significantly (p<0.001) more frequently than the YNC group directly after hitting the serve.

62% _59% 57% 31% 34% 25% 5% _2% 5% _2% 12% 7% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% AM FF YNC

Shot after serve

(33)

29

Figure 13. Distribution of placement of shot after serve for the AM, FF and YNC groups. CR= Cross, DTL=

Down the line & MC= Midcourt.

Considering placement of the shot after serve (see Figure 13) both the AM and FF groups placed more shots to CR and slightly less shots to MC than the YNC group.None of the differences in parameter distribution between groups were statistically significant at p<0.001.

33% 33% 28% 20% 22% 23% 47% _45% 49% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% AM FF YNC

Shot after serve placement

(34)

30

4.4 Number of valid shot characteristics

Number of valid shots

1-2 3-4 5-6 7-8 9-10 11 more YNC vs. AM <0.001 <0.001 0.584 0.838 0.566 0.126 <0.001

YNC vs. FF 0.001 0.002 0.193 0.196 0.233 0.167 0.001

AM vs. FF 0.625

Figure 14. Distribution of number of valid shots hit in rallies for the AM, FF and YNC groups. Inset table shows

the level of significance for overall effect and specific effects between group parameter distributions.

Figure 14 shows that the frequency of rallies that are finished within one or two valid shots over the net (serve return points) was most common in all three groups. Both in the AM and FF groups the frequency of rallies of eleven and more shots was significantly (p<0.001) higher compared to the YNC group. When comparing rallies of one-two shots the frequency was significantly (p<0.001) higher in the YNC group when compared to the AM group. Comparing the FF group with the YNC group rallies of one-two shots were more frequent in the YNC group (p<0.002) even though it didn´t reach the significance level of p<0.001).

39% 40% 47% 23% 24% _22% 16% _14% 16% 7%_5% 8%_5% _6% 4% 10% 9% 5% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% AM FF YNC

Number of valid shots

(35)

31

4.5 2

nd

last and last shot characteristics

In the following chapters the characteristics for the last two shots in a rally by the point winner will be presented.

4.5.1 Type of 2nd_{last and last shot.}

2nd last shot

Group Overall effect

FH BH SL VO SM DS LB

YNC vs. AM <0.001 0.825 <0.001 <0.001 0.088 0.706 0.293 0.949 YNC vs. FF <0.001 0.936 <0.001 <0.001 0.011 0.494 0.863 0.133 AM vs. FF 0.322

Figure 15. Distribution of shots used on the 2nd_{last shot by the point winner in a rally for the AM, FF and YNC}

groups. FH= Forehand, BH= Backhand, SL= Slice, VO= Volley, SM= Smash, DS= Drop shot & LB= Lob. Inset table shows the level of significance for overall effect and specific effects between group parameter

distributions.

As can be seen in Figures 15 & 16 (see next page) the most frequently used shots on the 2nd last shot and last shot by the point winner are FH followed by BH for all three groups. The YNC group used BH significantly (p<0.001) less frequently and SL significantly (p<0.001) more frequently compared to AM and FF groups on the 2nd last shot (see Figure 15).

Concerning last shot by the point winner it can be seen in Figure 16 that the YNC group used SL significantly (p<0.001) more frequently than the FF group and the FF group used FH significantly (p<0.001) more frequently than the YNC group.

47% 48% 48% 38% 42% 26% 10% _8% 20% 2% 1% 2% 1% _0% 1% 4%0,5%1% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% AM FF YNC

2nd last shot

FH BH SL VO SM DS LB

(36)

32

Last shot

Group Overall effect

FH BH SL VO SM DS LB PS

YNC vs. AM <0.001 0.022 0.236 0.003 0.013 0.304 0.012 0.291 0.175 YNC vs. FF <0.001 <0.001 0.673 <0.001 0.011 0.100 0.611 0.025 0.642 AM vs. FF 0.003 0.003 0.431 0.018 0.950 0.566 0.003 0.230 0.356

Figure 16. Distribution of shots used on the last shot by the point winner in a rally for the AM, FF and YNC

groups. FH= Forehand, BH= Backhand, SL= Slice, VO= Volley, SM= Smash, DS= Drop shot, LB= Lob & PS= Passing shot. Inset table shows the level of significance for overall effect and specific effects between group parameter distributions.

Figure 17. Distribution of the four most frequently used two last shot combinations by the point winner for the

AM, FF and YNC groups. BH/BH= Both 2nd_{last and last shot are a backhand. BH/FH= 2}nd_{last shot is a} backhand and last shot is a forehand. FR/BH= 2nd_{last shot is a forehand and last shot is a backhand. FH/FH=} Both 2nd_{last and last shot are a forehand.}

The four most frequently used shot combinations by the point winner can be seen in Figure 17. In all three groups the FH/FH was the most frequently used combination. None of the differences in parameter distribution between groups were statistically significant at p<0.001.

45% 52% 39% 30% 29% 28% 8%_6% _5%_6% 13%9% 3%5%1%2% 2%2%0%4% 4%2%1%4% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% AM FF YNC

Last shot

FH BH SL VO SM DS LB PS 22% _18% 16% 25% 32% _21% 20% 14% 22% 33% 36% 41% 0% 20% 40% 60% 80% 100% AM FF YNC

Combinations two last shots

(37)

33 4.5.2 Hitting zone characteristics of 2nd_{last and last shots}

Hitting zone 2nd last shot

Group O. E. Specific effects (p-value)

Z A Z B Z C Z D Z E Z F Z G Z H Z I Z J Z K YNC vs AM <0.001 0.141 0.981 0.071 0.486 <0.001 0.203 0.236 0.228 0.011 0.663 0.055 YNC vs FF <0.001 0.488 0.168 0.256 0.724 <0.001 0.104 0.003 0.005 0.818 0.317 0.001 AM vs FF 0.064

Figure 18. Distribution of hitting zones by the point winner on the 2nd_{last shot for the AM, FF and YNC groups.}

Inset table shows the level of significance for overall effect and specific effects between group parameter distributions.

As can be seen in Figure 18 the most frequent hitting zone for the 2nd last shot by the point winner is Zone I for all three groups. In the YNC group the point winner was hitting significantly (p<0.001) more 2nd last shots from Zone E compared to the AM and YNC groups. In addition the point winner in the FF group was hitting significantly (p<0.001) more 2nd last shots from Zone K compared to the YNC group.

0%3%_0% 0%2%_0% 1%3%1% 8%10% 8%11% 7% 24% 5%6% 5%9% 7%4% 15% 19% _12% 40% 30% 31% 10% 11% _9% 3% 5% _1% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% AM FF YNC

Hitting zone 2nd last shot

(38)

34

Hitting zone last shot

Group O. E. Specific effects (p-value)

Z A Z B Z C Z D Z E Z F Z G Z H Z I Z J Z K YNC vs AM 0.004 0.676 0.007 0.028 0.519 0.009 0.993 0.848 0.605 0.007 0.629 0.149 YNC vs FF <0.00 1 0.732 <0.001 0.323 0.172 0.004 0.729 0.043 0.335 0.005 0.247 0.146 AM vs FF 0.603

Figure 19. Distribution of hitting zones by the point winner on the last shot for the AM, FF and YNC groups.

Inset table shows the level of significance for overall effect and specific effects between group parameter distributions.

As can be seen in Figure 19 the most frequent hitting zone for the last shot by the point winner is Zone I for all three groups. In the YNC group the point winner was hitting significantly (p<0.001) more last shots from Zone B compared to the FF group.

1% 1% 1% 8% 7% 13% 3%9% _2%8% _1% 10% 15% 15% 20% 7%_4%13% _6%7%13% 7%_4%12% 27% 27% 21% 9% 10% 8% 4% 4% 3% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% AM FF YNC

Hitting zone last shot

(39)

35 4.5.3 Placement characteristics of 2nd_{last and last shot}

Figure 20. Distribution of placement of the 2nd_{last shot for the AM, FF and YNC groups. CR= Cross,}

DTL= Down the line & MC= Midcourt.

In Figures 20 & 21 it can be seen that the most frequent placement of the 2nd last and last shot by the point winner is MC followed by CR and DTL for all groups. The point winner in the AM group placed significantly (p<0.001) more last shots (see Figure 21) to CR compared to the YNC group. In addition the point winner in the AM group also placed (p=0.003) clearly more last shots to CR compared to the FF group even if the difference was not statistically significant at p<0.001.

Last shot placement

CR DTL MC

YNC vs. AM 0.002 0.001 0.656 0.004

YNC vs. FF 0.242

AM vs. FF 0.009 0.003 0.084 0.182

Figure 21. Distribution of placement of the last shot for the AM, FF and YNC groups. CR= Cross, DTL= Down

the line & MC= Midcourt. Inset table shows the level of significance for overall effect and specific effects between group parameter distributions.

24% 23% 30% _21% 23% _17% 53% _49% 60% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% AM FF YNC

2nd last shot placement

CR DTL MC 37% 29% 28% 24% 29% 25% 39% 42% 47% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% AM FF YNC

Last shot placement

(40)

36

Figure 22. Distribution of placement combinations of the last two shots for the AM, FF and YNC groups.

Both CR or DTL= both the 2nd_{last and last shot by the point winner are placed CR or DTL. Both MC= both the} 2nd_{last and last shot by the point winner are placed to MC. One MC & one CR or DTL= One of the last two} shots by the point winner is placed to MC and one of the two last shots by the point winner is placed to CR or DTL.

In Figure 22 it can be seen that the YNC group used the “Both MC” placement combination more frequently compared to the AM and FF groups. The AM and FF groups used the “Both CR or DTL” placement combinations more frequently than the YNC group. None of the differences in parameter distribution between groups were statistically significant at p<0.001.

32% 34% 26% 23% _21% 28% 45% 45% 46% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% AM FF YNC

Placement combinations last two shots

(41)

37

Figure 23. Distribution of length of the 2nd_{last shot for the AM, FF and YNC groups.}

The most frequent length of both the 2nd last and last shot for the point winner was medium for all groups (see Figures 23 & 24). A difference between the groups can be seen (see Figure 23) in the YNC group who more frequently placed the 2nd last shot short compared to the AM and FF groups. None of the differences in parameter distribution between groups were

statistically significant at p<0.001.

(42)

38

4.6 Movement characteristics of last two shots

Table 3. The four most frequent hitting zones for the last two shots by the point winner in a rally for the AM, FF

and YNC groups.

Groups (n) % (n) % (n) % (n) % Total (n)

Percent of all movements (%)

AM H-I H-I I-I I-I I-J I-J I-H I-H

21 5 % 54 12 % 27 6 % 36 8 % 138 31 % FF H-I H-I I-I I-I J-I J-I I-H I-H

30 6 % 53 11 % 20 4 % 26 5 % 129 26 % YNC E-B E-B I-I I-I I-E I-E E-I E-I

38 9 % 41 10 % 27 6 % 20 5 % 126 30 %

In Table 3 we can see between which zones of the court the point winner most commonly moved when hitting the 2nd last and last shot. For all three groups the most common zone to move in between 2nd last and last shot is hitting zone I (I-I). The AM and FF groups moved a lot between hitting zones I, H and J. In the YNC group players move a lot between hitting zones I, E and B. None of the differences in parameter distribution between groups were statistically significant at p<0.001.

(43)

39

Movement long distance last two shots

LM VM

YNC vs AM 0.111 YNC vs FF 0.121

AM vs FF <0.001 <0.001 <0.001

Figure 25. Lateral and vertical movement long distance (for definition see Figure 26) by point winner between

2nd_{last and last shot for the AM, FF and YNC groups. LM= see definition Figure 26. VM= see definition Figure} 26. Inset table shows the level of significance for overall effect and specific effects between group parameter distributions.

In Figure 25 we can see that the AM and FF groups moved more frequently longer distances in lateral and vertical direction between 2nd_{last and last shot when scoring points compared to} the YNC group even though differences were not significant. The AM group moved

significantly (p<0.001) more frequently vertically (VM) compared to the FF group and the FF group moved significantly more frequently lateral (LM) compared to the AM group.

Figure 26. Definitions of lateral movement (LM) and vertical movement (VM) long distance between hitting

zones on the 2nd_{last and last shot by the point winner.}

Lateral and vertical movements are done all the time by players. But in Figure 26 we refer to LM and VM movements that are done over long distances. See Figure 5 in chapter 3.2.2 to get a picture of what is meant by long distance movement in lateral (e.g. move from G-K) or vertical (e.g. move from G-C) directions.

2% 7% 5% 3% 1% 2% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% AM FF YNC

Lateral & vertical movement

LM VM

G-K K-G G-J J-G K-H H-K K-D D-K G-F F-G

G-B G-C H-A H-B H-C I-A I-B I-C J-A J-B J-C K-A K-B K-C LM= Definition of lateral movement long distance between 2nd last and last shot

VM= Definition of vertical movement long distance between 2nd last and last shot Hitting

zone Hitting

(44)

40

4.7 Point outcome characteristics

Figure 27: Distribution of point outcome for the AM, FF and YNC groups. UF net= Unforced error net, F net=

Forced error net, UF out= Unforced error out, F out= Forced error out & Winner (for definitions see chapter 1).

The AM group was hitting more winners and committed less unforced errors compared to the FF and YNC groups. Considering forced errors the differences between groups were minimal (see Figure 27). None of the differences in parameter distribution between groups were statistically significant at p<0.001.

(45)

41

4.8 Results summary

The AM group placed significantly (p<0.001) more 1st_{serves to Wide and T-line area and less} 1st_{serves to Mid in the service box than the YNC group. The AM group also placed}

significantly (p<0.001) less 1st_{serves to Mid than the FF group. In addition the FF group} placed significantly (p<0.001) more 1st_{serves to T-line and less serves to mid than the YNC} group. The AM group placed significantly (p<0.001) more 2nd serves to T-line than the FF and YNC groups. Concerning serve outcome, return of serve and return placement no statistical differences were observed between groups.

The YNC group used SL and VO significantly (p<0.001) more frequently on the shot after the serve compared to the AM and FF groups. In addition the FF group used BH significantly (p<0.001) more frequently than the YNC group directly after hitting the serve. Concerning placement of the shot after serve no statistical differences between groups were observed. Both in the AM and FF groups the frequency of rallies of eleven and more shots was significantly (p<0.001) higher compared to the YNC group. When comparing rallies of one-two shots the frequency was significantly (p<0.001) higher in the YNC group when compared to the AM group. The point winner in the YNC group used BH significantly (p<0.001) less frequently and SL significantly (p<0.001) more frequently compared to the AM and FF groups on the 2nd last shot. On the last shot by the point winner it can be seen in Figure 16 that the YNC group used SL significantly (p<0.001) more frequently than the FF group and FF group used FH significantly (p<0.001) more frequently than YNC group. Concerning shot combinations by the point winner no statistical differences between groups were observed. In the YNC group the point winner was hitting significantly (p<0.001) more 2nd last shots from Zone E compared to the AM and YNC groups. In addition the point winner in the FF group was hitting significantly (p<0.001) more 2nd last shots from Zone K compared to the YNC group. In the YNC group the point winner was hitting significantly (p<0.001) more last shots from Zone B compared to the FF group. For all three groups the most common zone to move in between 2nd last and last shot is hitting zone I (I-I). The AM and FF groups moved a lot between hitting zones I, H and J. In the YNC group players move a lot between hitting zones I, E and B. Differences between groups were not statistically significant. The point winner in the AM and FF groups moved more frequently longer distances in lateral and vertical direction on the tennis court between 2nd last and last shot compared to the YNC group even though differences were not significant. The AM group moved longer distances

Technical performance on ATP top level, future level and Swedish youth national level male tennis tournaments : Notational analysis of point characteristics in three different tournaments on three different performance levels