Small-sided games (SSGs) are commonly used in soccer to improve the interaction among players and focus on technical and tactical abilities because they allow more time on the ball under game-like conditions compared with the genuine game. Thus, most exercise sessions in soccer have SSGs played with a reduced number of players on a smaller pitch (17). In recent years, the physiological stress generated in SSGs has been examined with respect to its potential to improve aerobic fitness (9,16). The advantages of SSGs are commonly considered as reaching intensities of 90–95% of the maximum heart rate (HRmax), which is proposed to enhance soccer-specific endurance capacity, develop game-specific muscles, improve technical and tactical abilities in game-specific conditions, and assume an effective transfer to match play (6,11,12). Impellizzeri et al. have shown that the effectiveness of SSGs regarding the improvement of aerobic fitness is similar to the use of common fitness training such as interval running with an intensity of 90–95% of HRmax (12). Dellal et al. reported that some SSG formats reveal HR responses comparable with short-duration intermittent running (3). Because of the small variability in physiological responses during SSGs, it is suggested that these methods are suitable to stimulate aerobic training (7). Furthermore, it is assumed that players perceive SSGs as less intense than generic soccer training (8). Therefore, SSGs appear to be especially time effective by combining motor learning, team cohesion components, and aerobic fitness training.
However, the physiological response observed in SSGs change because of its formats. For example, Katis and Kellis showed that playing with 3 vs. 3 players provides a higher stimulus for aerobic fitness than the 6 vs. 6 does (13). Furthermore, Hill-Haas et al. stated that 2 vs. 2 reveals significantly higher responses in the HR and blood lactate as compared with 4 vs. 4 or 6 vs. 6 when the relative pitch area is kept constant. Therefore, they suggested using 2 vs. 2 to increase aerobic fitness (9).
However, some of the SSGs did not reveal game-like intensities and therefore might not be suitable to maximize the development of soccer-specific aerobic fitness. Given the published findings on SSG formats and their associated physiological responses, we intended to create 3 SSG formats, differing in the number of players, with intensities well comparable with soccer-specific game intensities. Therefore, the findings of our study could assist soccer coaches in accurately choosing specific SSG formats if the development of soccer-specific aerobic fitness reflects the key aspect of the training session. Although Hill-Haas et al. did not find significant differences in the physiological response revealed by continuous or intermittent SSGs (10), we favor an intermittent structure because it has been reported that high-intensity interval training maximizes the cardiac adaptations (20). Thus, we hypothesize that intermittent SSG formats of 2 vs. 2, 3 vs. 3, and 4 vs. 4 players with a constant grid ratio per player of 1:150 reveal game-like physiological responses and therefore are well suited to increase soccer-specific aerobic fitness.
Experimental Approach to the Problem
In this study, we investigated the physiological responses and time-motion characteristics of 3 SSG formats in elite youth soccer players. The SSGs were designed according to the existing knowledge on the parameters influencing the physiological response of these games. We hypothesized that all game formats reveal game-like intensities and are able to enhance the aerobic fitness, for example, by provoking HRs > 90% of individuals' maximum.
The physiological responses and time-motion characteristics (e.g., time spent in different speed zones) were analyzed by global positioning system (GPS) technology including HR monitoring (GPS, Forerunner 305, Garmin Inc., Olathe, KS, USA). Furthermore, blood samples were extracted to analyze the concentration of blood lactate. Differences in the physiological responses and time-motion characteristics between the SSG formats were analyzed by repeated-measures analysis of variance (ANOVA) and effect sizes according to Batterham and Hopkins (2).
All the investigations were carried out after preseason training and within 4 weeks of the first quarter of the competitive season (September 2009). Regarding specific soccer skills, each player was ranked by the coach with respect to his abilities in passing, close ball control, shooting, and game sense using a 5-point Likert scale (1 = “outstanding,” 5 = “below average”). Regarding their aerobic fitness level, the players were ranked by their relative maximum oxygen uptake (V[Combining Dot Above]O2max [milliliters per minute per kilogram]). Therefore, in the first and second weeks of the investigation, V[Combining Dot Above]O2max was estimated during a treadmill test until maximum exhaustion (ELG 2, Woodway, Germany). All the subjects started with a velocity of 8 km·h−1 and an incline of 1%. Every 3 minutes, a blood sample was drawn from the earlobe before the velocity was increased by 2 km·h−1. V[Combining Dot Above]O2 and HR were measured continuously with a mobile oxygen analyzer including a wireless HR monitor (MetaMax3b, Cortex Biophysik GmbH, Germany). Breath-by-breath values were processed by determining the 3-point median to exclude outliers and were then averaged over 15 seconds. Finally, the overall ranking was determined by the calculation of (skill rank × 0.75) + (aerobic fitness rank × 0.25). This was used to allocate players into well-matched teams. Furthermore, HRmax was determined by the treadmill test.
Seventeen male soccer players (age 14.9 ± 0.7 years, height 179.0 ± 7.6 cm, body mass 70.4 ± 7.5 kg, body mass index 21.9 ± 1.4 kg·m−2, V[Combining Dot Above]O2max 61.4 ± 4.5 ml·kg−1·min−1, HRmax 199.6 ± 7.3 b·min−1) were recruited from a local high performance center. All the players were members of the same youth team (<16 years) competing in the top-level age group competition in Germany, with a minimum playing experience of 7 years. Training frequency of the youth squad was 4 d·wk−1. The subjects and parents were carefully familiarized with the aims, procedures, and risks of the study. Afterward, the parents and youth players gave written informed consent. It was clearly stated that all the procedures of the study were in accordance with the ethical principles for medical research involving human subjects as published by the World Medical Association in the Declaration of Helsinki (21).
The games were played at the same time of the day, at the beginning of the training session. Therefore, hydration status, nutrition, and activity profile were assumed to be consistent. The order of the games was 3 vs. 3 players on Monday in the third week, 4 vs. 4 players on Thursday in the third week, and 2 vs. 2 players on Monday in the fourth week of the investigation. As requested by the coach, SSGs were not scheduled the day before a competitive game. When the ball was kicked out of play, immediate access to replacement balls was made possible along the entire pitch. Before the games were started, the coach conducted a 15-minute warm-up. The players performed low-intensity activities before they had to play the SSG.
A brief overview about SSG variables is given in Table 1. In this study, we altered the number of players and adjusted the pitch to achieve a constant ratio of pitch area per player, similar to that in previous studies (15,16,19). Furthermore, the game duration was increased from 12 minutes in 2 vs. 2 to 18 minutes in 4 vs. 4. The overall game duration was split into thirds. The 1.5-minute break was used to retrieve blood samples for analysis of blood lactate concentration during the SSG. The end of each third was announced by a whistle. The players were instructed to proceed quickly to their assigned researcher who took the blood sample.
Heart rate and time-motion parameters were measured using mobile GPS units including a coded wireless HR monitor (Forerunner 305, Garmin Inc.) collecting data at 1 Hz. The GPS unit was taped firmly on the left shoulder close to the neck using adhesive tape. This allowed for unrestricted game participation and for measuring time-motion parameters of the trunk. The HR monitor was strapped around the thorax after moistening the electrodes to ensure a solid transmission of the HR. Exercise intensity was expressed as a percentage of HRmax using 4 defined intensity zones: <75% HRmax, 75–84% HRmax, 85–89% HRmax, and >90% HRmax (5). Because of the time delay between external loading and adjustment of the HR, HR values of the first 30 seconds of each third were removed from analysis.
To differentiate the speed of the players, 7 speed zones were chosen as done in previous studies (9,12,16): walking (0–5.2 km·h−1), jogging (5.3–7.6 km·h−1), low-speed running (7.7–10.2 km·h−1), moderate-speed running (10.3–13.9 km·h−1), high-speed running (14.0–17.1 km·h−1), sprint running (17.2–26.7 km·h−1), and maximum sprinting (>26.7 km·h−1).
The GPS units were started by the investigators. Afterward, the players had to stand at the baseline of the court for 20 seconds until the third was started by a whistle. Combined with the fixed time protocol of the SSGs, this allowed for identifying and excluding activities conducted during the breaks from GPS and heart frequency data.
The dependent variables, including HR, HRmax, blood lactate, and various time-motion parameters were analyzed for significant differences by one-way repeated-measures ANOVA (PASW Statistics, IBM Corporation, Armonk, NY, USA). The independent variable was the SSG format. When required, Scheffés post hoc comparisons were applied to identify significant differences between specific SSG formats.
Furthermore, effect sizes between each SSG format were calculated according to Batterham and Hopkins (2). Therefore, the minimal clinically important difference (MCID) was defined as the smallest difference perceived to be clinically significant. The MCID values were a difference of 0.5 mmol·L−1 for blood lactate, 2.5 b·min−1 for HR, 2.5% for HRmax, and 0.25 km·h−1 for average velocity. For the time spent in the 7 speed zones and in the 4 intensity zones with respect to HRmax, the MCID value was set to 1% of the game duration. For the number of sprints, average sprint duration and sprinting distance, the MCID was set to 1, 1 second, and 1 m, respectively. The probabilities that the effect between the SSG formats was a decrease, trivial, or an increase of the observed parameter were expressed as percentages, reflecting the following descriptors: <0.5% = almost certainly not; 0.5–5% = very unlikely; 5–25% = unlikely; 25–75% = possibly; 75–95% = likely; 95–99% very likely; >99% = almost certainly (2,4). The level of significance was set at p ≤ 0.05, and all results are given as mean ± SD.
The SSGs of 2 vs. 2 players revealed the highest responses in the HR (186 ± 7 b·min−1), HRmax (93.3 ± 4.2%), and blood lactate (5.5 ± 2.4 mmol·L−1) but showed the lowest average velocity (6.5 ± 0.4 km·h−1). Descriptive data of the HR, HRmax, blood lactate, and velocity and the results of the repeated-measures ANOVA for all SSGs are given in Table 2. The decrease in HR observed between 2 vs. 2 and 3 vs. 3 players remained very likely trivial (3% decrease/96% trivial/1% increase), whereas the HR was likely to decrease between 2 vs. 2 and 4 vs. 4 (91% decrease/9% trivial/0% increase) and between 3 vs. 3 and 4 vs. 4 players (76% decrease/24% trivial/0% increase). Similar to the absolute HR, the average percentage of HRmax was very likely trivial between format 2 vs. 2 and 3 vs. 3 (3% decrease/96% trivial/1% increase) but was likely to decrease between formats 2 vs. 2 and 4 vs. 4 (91% decrease/9% trivial/0% increase) and between 3 vs. 3 and 4 vs. 4 (77% decrease/23% trivial/0% increase). The concentration of blood lactate was very likely to decrease from 2 vs. 2 to 3 vs. 3 (96% decrease/3% trivial/1% increase) and almost certainly decreased from 2 vs. 2 to 4 vs. 4 (100% decrease). An unclear development of the lactate concentration was found between formats 3 vs. 3 and 4 vs. 4 (39% decrease/37% trivial/25% increase). The mean velocity performed by the players was almost certain to increase between formats 2 vs. 2 and 3 vs. 3 (100% increase) and between formats 2 vs. 2 and 4 vs. 4 (100% increase). The mean velocity possibly increased between formats 3 vs. 3 and 4 vs. 4 (25% decrease/18% trivial/58% increase).
The relative distribution of the intensity zones expressed by the percentage of HRmax illustrates that the players performed 78.8 ± 26.3% of the game duration of SSG 2 vs. 2 with an HR >90% of HRmax, whereas SSGs 3 vs. 3 and 4 vs. 4 were performed with at least 90% of HRmax for 67.8 ± 26.6 and 52.8 ± 26.5% of the game duration, respectively. The difference in the time spent in the highest intensity zone between formats 2 vs. 2 and 3 vs. 3 is possibly trivial, but between formats 2 vs. 2 and 4 vs. 4 and between formats 3 vs. 3 and 4 vs. 4 the time acting >90% of HRmax was very likely to decrease. A detailed overview about the relative time spent in each intensity zone including p values and effect sizes is given in Figure 1.
The time spent in speed zone “walking” (≤5.2 km·h−1) in relation to the game duration dropped from 46.2 ± 3.6% in SSG 2 vs. 2 to 41.1 ± 6.8% in SSG 4 vs. 4 (p = 0.008). The time spent in the remaining speed zones increased, but only “moderate-speed running” and “maximum sprinting” reached the level of significance. A detailed overview about the changes in the speed zones including p values and effect sizes is given in Figure 2. Generally, the differences between 2 vs. 2 and 4 vs. 4 revealed the highest effect sizes. The difference between formats 2 vs. 2 and 3 vs. 3 was more pronounced than the difference between 3 vs. 3 and 4 vs. 4.
The number of sprints (≥17.2 km·h−1) per minute did not significantly change within the SSG formats, but significant differences were found between 2 vs. 2 and 3 vs. 3 and between 2 vs. 2 and 4 vs. 4 for the maximum sprint duration (p = 0.037 and p < 0.001, respectively) and for the maximum sprint distance (p < 0.001 and p < 0.001, respectively, Figure 3). The number of sprints per minute was possibly trivial between the formats 2 vs. 2, 3 vs. 3 and 4 vs. 4 (effect sizes not shown). Generally, the sprint duration and the sprinting distance were likely or very likely to increase between formats 2 vs. 2, 3 vs. 3 and 4 vs. 4 (effect sizes not shown), except for the sprinting distance between formats 3 vs. 3 and 4 vs. 4, which was found to be possibly trivial (0% decrease/52% trivial/48% increase).
Data losses during HR recording because of malfunction were found in 2 players during 3 vs. 3 and in 1 player during 2 vs. 2 and were excluded from data analysis.
This study aimed to identify soccer-specific SSG formats with intensities >90% HRmax, which is considered to be ideally suited to improve aerobic fitness and to describe time-motion characteristics of these formats. Generally, the findings suggest that the investigated formats 2 vs. 2, 3 vs. 3, and 4 vs. 4 are suitable for aerobic fitness training with different accentuations.
Heart rate monitoring during exercise offers a sound method to determine exercise intensity and, furthermore, is able to distinguish between recovery training, low-intensity, and high-intensity training by the mean of %HRmax (18). When applying these criteria, all the formats investigated in this study could be classified as high-intensity exercises as their mean HRmax reached (4 vs. 4) or exceeded (2 vs. 2, 3 vs. 3) 90% of HRmax. Our findings that smaller game formats increase exercise intensity are in agreement with those of most previously published studies (e.g., [15,17]).
Soccer is characterized as an intermittent sport provoking high demands on the anaerobic and aerobic energy system with mean and peak HRmax rates of around 85 and 98%, respectively (1). Furthermore, it has been stated that blood lactate is not well suited for estimating muscle lactate during soccer matches, but it allows for approximating the overall accumulation of lactate production (14). All SSG formats investigated in this study reached mean HRs >85% HRmax and mean blood lactate values of at least 4 mmol·L−1. Thus, one could consider these formats as a sufficient stimulus for improving aerobic fitness. However, the HRmax and blood lactate values indicate that playing with 2 vs. 2 players induces peak loading including pronounced anaerobic energy supply. The players performed nearly 80% of the game with an HR >90% HRmax. Consequently, this format could be used if intensities similar to peak loadings during match play are intended. In contrast, the players performed around 70 and 50% of the game duration with an HR >90% HRmax during 3 vs. 3 and 4 vs. 4, respectively. The findings of the statistical analysis suggest that both formats do not predominantly challenge the anaerobic energy supply and provoke similar lactate concentrations. Furthermore, the effect sizes demonstrate that both formats mainly differ in the HR response. Thus, an SSG of 3 vs. 3 players with a game duration of 3 × 5 and 1.5 minutes of passive rest could be recommended for a general improvement of aerobic fitness. Because the mean HR response of format 4 vs. 4 is around 90% of HRmax, we assume that the format will provoke less adaptations of the aerobic fitness as compared with format 3 vs. 3.
The time-motion parameters observed in our study indicate a shift toward higher velocities if the number of players and the grid total area increase, suggesting that the players make use of the larger pitch and off-the-ball movements as reflected by the increased sprint duration and maximum sprint duration. Consistent with the physiological responses, the differences are more pronounced between SSG formats 2 vs. 2 and 3 vs. 3 as compared with the differences between 3 vs. 3 and 4 vs. 4. These findings are in accordance with the study by Hill-Haas et al., who also found a decreased distance covered within a velocity lower than 6.9 km·h−1 but an increased running or sprinting distance (9). Furthermore, the overall distance covered by the players in relation to game duration did not show significant differences or noteworthy effect sizes (data not shown). Thus, we agree with Hill-Haas et al. that the total distance is an inappropriate indicator of the work rate in SSGs (9).
The subjects investigated in this study were of a similar age but showed superior V[Combining Dot Above]O2max (61.4 vs. 54.8 ml·kg−1·min−1) compared with the subjects measured with a similar instrumentation by Hill-Haas et al. (9). Furthermore, the number of sprints per minute, the duration, and the sprinting distance were remarkably higher in our subjects. Additionally, the mean HRmax of SSG 2 vs. 2 increased in our study to 93.3% compared with 89.0% as described by Hill-Haas et al. (9). The removal of the HR during the first 30 seconds of each third, as conducted in our study, could explain this discrepancy to a lesser extent, although continuously played SSGs were shown to result in higher HRmax responses when compared with intermittent SSGs (10). Therefore, we assume that a pronounced aerobic fitness in our subjects explains—to a large extent—the higher intensity of the SSG formats in our subjects. Thus, more information is needed to show whether and to what extent fitness level, technical and tactical abilities alter the physiological responses of the subject and time-motion characteristics when performing SSGs.
The use of GPS systems offers a practical method to describe time-motion characteristics during SSGs. However, because of the low sampling rate of 1 Hz, the reliability is reported to be good if the players move with low or moderate velocities but is poor if the players go faster than 20 km·h−1 (9). Thus, the top speed players run by a fraction of a second are not captured by the GPS system. However, the study demonstrates that players spent most of the game in the slower categories; thus, a general description of the time-motion parameters is possible using GPS systems.
The blood lactate values could potentially be biased because of inconsistencies of the blood sampling because 2 players were assigned to 1 researcher who took blood samples. Although the researchers were highly trained and obtained both samples within 30 seconds, the concentrations of blood lactate could have been altered according to the time between the end of the play time and drawing the blood sample. However, according to the lactate concentration observed, we only assume a small decay of the absolute blood lactate concentration.
The results of this study show pronounced physiological responses compared with previously published findings of Hill-Haas et al. (9) and Katis et al. (13). Both previous studies reported HR values ranging from 83% (for SSG formats 6 vs. 6) to 88% (3 vs. 3) or 89% (2 vs. 2) of HRmax. Our findings suggest that the SSG formats 2 vs. 2, 3 vs. 3, and 4 vs. 4 with the characteristics given in Table 1 reach 93, 92, and 90% of HRmax, respectively. Because it is has been stated that HRs of 90–95% of HRmax may effectively increase soccer-specific aerobic fitness, soccer coaches may prefer these game formats if the improvement of the aerobic fitness is a major objective of the training session. Moreover, our findings suggest that the physiological responses of the game formats tested in this study are well comparable with the physiological responses in match play. Thus, these SSG formats ensure practicing under game-like physiological conditions and may optimize the transfer to the genuine game.
However, our findings suggest small differences in the physiological response although we attempted to create game formats revealing very similar physiological responses. The characteristics of SSG 2 vs. 2 resulted in a pronounced accumulation of blood lactate around 5.5 mmol·L−1. Instead, SSGs 3 vs. 3 and 4 vs. 4 showed a similar blood lactate concentration of 4.3 and 4.4 mmol·L−1, respectively. Therefore, soccer coaches may prefer 2 vs. 2 when aiming for higher loads leading to a greater demand of the anaerobic energy supply.
Small-sided game 3 vs. 3 challenges the anaerobic energy supply to a lesser extent compared with 2 vs. 2 but reveals a remarkably higher HR response compared with 4 vs. 4. Thus, we recommend using SSG 3 vs. 3 for the development of aerobic fitness, because it remains a predominantly aerobic activity and players spend nearly 70% of the game duration >90% of HRmax.
Partial funding was received from the University of Bremen (ZF-09/606/09).
1. Bangsbo J, Mohr M, Krustrup P. Physical and metabolic demands of training and match-play in the elite football player. J Sports Sci 24: 665–674, 2006.
2. Batterham AM, Hopkins WG. Making meaningful inferences about magnitudes. Int J Sports Physiol Perform 1: 50–57, 2006.
3. Dellal A, Chamari K, Pintus A, Girard O, Cotte T, Keller D. Heart rate responses during small-sided games and short intermittent running training in elite soccer players: A comparative study. J Strength Cond Res 22: 1449–1457, 2008.
4. Gabbett TJ, Mulvey MJ. Time-motion analysis of small-sided training games and competition in elite women soccer players. J Strength Cond Res 22: 543–552, 2008.
5. Gore C. Physiological Tests for Elite Athletes. Champaign, IL: Human Kinetics, 2000.
6. Helgerud J, Engen LC, Wisloff U, Hoff J. Aerobic endurance training improves soccer performance. Med Sci Sports Exerc 33: 1925–1931, 2001.
7. Hill-Haas S, Coutts A, Rowsell G, Dawson B. Variability of acute physiological responses and performance profiles of youth soccer players in small-sided games. J Sci Med Sport 11: 487–490, 2008.
8. Hill-Haas SV, Coutts AJ, Rowsell GJ, Dawson BT. Generic versus small-sided game training in soccer. Int J Sports Med 30: 636–642, 2009.
9. Hill-Haas SV, Dawson BT, Coutts AJ, Rowsell GJ. Physiological responses and time-motion characteristics
of various small-sided soccer games in youth players. J Sports Sci 27: 1–8, 2009.
10. Hill-Haas SV, Rowsell GJ, Dawson BT, Coutts AJ. Acute physiological responses and time-motion characteristics
of two small-sided training regimes in youth soccer players. J Strength Cond Res 23: 111–115, 2009.
11. Hoff J. Training and testing physical capacities for elite soccer players. J Sports Sci 23: 573–582, 2005.
12. Impellizzeri FM, Marcora SM, Castagna C, Reilly T, Sassi A, Iaia FM, Rampinini E. Physiological and performance effects of generic versus specific aerobic training in soccer players. Int J Sports Med 27: 483–492, 2006.
13. Katis A, Kellis E. Effects of small-sided games on physical conditioning and performance in young soccer players. J Sports Sci Med 8: 374–380, 2009.
14. Krustrup P, Mohr M, Steensberg A, Bencke J, Kjaer M, Bangsbo J. Muscle and blood metabolites during a soccer game: Implications for sprint performance. Med Sci Sports Exerc 38: 1165–1174, 2006.
15. Owen A, Twist C, Ford P. Small-sided games: The physiological and technical effect of altering pitch size and player numbers. Insight 7: 50–53, 2004.
16. Rampinini E, Coutts AJ, Castagna C, Sassi R, Impellizzeri FM. Variation in top level soccer match performance. Int J Sports Med 28: 1018–1024, 2007.
17. Rampinini E, Impellizzeri FM, Castagna C, Abt G, Chamari K, Sassi A, Marcora SM. Factors influencing physiological responses to small-sided soccer games. J Sports Sci 25: 659–666, 2007.
18. Reilly T. The Science of Training—Soccer: A Specific Approach to Developing Strength, Speed and Endurance. London, United Kingdom: Routledge, 2007.
19. Williams K, Owen A. The impact of player numbers on the physiological responses to small sided games. J Sports Sci Med 6 (Suppl. 10): 100, 2007.
20. Wisloff U, Ellingsen O, Kemi OJ. High-intensity interval training to maximize cardiac benefits of exercise training? Exerc Sport Sci Rev 37: 139–146, 2009.
21. World Medical Association. Ethical Principles for Medical Research Involving Human Subjects. 2010; Available from: http://www.wma.net/en/30publications/10policies/b3/index.html
. Accessed February25, 2010.