Secondary Logo

Journal Logo

Original Research

Familiarization, Reliability, and Evaluation of a Multiple Sprint Running Test Using Self-Selected Recovery Periods

Glaister, Mark1; Witmer, Chad2; Clarke, Dustin W2; Guers, John J2; Heller, Justin L2; Moir, Gavin L2

Author Information
Journal of Strength and Conditioning Research: December 2010 - Volume 24 - Issue 12 - p 3296-3301
doi: 10.1519/JSC.0b013e3181bac33c
  • Free



Tests of multiple sprint performance are a popular means of evaluating the performance capabilities of athletes involved in field and court sports. On the basis of the results of several time-motion analyses, these tests have typically comprised several (5 ≤ n ≤ 20) short (≤6 s) sprints interspersed with relatively short (≤60 s) passive recovery periods (14). The key performance determinants arising from these tests are a) the ability to produce a high sprint speed and b) the ability to resist fatigue and thereby maintain a high sprint speed for the duration of the test. Although measures of the former have been shown to have good test-retest reliability, the same is not true of the latter (5,7). In fact, of 8 different approaches used to quantify fatigue in multiple sprint work, the best only gives a test-retest coefficient of variation (CV) of approximately 30% (8).

Because the recovery of sprint performance is an aerobic process, primarily involving the resynthesis of phosphocreatine (PCr) and the removal of accumulated intracellular inorganic phosphate (6), it follows that individuals with a high level of aerobic fitness should have an enhanced capacity to recover between sprints and thereby maintain performance (6). However, although there is some evidence that endurance-trained athletes display less fatigue in multiple sprint tests than team-sport players (1,11), the effects of endurance training on repeated sprint ability are inconclusive (4,10). Similar contradictions exist in the results of investigations into the relationship between one of the key parameters of endurance fitness, namely maximal oxygen uptake (O2max) and fatigue during multiple sprint work (6). Because many of these discrepancies may be the result of the large variability associated with fatigue measures, an alternative and somewhat radical approach to address this problem may be to allow individuals to choose their own recovery time in a multiple sprint test, based on individual perceptions of recovery, and to use mean recovery time as an index of fatigue. In effect, it is anticipated that those individuals with the highest levels of fatigue would typically choose the longest recovery times.

To evaluate the efficacy of using the above approach as a means of reliably quantifying repeated sprint ability, 3 aims were identified. First, if individuals are able to accurately determine their own recovery time, it was essential to investigate whether there were any learning effects associated with the process. Second, it was important to establish how reliable self-selected recovery is once any of the aforementioned learning effects have been reconciled, particularly if this approach is to be used as a routine means of evaluating repeated sprint ability. Third, if the duration of self-selected recovery was indeed related to an individual's level of aerobic fitness, it was important to evaluate the magnitude of that relationship.


Experimental Approach to the Problem

To provide sufficient data for familiarization and reliability analysis, all subjects completed 4 trials of the multiple sprint test, which consisted of 12 × 30 m straight-line sprints on an indoor synthetic running surface, interspersed with self-selected recovery periods. After completion of the multiple sprint trials, subjects completed a graded exercise test on a motorized treadmill (Q-Stress TM55, Quinton, Inc., Bothell, WA, USA) to evaluate the relationship between self-selected recovery time and O2max. All trials were completed at approximately the same time of day with 7 days between trials 1 and 2 (to allow recovery from any initial postexercise muscle soreness) and a minimum of 48 hours between the remaining trials. Subjects were instructed to avoid food and drink in the hour before testing and to avoid strenuous exercise and caffeine consumption 24 hours before each trial. Heart rate and ratings of perceived exertion (RPE) were recorded through each multiple sprint trial to provide an indication of physiological and psychologic strain, respectively.


Twenty male exercise science students volunteered for the study, which was approved by St. Mary's University College Ethics Committee and the Institutional Review Board (for the Use of Human Subjects) of East Stroudsburg University. Before testing, subjects received written and verbal instructions regarding the nature of the investigation and completed a training history questionnaire, which indicated that all had been actively involved in sport for approximately 14 years and that most (n = 16) regularly participated in some form of multiple sprint sport. Mean times spent training and competing each week were reported as 8.9 ± 4.1 hours and 8.0 ± 4.9 hours, respectively. Before starting, all subjects completed a health-screening questionnaire and provided written informed consent. Means ± SD for age, height, body mass, body fat (3), and O2max of the subjects were 21 ± 2 years, 1.79 ± 0.09 m, 83.7 ± 10.8 kg, 16.6 ± 3.9%, and 52.7 ± 7.2 ml·kg−1·min−1, respectively.


Before each multiple sprint test, subjects performed a standardized warm-up (approximately 5 min) comprising 400 m of jogging (self-selected pace), a series of sprint drills (3 × 10 m each of high-knees, heel-flicks, and walking lunges), and 3 practice sprints. After the warm-up, subjects were given 5 minutes to stretch and prepare themselves for the multiple sprint test. Each sprint was initiated from a line 30 cm behind the start line (to prevent false triggering of the first timing gate), and all sprint and recovery times were recorded electronically by way of twin-beam photocells (Swift Performance Equipment, Lismore, Australia) placed at each end of the 30-m runway. This equipment has been shown to have very good test-retest reliability (CV = 1.51%; intraclass correlation coefficient [ICC] = 0.91) (7). Alternate sprints were performed in the opposite direction to enable subjects to maximize the passive recovery time between sprints. Before the start of each trial, subjects were instructed to perform each sprint with maximal effort and to allow sufficient recovery time between sprints to enable performance to be maintained such that sprint 12 was as fast as sprint 1. All timepieces were removed from the testing environment so that subjects had no external reference of recovery time. Heart rate was monitored throughout each trial (Polar Accurex Plus: Polar Electro Oy, Kempele, Finland), with RPE recorded after every 3 sprints using a 15-point scale (2). Fatigue during each trial was calculated from 30-m sprint times using the percentage decrement calculation (5):

Fatigue = (100 × (total sprint time ÷ ideal sprint time)) - 100,

where total sprint time = sum of sprint times from all sprints, and ideal time = number of sprints × fastest sprint time.

After the multiple sprint trials subjects completed the graded exercise test, which began with a 5-minute warm-up at 8 km·h−1 on a 1% gradient. After a further 5-minute rest period, the test began, again on a 1% gradient and at a speed estimated to achieve exhaustion in 8 to 15 minutes. Every minute during the test, the treadmill gradient was increased by 1% until subjects reached volitional exhaustion. During the tests, respiratory gases were analysed breath-by-breath using an online gas analyzer (TrueOne 2400, Parvomedics, Sandy, UT, USA), which was calibrated before every test in accordance with the manufacturer's instructions. The O2max was determined as the highest 30-second average O2 observed during the test provided that at least 2 of the following criteria had been met:

  • ▪ A plateau in O2, as determined by an increase of less than 2 ml·kg−1·min−1 over the previous stage.
  • ▪ An repiratory exchange ratio (RER) of 1.15 or greater.
  • ▪ A heart rate within 10 b·min−1 of age predicted maximum.
  • ▪ An RPE of 19 or greater.

Statistical Analyses

All statistical analyses were conducted using the Statistical Package for the Social Sciences (SPSS for Windows, SPSS, Inc., Chicago, IL, USA). Measures of centrality and spread are presented as means ± SD. The process of familiarization was examined in 3 ways. First, because recovery time was only relevant if subjects were able to maintain sprint performance, the ability of subjects to achieve this goal was quantified by the attainment of 2 criteria:

  1. The absence of an obvious pattern of fatigue.
  2. A within-trial CV of 2.02% or less (the upper confidence limit of the CV of fastest sprint time in this type of exercise) (7).

In effect, if subjects were not able to achieve the above criteria in the early trials, this was considered evidence of learning effects.

Second, learning effects were evaluated from between-trial differences in mean recovery time assessed by way of a one-way analysis of variance (ANOVA). Third, learning effects were evaluated by examining changes in the reliability of mean recovery time between consecutive pairs of trials using a two-way ANOVA as described by Schabort et al. (13), with mean recovery time as the dependent variable in each model, subject number included as a random effect, and trial number as a fixed effect.

After determining the number of trials required to limit the effects of familiarization and after eliminating those subjects who had failed to maintain sprinting performance on every trial, reliability was evaluated across the remaining trials, again using a two-way ANOVA, with measures of reliability determined as CV and ICC. The 95% confidence limits for CV and ICC were calculated using chi square and McGraw and Wong (12) estimates, respectively.

The pattern of the recovery times and the RPE responses was investigated by performing a one-way repeated measures ANOVA on trial 4, with a Pearson correlation used to investigate the relationship between O2max and mean recovery time. Significant main effects for all ANOVA were followed up using Bonferonni adjustments. The α was set at 5% for all analyses.



The number of subjects failing to meet the absence of fatigue criteria during trials 1 to 4 were 9, 10, 5, and 3, respectively. However, despite the trend for an increase in mean self-selected recovery time between trials 1 and 2 and a subsequent decrease between trials 2 and 3 (Table 1), the difference in mean recovery time between trials was not statistically significant (F(2.07,39.35) = 2.101; p = 0.134). Nevertheless, between-trial test-retest reliability revealed better reliability between trials 2 and 3 and 3 and 4 than between trials 1 and 2 (Table 2). To err on the side of caution, familiarization effects were considered evident across the first 2 trials, and therefore reliability of mean recovery time was evaluated across trials 3 to 4 after excluding those subjects (n = 5) who had failed to meet the required inclusion criteria in trial 3.

Table 1:
Mean sprint times, recovery times, and fatigue data for 12 × 30 m sprints repeated at self-selected recovery periods (n = 20).
Table 2:
Reliability of mean recovery time for 12 × 30 m sprints repeated at self-selected recovery periods (n = 20).*,†


Test-retest reliability of mean recovery time across trials 3 to 4 (n = 15) revealed a good level of reliability as evidenced by a CV of 11.1% (95% likely range: 8.0-18.1%) and an ICC of 0.76 (95% likely range: 0.40-0.91).

Pattern of Recovery

The pattern of the self-selected recovery times is presented in Figure 1. Analysis of the recovery data revealed a significant effect of time (F(4.13,66.12) = 8.405; p < 0.001), with post hoc comparisons revealing significant differences only in those contrasts involving recovery times from the first 2 sprints. Analysis of within-subject recovery time revealed a mean CV of 17.0 ± 6.1% when considering recovery data from all 12 sprints, which reduced to 14.1 ± 6.0% when the first 2 recovery times were excluded from the analysis. The correlation between relative O2max and mean recovery time was 0.14 (95% likely range: −0.37-0.58).

Figure 1:
Self-selected recovery times from trial 4 of a 12 × 30 m multiple sprint running protocol (n = 17). Values are means; bars are SDs. *Significantly different from remaining data.

RPE Responses

The pattern of perceived exertion throughout the multiple sprint protocol is presented in Figure 2. Despite no decline in multiple sprint performance across the trial, analysis of the RPE data revealed a significant effect of time (F(1.26,20.19) = 65.646; p < 0.001), with post hoc tests revealing significant (p < 0.001) differences between all contrasts.

Figure 2:
Ratings of perceived exertion during a 12 × 30 m multiple sprint running protocol (n = 17) using self-selected recovery periods. Values are means; bars are SDs. Note: all means are significantly (p < 0.05) different from each other.

Heart Rate Responses

The pattern of the heart rate response to the multiple sprint protocol is presented in Figure 3. Maximum heart rate during the multiple sprint tests was 171.2 ± 10.4 b·min−1, with mean heart rate recovery between sprints being 27.1 ± 9.2 b·min−1. Maximum heart rate during the O2max tests was 193.1 ± 8.5 b·min−1.

Figure 3:
Heart rate response during a 12 × 30 m multiple sprint running protocol (n = 17) using self-selected recovery periods. Solid line represents mean heart rate response; dashed lines represent SDs. Note: recovery heart rate data are presented as percentage of total test time to allow direct comparisons between subjects.


The aims of the present study were to investigate the process of self-selected recovery in a multiple sprint test with a view to using self-selected recovery time as a means of reliably quantifying an individual's ability to resist fatigue in this type of exercise. Despite having no external reference of elapsed time, the results showed that after the completion of 2 familiarization trials, participants were able to maintain sprint performance with a relatively short and consistent recovery. These findings compare well with those of Glaister et al. (8), which showed a relatively small amount of fatigue in individuals performing 12 × 30 m sprints repeated at 65-second intervals, particularly when compared with the same protocol repeated at 35-second intervals. Nevertheless, despite the absence of fatigue in the present study, RPE values suggest that although individuals believed they had recovered sufficiently to enable sprint performance to be maintained, they were progressively finding the test more difficult. Given that subjects were instructed to give themselves sufficient recovery time to enable sprint performance to be maintained, it is difficult to elucidate the reasons for this response or to speculate on what would have happened to performance if the number of sprints had been extended. It is, however, possible that the steady increase in RPE reflected the fact that subjects were only just giving themselves sufficient recovery time based upon the fact that they knew the number of sprints they were required to perform. As such, increasing or decreasing the number of sprints may have resulted in the same RPE response across the trial.

The energetics of a sprint as short as that performed in the present study are reported to be fuelled primarily by PCr degradation and anaerobic glycolysis, with the former providing the larger (approximately 60%) contribution (6,14). As sprints are repeated, the ability to maintain performance is determined by the ability to return to homeostasis during the intervening recovery periods. Because PCr off-kinetics follow a biexponential pattern of resynthesis with peak resynthesis rates of approximately 1.3 mmol·kg dry muscle−1·s−1 (15), it would appear that the recovery periods chosen by the subjects in the present study would have been sufficient to allow PCr to continue to make the same contribution to adenosine triphosphate (ATP) provision throughout each sprint. Moreover, with such a relatively short time course, anaerobic glycolysis would not be impaired by glycogen availability, although a corresponding increase in acidosis may have impaired the rate of ATP provision. Unfortunately, because neither muscle nor blood pH levels were evaluated in this investigation, this latter point remains speculative and is an issue requiring further investigation. However, the idea of a progressive increase in acidosis is plausible given the magnitude of the glycolytic contribution to each sprint and the much slower rate of intramuscular pH recovery relative to that of PCr (9). Indeed, an increase in acidosis, along with a number of other mediating factors (such as muscle damage), also provide a possible explanation for the progressive increase in RPE observed throughout each test.

The duration of the recovery periods chosen by the subjects had a much more distinct effect on heart rate than the fixed, and considerably shorter, recovery periods used in previous research (9). In fact, in multiple sprint tests with 10-second recovery periods, the recovery of heart rate between sprints has been shown to be barely identifiable (9). However, despite individual differences in the duration of the recovery periods, the correlation between recovery duration and O2max was poor. In effect, although the rapid phase of postexercise recovery is fuelled by aerobic metabolism (ATP and PCr resynthesis and restoration of muscle and blood oxygen stores), those individuals with the greatest capacity to use oxygen did not typically choose the shortest recovery periods. One of the main limitations with this study is that the duration of the self-selected recovery periods was likely to have, in-part, been influenced by the sprinting ability of each subject. In effect, those subjects with the fastest sprint times were likely to have encountered the largest amount of physiologic/metabolic strain and as such required a longer recovery time compared with their less anaerobic counterparts, regardless of their level of aerobic fitness. Although this argument cannot be substantiated from the data collected in this investigation, it is a confounding factor that may explain why the correlation between recovery duration and O2max was lower than anticipated.

Practical Applications

The results of the present study show that after the completion of 2 familiarization trials, the ability to maintain sprinting performance in a series of repeated sprints can be self-regulated by an athlete to a high degree of accuracy without the need for external timepieces. These findings have 2 practical applications. First, for those athletes involved in multiple sprint sports, the use of self-selected recovery periods provides an alternative and reliable approach to quantifying their ability to recover between sprints, and thereby resist fatigue, in this type of activity. Second, if the goal of a sprint training session is to maintain quality, the use of self-selected recovery periods provides coaches, who would otherwise use fixed recovery periods, with a way of maintaining that quality tailored to the abilities of each athlete.


The authors express their gratitude to Dr. Shala Davis and Dr. Gregory Dwyer for their help in the data collection for this investigation.


1. Bishop, D and Spencer, M. Determinants of repeated-sprint ability in well-trained team-sport athletes and endurance-trained athletes. J Sports Med Phys Fitness 44: 1-7, 2004.
2. Borg, G. Perceived exertion as an indicator of somatic stress. Scand J Rehab Med 2: 92-98, 1970.
3. Durnin, JV and Womersley, J. Body fat assessed from total body density and its estimation from skinfold thickness: measurements on 481 men and women aged from 16 to 72 years. Br J Nutr 32: 77-97, 1974.
4. Edge, J, Bishop, D, Goodman, C, and Dawson, B. Effects of high- and moderate-intensity training on metabolism and repeated sprints. Med Sci Sports Exerc 37: 1975-1982, 2005.
5. Fitzsimmons, M, Dawson, B, Ware, D, and Wilkinson, A. Cycling and running tests of repeated sprint ability. Aust J Sci Med Sport 25: 82-87, 1993.
6. Glaister, M. Multiple sprint work: physiological responses, mechanisms of fatigue and the influence of aerobic fitness. Sports Med 35: 757-777, 2005.
7. Glaister, M, Howatson, G, Lockey, RA, Abraham, C, Goodwin, J, and McInnes, G. Familiarisation and reliability of multiple sprint running performance indices. J Strength Cond Res 21: 857-859, 2007.
8. Glaister, M, Howatson, G, Pattison, JR, and McInnes, G. The reliability and validity of fatigue measures during multiple sprint work: an issue revisited. J Strength Cond Res 22: 1597-1601, 2008.
9. Glaister, M, Stone, MH, Stewart, AM, Hughes, M, and Moir, GL. The influence of recovery duration on multiple sprint cycling performance. J Strength Cond Res 19: 831-837, 2005.
10. Glaister, M, Stone, MH, Stewart, AM, Hughes, M, and Moir, GL. The Influence of Endurance Training on Multiple Sprint Cycling Performance. J Strength Cond Res 21: 606-612, 2007.
11. Hamilton, AL, Nevill, ME, Brooks, S, and Williams, C. Physiological responses to maximal intermittent exercise: differences between endurance-trained runners and games players. J Sports Sci 9: 371-382, 1991.
12. McGraw, KO and Wong, SP. Forming inferences about some intraclass correlation coefficients. Psychol Methods 1: 30-46, 1996.
13. Schabort, EJ, Hawley, JA, Hopkins, WG, and Blum, H. High reliability of performance of well-trained rowers on a rowing ergometer. J Sports Sci 17: 627-632, 1999.
14. Spencer, M, Bishop, D, Dawson, B, and Goodman, C. Physiological and metabolic responses of repeated-sprint activities: specific to field-based team sports. Sports Med 35: 1025-1044, 2005.
15. Walter, G, Vandenborne, K, McCully, KK, and Leigh, JS. Noninvasive measurement of phosphocreatine recovery kinetics in single human muscles. Am J Physiol 272: C525-C534, 1997.

repeated sprint ability; intermittent; perceived recovery; multiple sprint work

© 2010 National Strength and Conditioning Association