This paper explores the notion that the availability and analysis of large data sets have the capacity to improve practice and change the nature of science in the sport and exercise setting. The increasing use of data and information technology in sport is giving rise to this change. Web sites hold large data repositories, and the development of wearable technology, mobile phone applications, and related instruments for monitoring physical activity, training, and competition provide large data sets of extensive and detailed measurements. Innovative approaches conceived to more fully exploit these large data sets could provide a basis for more objective evaluation of coaching strategies and new approaches to how science is conducted. An emerging discipline, sports analytics, could help overcome some of the challenges involved in obtaining knowledge and wisdom from these large data sets. Examples of where large data sets have been analyzed, to evaluate the career development of elite cyclists and to characterize and optimize the training load of well-trained runners, are discussed. Careful verification of large data sets is time consuming and imperative before useful conclusions can be drawn. Consequently, it is recommended that prospective studies be preferred over retrospective analyses of data. It is concluded that rigorous analysis of large data sets could enhance our knowledge in the sport and exercise sciences, inform competitive strategies, and allow innovative new research and findings.
Louis Passfield and James G. Hopker
Richard Ebreo, Louis Passfield, and James Hopker
Purpose: To evaluate the reliability of calculating gross efficiency (GE) conventionally and using a back extrapolation (BE) method during high-intensity exercise (HIE). Methods: A total of 12 trained participants completed 2 HIE bouts (P1 = 4 min at 80% maximal aerobic power [MAP]; P2 = 4 min at 100%MAP). GE was calculated conventionally in the last 3 minutes of submaximal (50%MAP) cycling bouts performed before and after HIE (Pre50%MAP and Post50%MAP). To calculate GE using BE (BGE), a linear regression of GE submaximal values post-HIE were back extrapolated to the end of the HIE bout. Results: BGE was significantly correlated with Post50%MAP GE in P1 (r = .63; P = .01) and in P2 (r = .85; P = .002). Reliability data for P1 and P2 BGE demonstrate a mean coefficient of variation of 7.8% and 9.8% with limits of agreement of 4.3% and 4.5% in relative GE units, respectively. P2 BGE was significantly lower than P2 Post50%MAP GE (18.1% [1.6%] vs 20.3% [1.7%]; P = .01). Using a declining GE from the BE method, there was a 44% greater anaerobic contribution compared with assuming a constant GE during 4-minute HIE at 100%MAP. Conclusion: HIE acutely reduced BGE at 100%MAP. A greater anaerobic contribution to exercise as well as excess postexercise oxygen consumption at 100%MAP may contribute to this decline in efficiency. The BE method may be a reliable and valid tool in both estimating GE during HIE and calculating aerobic and anaerobic contributions.
Antonis Kesisoglou, Andrea Nicolò, and Louis Passfield
Purpose: To examine the effect of cycling exercise intensity and duration on subsequent performance and to compare the resulting acute performance decrement (APD) with total work done (TWD) and corresponding training-load (TL) metrics. Methods: A total of 14 male cyclists performed a 5-minute time trial (TT) as a baseline and after 4 initial exercise bouts of varying exercise intensity and duration. The initial exercise bouts were performed in a random order and consisted of a 5- and a 20-minute TT and a 20- and a 40-minute submaximal ride. The resulting APD was calculated as the percentage change in 5-minute TT from baseline, and this was compared with the TWD and TL metrics for the corresponding initial exercise bout. Results: Average power output was different for each of the 4 initial exercise bouts (
Erin Calaine Inglis, Danilo Iannetta, Louis Passfield, and Juan M. Murias
Purpose: To (1) compare the power output (PO) for both the 20-minute functional threshold power (FTP20) field test and the calculated 95% (FTP95%) with PO at maximal lactate steady state (MLSS) and (2) evaluate the sensitivity of FTP95% and MLSS to training-induced changes. Methods: Eighteen participants (12 males: 37  y and 6 females: 28  y) performed a ramp-incremental cycling test to exhaustion, 2 to 3 constant-load MLSS trials, and an FTP20 test. A total of 10 participants returned to repeat the test series after 7 months of training. Results: The PO at FTP20 and FTP95% was greater than that at MLSS (P = .00), with the PO at MLSS representing 88.5% (4.8%) and 93.1% (5.1%) of FTP and FTP95%, respectively. MLSS was greater at POST compared with PRE training (12  W) (P = .002). No increase was observed in mean PO at FTP20 and FTP95% (P = .75). Conclusions: The results indicate that the PO at FTP95% is different to MLSS, and that changes in the PO at MLSS after training were not reflected by FTP95%. Even when using an adjusted percentage (ie, 88% rather than 95% of FTP20), the large variability in the data is such that it would not be advisable to use this as a representation of MLSS.
Andy Galbraith, James Hopker, Stephen Lelliott, Louise Diddams, and Louis Passfield
To compare critical speed (CS) measured from a single-visit field test of the distance–time relationship with the “traditional” treadmill time-to-exhaustion multivisit protocol.
Ten male distance runners completed treadmill and field tests to calculate CS and the maximum distance performed above CS (D′). The field test involved 3 runs on a single visit to an outdoor athletics track over 3600, 2400, and 1200 m. Two field-test protocols were evaluated using either a 30-min recovery or a 60-min recovery between runs. The treadmill test involved runs to exhaustion at 100%, 105%, and 110% of velocity at VO2max, with 24 h recovery between runs.
There was no difference in CS measured with the treadmill and 30-min- and 60-minrecovery field tests (P < .05). CS from the treadmill test was highly correlated with CS from the 30- and 60-min-recovery field tests (r = .89, r = .82; P < .05). However there was a difference and no correlation in D′ between the treadmill test and the 30 and 60-min-recovery field tests (r = .13; r = .33, P > .05). A typical error of the estimate of 0.14 m/s (95% confidence limits 0.09–0.26 m/s) was seen for CS and 88 m (95% confidence limits 60–169 m) for D′. A coefficient of variation of 0.4% (95% confidence limits: 0.3–0.8%) was found for repeat tests of CS and 13% (95% confidence limits 10–27%) for D′.
The single-visit method provides a useful alternative for assessing CS in the field.
Andy Galbraith, James Hopker, Marco Cardinale, Brian Cunniffe, and Louis Passfield
To examine the training and concomitant changes in laboratory- and field-test performance of highly trained endurance runners.
Fourteen highly trained male endurance runners (mean ± SD maximal oxygen uptake [VO2max] 69.8 ± 6.3 mL · kg−1 · min−1) completed this 1-y training study commencing in April. During the study the runners undertook 5 laboratory tests of VO2max, lactate threshold (LT), and running economy and 9 field tests to determine critical speed (CS) and the modeled maximum distance performed above CS (D′). The data for different periods of the year were compared using repeated-measures ANOVA. The influence of training on laboratory- and field-test changes was analyzed by multiple regression.
Total training distance varied during the year and was lower in May–July (333 ± 206 km, P = .01) and July–August (339 ± 206 km, P = .02) than in the subsequent January–February period (474 ± 188 km). VO2max increased from the April baseline (4.7 ± 0.4 L/min) in October and January periods (5.0 ± 0.4 L/min, P ≤ .01). Other laboratory measures did not change. Runners’ CS was lowest in August (4.90 ± 0.32 m/s) and highest in February (4.99 ± 0.30 m/s, P = .02). Total training distance and the percentage of training time spent above LT velocity explained 33% of the variation in CS.
Highly trained endurance runners achieve small but significant changes in VO2max and CS in a year. Increases in training distance and time above LT velocity were related to increases in CS.
Marco Arkesteijn, Simon Jobson, James Hopker, and Louis Passfield
Previous research has shown that cycling in a standing position reduces cycling economy compared with seated cycling. It is unknown whether the cycling intensity moderates the reduction in cycling economy while standing.
The aim was to determine whether the negative effect of standing on cycling economy would be decreased at a higher intensity.
Ten cyclists cycled in 8 different conditions. Each condition was either at an intensity of 50% or 70% of maximal aerobic power at a gradient of 4% or 8% and in the seated or standing cycling position. Cycling economy and muscle activation level of 8 leg muscles were recorded.
There was an interaction between cycling intensity and position for cycling economy (P = .03), the overall activation of the leg muscles (P = .02), and the activation of the lower leg muscles (P = .05). The interaction showed decreased cycling economy when standing compared with seated cycling, but the difference was reduced at higher intensity. The overall activation of the leg muscles and the lower leg muscles, respectively, increased and decreased, but the differences between standing and seated cycling were reduced at higher intensity.
Cycling economy was lower during standing cycling than seated cycling, but the difference in economy diminishes when cycling intensity increases. Activation of the lower leg muscles did not explain the lower cycling economy while standing. The increased overall activation, therefore, suggests that increased activation of the upper leg muscles explains part of the lower cycling economy while standing.
Antonis Kesisoglou, Andrea Nicolò, Lucinda Howland, and Louis Passfield
Purpose: To examine the effect of continuous (CON) and intermittent (INT) running training sessions of different durations and intensities on subsequent performance and calculated training load (TL). Methods: Runners (N = 11) performed a 1500-m time trial as a baseline and after completing 4 different running training sessions. The training sessions were performed in a randomized order and were either maximal for 10 minutes (10CON and 10INT) or submaximal for 25 minutes (25CON and 25INT). An acute performance decrement (APD) was calculated as the percentage change in 1500-m time-trial speed measured after training compared with baseline. The pattern of APD response was compared with that for several TL metrics (bTRIMP, eTRIMP, iTRIMP, running training stress score, and session rating of perceived exertion) for the respective training sessions. Results: Average speed (P < .001,
Ciaran O’Grady, Louis Passfield, and James G. Hopker
Purpose: Rating of perceived exertion (RPE) as a training-intensity prescription has been extensively used by athletes and coaches. However, individual variability in the physiological response to exercise prescribed using RPE has not been investigated. Methods: Twenty well-trained competitive cyclists (male = 18, female = 2, maximum oxygen consumption =55.07 [11.06] mL·kg−1·min−1) completed 3 exercise trials each consisting of 9 randomized self-paced exercise bouts of either 1, 4, or 8 minutes at RPEs of 9, 13, and 17. Within-athlete variability (WAV) and between-athletes variability (BAV) in power and physiological responses were calculated using the coefficient of variation. Total variability was calculated as the ratio of WAV to BAV. Results: Increased RPEs were associated with higher power, heart rate, work, volume of expired oxygen (VO2), volume of expired carbon dioxide (VCO2), minute ventilation (VE), deoxyhemoglobin (ΔHHb) (P < .001), and lower tissue saturation index (ΔTSI%) and ΔO2Hb (oxyhaemoglobin; P < .001). At an RPE of 9, shorter durations resulted in lower VO2 (P < .05) and decreased ΔTSI%, and the ΔHHb increased as the duration increased (P < .05). At an RPE of 13, shorter durations resulted in lower VO2, VE, and percentage of maximum oxygen consumption (P < .001), as well as higher power, heart rate, ΔHHb (P < .001), and ΔTSI% (P < .05). At an RPE of 17, power (P < .001) and ΔTSI% (P < .05) increased as duration decreased. As intensity and duration increased, WAV and BAV in power, work, heart rate, VO2, VCO2, and VE decreased, and WAV and BAV in near-infrared spectroscopy increased. Conclusions: Self-paced intensity prescriptions of high effort and long duration result in the greatest consistency on both a within- and between-athletes basis.
Arthur H. Bossi, Ciaran O’Grady, Richard Ebreo, Louis Passfield, and James G. Hopker
Purpose : To describe pacing strategy and competitive behavior in elite-level cyclo-cross races. Methods: Data from 329 men and women competing in 5 editions (2012–2016) of Union Cycliste Internationale Cyclo-Cross World Championships were compiled. Individual mean racing speeds from each lap were normalized to the mean speeds of the whole race. Lap and overall rankings were also explored. Pacing strategy was compared between sexes and between top- and bottom-placed cyclists. Results: A significant main effect of laps was found in 8 out of 10 races (4 positive, 3 variable, 2 even, and 1 negative pacing strategies), and an interaction effect of ranking-based groups was found in 2 (2016, male and female races). Kendall tau-b correlations revealed an increasingly positive relationship between intermediate and overall rankings throughout the races. The number of overtakes during races decreased from start to finish, as suggested by significant Friedman tests. In the first lap, normalized cycling speeds were different in 3 out of 5 editions—men were faster in 1 and slower in 2 editions. In the last lap, however, normalized cycling speeds of men were lower than those of women in 4 editions. Conclusions : Elite cyclo-cross competitors adopt slightly distinct pacing strategies in each race, but positive pacing strategies are highly probable in most events, with more changes in rankings during the first laps. Sporadically, top- and bottom-placed groups might adopt different pacing strategies during either men’s or women’s races. Men and women seem to distribute their efforts differently, but this effect is of small magnitude.