Introduction

Biathlon is an Olympic winter sport that combines rifle shooting with cross-country (XC) skiing in the skate technique. Final performance in biathlon is determined from both skiing and shooting components [12], where skiing speed alone can explain between approximately 50% to as much as 84% of the variation in the final biathlon performance, depending on the specific event (i.e., sprint, individual, mass-start, pursuit) [13,14,15,16].

Biathlon competition is performed over 6–20 km of XC skiing divided into 3 or 5 laps (depending on the specific event) interspersed with shooting mixed between the prone and standing positions [9]. During the XC skiing component of biathlon, athletes must also carry a rifle with a minimum mass of 3.5 kg harnessed on their back which has been shown to increase the physiological demands during both roller-skiing on a treadmill [9, 21, 24] and also during on-snow skiing [23]. Further, it has also been observed that rifle carriage influences skiing biomechanics [10, 24]. For example, when roller-skiing on a treadmill whilst carrying the biathlon rifle, skiing cycle time, cycle length and cycle forces were reduced, whilst cycle frequency was increased [24].

A common method of performance analysis in biathlon and XC skiing is to analyse pacing strategies [1,2,3,4, 25]. For example, the lap-to-lap pacing strategies in both XC skiing and biathlon are relatively well understood [1,2,3,4, 25]. Further, more recent research has utilised statistical parametric mapping (SPM) as a way to analyse micro-pacing strategies (i.e., within-lap pacing) in both a classic sprint [7] and a freestyle distance [22] XC ski race. SPM is a novel statistical technique which has been promoted for analysis of smooth biomechanical data [19, 20]. These studies [7, 22] identified specific track sections that were particularly crucial for successful race performance. For example, in a men’s distance XC ski race the fastest man gained more than 20 s on the slowest man in a 400 m section of an uphill [22]. Analyses such as these provide crucial insights for athletes and coaches to guide training programmes and pacing strategies during competitions.

Despite this, SPM analyses have not yet been used to identify micro-pacing strategies in biathlon. Although biathletes also use the skating technique, rifle carriage affects both physiological and biomechanical aspects of skiing which makes it different to XC skiing. Due to this, biathletes might adopt different micro-pacing strategies. However, it remains unclear how biathletes use micro-pacing and how rifle carriage might affect micro-pacing strategies. A greater understanding of how biathletes use micro-pacing and the effect of rifle carriage on these strategies can assist coaches and athletes in developing race performance plans and training programmes. Accordingly, this study aimed to examine the micro-pacing strategies during biathlon skiing with and without the biathlon rifle.

Methods

Participants

Twenty biathletes (7 females: Age: 19 ± 2 years, Stature: 167 ± 7 cm, Body mass: 71 ± 10 kg; 13 males: Age: 19 ± 3 years; Stature: 180 ± 6 cm; Body mass: 72 ± 5 kg) participated in this study. The biathletes were tier 3 athletes representing provincial/state academy programmes [17]. The regional ethical review board in (The regional ethical review board in Umeå, Sweden (registration number: 2016-506- 31M) preapproved the research techniques and experimental protocol. All participants provided written informed consent and agreed to participate in this study. All research was conducted in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki).

Design

Biathletes completed a standardised 15-min warm-up prior to the time-trials which involved familiarisation of the time-trial course, dynamic bodily movements and some high intensity skiing actions. Thereafter, participants completed two maximal effort XC skiing time-trials using the skating technique, completed both with and without the biathlon rifle, in a randomised-counterbalanced order. Participants were permitted at least 20-min of self-regulated active recovery between time-trial efforts, which previous research has demonstrated is sufficient for blood lactate clearance [18, 26]. During the time-trials, athletes wore a GNSS sensor (1 Hz; Forerunner 920; Garmin, Olathe, KS, USA) in order to record positioning and skiing speed which was used in order to analyse micro-pacing strategy.

Skiing Course

The biathlon skiing time-trials were performed at the Swedish National Biathlon Arena on a course similar to that which is regularly included in the IBU world cup event. The circuit was approximately 2300 m in distance with a total climb of 95 m. The course was divided into discrete uphill (U1, U2, U3, U4, U5, U6, U7), downhill (D1, D2, D3, D4, D5, D6, D7) and flat sections (F1, F2) for analyses, by visual inspection of the altitude profile from the positioning system worn by athletes.

Data Analyses

The x, y positioning coordinates of all GNSS data were interpolated to 10 Hz using a cubic spline and filtered using a second order low-pass Butterworth filter with a cut-off frequency of 4 Hz. This filter frequency was selected in order to remove any undesired noise from the positioning signal and is similar to cut-off filter frequencies previously used with positioning system data [8]. For the fastest athlete (time-trial performance), the interpolated and filtered GNSS coordinates were subsequently resampled to every 1 m integer also using a cubic spline in order to be used as a reference trajectory.

The GNSS coordinates of all other athletes were corrected using a trajectory correction method, which has been described in depth previously [5, 7, 22]. Briefly, the moment with the smallest Euclidean distance between the measured and reference trajectory was identified and the instantaneous skiing speed at the corresponding track location was calculated before being smoothed using a moving mean over 30 m. This permitted comparison of instantaneous skiing speed between athletes at the same 1 m integer along the time-trial course.

Statistical Analyses

The instantaneous speed curves (1-dimensional; 1D data) were analysed using a SPM procedure using open-source SPM 1D software [19, 20] in MATLAB R2020b (The MathWorks, Inc., Natick, Massachusetts, USA). For each section of the time-trial course, SPM 1D one-tailed linear regression models were applied to investigate the relationships between instantaneous speed and section time-trial time. This resulted in SPM{t} curves with a critical threshold set at α = 0.05. The SPM{t} curve represents the alpha value of the relationship between instantaneous speed and time-trial performance at every 1 m integer. Where the SPM{t} values exceeded the critical threshold, instantaneous speed was considered to be significantly related to time-trial performance. The course locations where the SPM{t} curve exceeded the critical threshold (i.e., the course locations where instantaneous speed and time-trial performance were significantly related) were computed. These sections are termed SPM ‘clusters’. For each cluster, the time difference between the fastest and the slowest skier was computed.

Results

Women

Figure 1 displays the skiing speed, altitude profile and the SPM{t} curve for each section of the time-trial course, along with the time gain between the fastest and slowest skier within the cluster for both No-Rifle (Panel A) and With-Rifle (Panel B).

Fig. 1
figure 1

Speed (top panel), Altitude (middle panel), and Statistical Parametric Mapping (SPM{t}) curves (bottom panel) for the women’s No-Rifle A and With-Rifle B time-trials. The shaded area on the SPM{t} curve shows the course locations where a significant relationship existed between time-trial performance and instantaneous speed. The red dotted line represents the threshold value to meet statistical significance. TG time gain between the fastest and slowest skier within that cluster, U Uphill, D Downhill, F Flat

Full size image

No-Rifle

The mean time-trial duration was 493 ± 71 s. Time-trial performance was significantly related to instantaneous skiing speed during clusters in D1, U2, U4 (two clusters), D4, D6 and U7. The cluster with the largest time gain was observed in D6, where the fastest skier gained 3.9 s over the slowest skier in a cluster of 26 m distance. In total, the fastest skier gained 16.1 s on the slowest skier throughout all clusters combined.

With-Rifle

The mean time-trial duration was 508 ± 91 s. Time-trial performance was significantly related to instantaneous skiing speed during clusters in F1, D1, U2, U3 (two clusters), D3, U4, U5, U6 (two clusters), D6 and U7. The cluster with the largest time gain was observed in D6, where the fastest skier gained 4.1 s over the slowest skier in a cluster of 62 m distance. In total, the fastest skier gained 25.8 s on the slowest skier throughout all clusters combined.

Men

Figure 2 displays the skiing speed, altitude profile and the SPM{t} curve for each section of the time-trial course, along with the time gain between the fastest and slowest skier within the cluster for both No-Rifle (Panel A) and With-Rifle (Panel B).

Fig. 2
figure 2

Speed (top panel), Altitude (middle panel), and Statistical Parametric Mapping (SPM{t}) curves (bottom panel) for the men’s No-Rifle A and With-Rifle B time-trials. The shaded area on the SPM{t} curve shows the course locations where a significant relationship existed between time-trial performance and instantaneous speed. The red dotted line represents the threshold value to meet statistical significance. TG time gain between the fastest and slowest skier within that cluster, U Uphill, D Downhill, F Flat

Full size image

No-Rifle

The mean time-trial duration was 371 ± 22 s. Time-trial performance was significantly related to instantaneous skiing speed during clusters in D1, U2 (three clusters), D2 (two clusters), U3 (two clusters), D3, U4 and F2 (two clusters). The cluster with the largest time gain was observed in U2, where the fastest skier gained 2.7 s over the slowest skier in a cluster of 26 m distance. In total, the fastest skier gained 21.9 s on the slowest skier throughout all clusters combined.

With-Rifle

The mean time-trial duration was 380 ± 24 s. Time-trial performance was significantly related to instantaneous skiing speed during clusters in U1, U2, D2, U3, D3, U4, D4, U5, D5, D6 and U7. The cluster with the largest time gain was observed in D6, where the fastest skier gained 2.9 s over the slowest skier in a cluster of 12 m distance. In total, the fastest skier gained 18.9 s on the slowest skier throughout all clusters combined.

Figure 3 displays the latitude and longitude positions of track sections U6 and D6, along with the alpha value and distance of all clusters within the sections and the corresponding time gain between the fastest and slowest skier. These track sections are unique because they are separated by a 180° turn. Further, for the With-Rifle condition, section D6 contained the cluster with the largest time gain between the fastest and slowest skier for both men and women (2.9 s and 4.1 s, respectively). In addition, section D6 contained the cluster with the largest time gain between the fastest and slowest skier for the women’s No-Rifle trial (3.9 s).

Fig. 3
figure 3

Latitude and longitude positions of track sections U6 and D6 for both With-Rifle (left panel) and No-Rifle (right panel), along with the alpha value and distance of all clusters within the sections and the corresponding time gain between the fastest and slowest skier. Red shaded areas represent the track location of SPM clusters for the women. Blue shaded areas represent the track location of SPM clusters for the men. Arrows indicate direction of travel. Dist Distance, TG time gain between the fastest and slowest skier within that cluster

Full size image

Discussion

This study is the first to use SPM analyses to investigate micro-pacing strategies in biathlon skiing. This study provides a unique insight into the effects of the rifle carriage on skiing performance in biathlon. The key findings of this study were: (1) the fastest biathletes skied with greater instantaneous speeds in specific clusters of the course, which included both uphill and downhill sections; (2) while carrying the rifle, there were additional clusters mostly in uphill sections where there were significant relationships between instantaneous skiing speed and time-trial performance, suggesting that rifle carriage particularly impacts skiing performance during uphill sections; (3) the largest time gains between the fastest and slowest biathletes were observed in a particular downhill section (D6) that was preceded by a 180° turn; (4) statistical parametric mapping can be used in biathlon to providing pacing and performance feedback to athletes and coaches.

Previous research has primarily focused on lap-to-lap pacing strategies in biathlon [4, 14,15,16]. These insights are indeed important for biathletes and their coaches. However, these studies provide little insight into how a biathlete should focus their efforts within each lap and which sections of the course are particularly crucial for faster ski times (i.e. micro-pacing). In line with previous research in both sprint [7] and distance [22] XC skiing races, the present study identified specific track sections where instantaneous speed was significantly related to time-trial performance during biathlon skiing with and without the rifle. These were primarily uphill sections. However, in contrast to previous research [22], the present study also identified some downhill sections with clusters where instantaneous skiing speed and time-trial performance were significantly related. These differences might reflect different micro-pacing strategies adopted by biathletes compared to XC skiers. Although the two sports are similar, previous studies have shown that biathletes adopt different lap-to-lap pacing strategies compared to XC skiers, likely due to the additional demand of precision shooting for successful performance [14,15,16]. Therefore, it is likely that biathletes also adopt different micro-pacing strategies. However, this study has observed biathletes only, so a direct comparison between the micro-pacing strategies of XC skiers and biathletes under the same conditions is not possible. In addition, cross-sectional comparisons of the micro-pacing strategy between the present study and the aforementioned research in XC skiing [7, 22], must consider the level of the athlete. While the present study has observed micro-pacing strategies in tier 3 biathletes, the previous research studies investigating micro-pacing in XC skiing examined professional FIS-sanctioned XC ski races, which likely represents tier 4 athletes [17].

Another finding of the present study is that during rifle carriage there were additional clusters during uphill sections with significant relationships between instantaneous skiing speed and time-trial performance. This suggests a greater inter-athlete variability in instantaneous skiing speed during uphill sections with rifle carriage. For example, for women during the no-rifle condition, there were clusters with significant relationships between instantaneous skiing speed and time-trial performance in uphill sections U2, U4 and U7. However, with-rifle there were clusters in these uphill sections as well as additional clusters in U3, U5 and U6. Moreover, for men during the no-rifle condition there were clusters with significant relationships between instantaneous skiing speed and time-trial performance in uphill sections U2, U3 and U4, but with-rifle, there were clusters in these uphill sections as well as additional clusters in U1 and U5. This is likely due to the added mass of the rifle which has a greater impact on skiing performance during uphill sections. This is probably because of the additional force (and therefore work) required to move a mass uphill [6, 11]. All taken together, this stresses the importance of uphill sections during biathlon skiing, which is in line with previous findings [15] that also demonstrated that uphill sections are particularly crucial to successful performance. Accordingly, in order to improve the XC skiing component of biathlon, it seems that coaches and athletes should focus their efforts on improving uphill skiing performance.

Despite this, the largest time gains between the fastest and slowest skiers was most often observed in downhill Sect. 6 (D6). Section D6 is particularly unique because it is preceded by a 180° turn (Fig. 3) at the top of an uphill section (U6). This 180° turn at the top of U6 also coincides with some of the slowest skiing speeds observed over the course (Figs. 1 and 2). Accordingly, it is likely that these large time differences observed in section D6 are related to the speed an athlete is able to carry through the 180° turn. Major speed reductions through the turn are likely having a major impact on the speeds through the subsequent downhill. This hypothesis is in line with previous findings in a sprint XC ski race, where faster times recorded on the downhill sections were associated with greater instantaneous speeds during the acceleration phase at the top of the downhill, as well as during the downhill turns [7].

This study also demonstrates that SPM analyses can provide practical information to optimise pacing strategies in biathlon. Coaches and sports scientists could use SPM analyses to identify the critical components for success on different courses and race formats to guide pacing strategies in advance of a competition. In addition, the strengths and weaknesses of individual athletes can be identified and used to guide training programmes.

A limitation of this study is that the study protocol did not include a shooting component. It is understood that biathletes adjust their pacing according to the shooting component to limit the physical/physiological exertion in order to maintain shooting performance. In demonstration of this, elite biathletes tend to use a J-shaped pacing strategy, where ski times increase on the final lap when there is no shooting to consider [13, 15]. It is likely that the biathletes in the current study used a different pacing strategy knowing that no shooting was required at the end of the time-trial. Furthermore, the time-trials in this study were completed over different testing days. Changes in weather and snow conditions are known to effect ski waxes and skiing speeds. Accordingly, this would also influence the results of the current study. Nonetheless, this study provides a unique insight into the effects of the biathlon rifle on biathlon skiing performance and presents a method with the ability to analyse pacing strategies with greater precision and detail. Future research should consider applying these methods in actual biathlon competitions or time-trials performed in the same weather conditions and also include shooting performance.

Conclusion

Specific clusters were identified where instantaneous skiing speed was significantly related to time-trial performance in biathlon skiing. These clusters might represent micro-pacing strategies in which the fastest athletes exerted more effort to maintain higher speeds. In these sections, the fastest skiers were able to gain significant time on the slowest skiers. Both uphill and downhill sections were identified, suggesting that the fastest skiers were able to separate themselves from the slowest skiers in both uphill and downhill sections. In biathlon skiing with the rifle, there were more clusters of significant relationships between instantaneous skiing speed and time-trial performance, particularly in the uphill sections, for both men and women, indicating the importance of uphill sections to biathlon skiing performance. Finally, SPM analyses are useful to analyse micro-pacing strategies in biathlon skiing and could be used to guide race strategies and training programmes.