Velocity-Based vs. Percentage-Based Training: Superior Effects on Acceleration and Explosive Power in High School Triple Jump Athletes

Abstract

This study compared velocity-based training (VBT) with percentage-based training (PBT) on acceleration (30-m sprint) and explosive power in high school triple jump athletes. Twelve male national-level athletes were randomized (1:1, concealed allocation; blinded assessors) to VBT (n = 6) or PBT (n = 6). Both groups completed identical lower-body resistance training three times per week for eight weeks; the VBT group additionally received real-time barbell-velocity feedback with velocity-loss (VL) based set termination (15–20%). Performance was assessed using 30-m sprint, standing long jump (SLJ), standing triple jump (STJ), and vertical jump (VJ) at pre- and post-test. Statistical analysis included baseline-adjusted ANCOVA and effect sizes (Hedges’ g). VBT improved 30-m sprint (−1.08%, d = 0.89), SLJ (+2.07%, d = 1.02), STJ (+1.64%, d = 0.63), and VJ (+6.01%, d = 1.39; all p < 0.001). PBT also improved SLJ (+1.03%, d = 0.69; p < 0.001) and showed a moderate, statistically significant within-group gain in STJ (+0.56%, d = 0.72; p = 0.001), while improvements in 30-m sprint and VJ were modest. Between-group effects favored VBT across all outcomes. These preliminary findings suggest that VBT may provide more targeted neuromuscular adaptations than PBT, particularly in explosive movements relevant to triple jump performance. However, due to the modest sample size and limited precision, the results should be interpreted with caution and confirmed in larger, adequately powered randomized trials. Nevertheless, this study offers practical insight into load prescription for youth jump athletes and represents one of the first randomized trials to directly compare VBT and PBT in this population.

1. Introduction

The triple jump is a technically demanding event consisting of a hop–step–jump sequence, where performance is determined by approach velocity, horizontal and vertical force at takeoff, and the ability to maintain rhythm and speed across phases. Particularly, the maximal velocity achieved during the approach and the minimization of speed loss in the last two strides directly influence overall performance, showing strong similarities to the long jump. Previous biomechanical analyses have consistently reported that approach velocity and takeoff mechanics are critical determinants of total distance achieved in elite jumpers [1]. Biomechanical studies have indicated that elite male triple jumpers typically achieve approach velocities of 9–10 m/s, which are highly correlated with jump distance. An optimal balance between horizontal and vertical velocity at each takeoff is essential, as excessive vertical speed disrupts phase transitions, whereas horizontal velocity loss limits performance. Thus, triple jump training should not be limited to technical repetitions but should instead integrate acceleration capacity, horizontal and vertical explosive strength, and elastic strength and stretch-shortening cycle (SSC) efficiency [2].

Therefore, strength and power development, acceleration capacity, and stretch-shortening cycle (SSC) efficiency represent key physical factors that should be systematically enhanced in triple jump athletes.

Biomechanical characteristics of the triple jump, including approach velocity, takeoff forces, and transition efficiency, have been widely identified as key determinants of performance in previous studies.

Adolescents, particularly high school athletes (16–18 years), possess distinct developmental characteristics compared to adults.

Their performance level corresponds to national-level U18 athletes, which is consistent with typical standards in competitive youth triple jump practice.

During this stage, physical growth coincides with neuromuscular adaptations, leading to rapid improvements in motor unit recruitment efficiency, rate of force development (RFD), and neuromuscular coordination. Although hypertrophic responses are relatively limited in this population, neural adaptations and skill acquisition are pronounced [3]. Recent youth resistance training studies support that these neural and coordinative gains occur even in the absence of significant hypertrophy [4].

Faigenbaum et al. reported that resistance training in youth athletes not only enhances muscular strength and motor skill proficiency but also contributes to injury prevention, improved sport performance, and long-term health outcomes. These findings underscore the importance of implementing appropriately designed resistance training programs during adolescence, as such interventions may provide a critical foundation for both athletic development and lifelong physical activity [5].

Traditionally, resistance training intensity has been prescribed using percentage-based training (PBT), in which loads are determined relative to a fixed percentage of the one-repetition maximum (1RM). While widely adopted, this method fails to account for daily fluctuations in fatigue, recovery, and readiness. Velocity-based training (VBT) has emerged as an alternative, using barbell velocity (mean or peak concentric velocity) to regulate intensity and manage fatigue via velocity loss thresholds. A core advantage of VBT is real-time velocity feedback, which not only enhances acute performance but also promotes long-term adaptations in strength and power. Multiple comparative studies have shown that VBT can achieve equal or superior gains in power and speed compared with traditional percentage-based training in trained athletes [6]. Recent meta-analyses confirm that VBT is at least as effective as, and often superior to, PBT in improving sprint, jump, and strength performance, particularly in sports requiring explosive outputs [7,8].

Given the characteristics of the triple jump, performance depends heavily on acceleration ability and explosive power. Accordingly, this study employed four key performance tests: the 30 m sprint, standing long jump (SLJ), standing triple jump (STJ), and vertical jump (VJ). These measures, respectively, reflect approach acceleration, horizontal explosive power, elastic strength under sequential jumps, and vertical explosive capacity. While VBT research has been conducted in adults and across mixed athlete populations, studies that systematically evaluate its effects on acceleration, horizontal and vertical explosive strength, and elastic strength in youth jumpers remain scarce.

Therefore, the purpose of this study was to investigate the effects of VBT compared with PBT on fundamental explosive performance indicators in high school triple jumpers. Specifically, we examined whether VBT would produce greater improvements in SLJ, STJ, and VJ, as well as reductions in 30 m sprint time, and explored its practical applicability in the training of youth jump athletes.

2. Materials and Methods

2.1. Participants

Twelve male high school triple jump athletes from the Republic of Korea, all of whom also competed in the long jump, volunteered to participate. The intervention spanned eight weeks (3 January–28 February 2024) and was conducted at Yecheon Sports Complex, Republic of Korea. All athletes were registered with the Korea Association of Athletics Federations (KAAF) in 2024 and had recent competitive experience at the National Sports Festival or other national-level championships. Baseline performance ranged from 13.35–15.09 m for the triple jump and 6.19–7.22 m for the long jump.

Recruitment. Athletes were recruited from a regional joint training camp for high-school jumpers held at Yecheon Sports Complex during the study period; all KAAF-registered attendees were screened, and consecutive volunteers meeting the eligibility criteria were enrolled.

Eligibility. Inclusion criteria were active participation in organized jump training, no musculoskeletal injury for ≥3 months, and no surgery within the previous 12 months. Athletes with neurological, cardiovascular, or orthopedic conditions affecting performance were excluded.

Allocation. Participants were randomly allocated to the VBT group (n = 6) or the PBT group (n = 6) (see Section 2.2 for randomization procedures). The groups were comparable at baseline; descriptive characteristics and between-group differences (with effect sizes) are reported in Table 1.

Table 1. Physical characteristics of the participants.

Group	n	Age (Years)	Height (cm)	Body Mass (kg)	Training Experience (Years)	Triple Jump (m)	Long Jump (m)
VBT	6	18.19 ± 1.88 SMD = +0.18	181.23 ± 4.56 SMD = +0.62	64.67 ± 5.62 SMD = +0.27	6.23 ± 0.61 SMD = +0.36	14.37 ± 0.75 SMD = +0.61	6.81 ± 0.26 SMD = +1.05
PBT	6	17.83 ± 2.01	177.85 ± 6.15	63.33 ± 4.34	5.66 ± 2.14	13.98 ± 0.51	6.51 ± 0.31

Note. SMDs (Cohen’s d) interpreted approximately as: small ≈ 0.2, moderate ≈ 0.5, large ≈ 0.8. SMD computed as (mean_VBT − mean_PBT)/SD_pooled.

The study was approved by the Institutional Review Board of Korea National Sport University (IRB No. 20231108-119) and conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from the athletes and, for minors, from their legal guardians, after a detailed explanation of the study purpose and procedures.

2.2. Study Design and Setting

This was an 8-week, parallel-group randomized controlled trial conducted entirely at Yecheon Sports Complex, Republic of Korea, which includes an outdoor synthetic 400-m track and a regulation long-jump sand pit. Athletes were randomized 1:1 to the velocity-based training (VBT) or percentage-based training (PBT) group.

The allocation sequence was generated by an investigator not involved in testing in R (version 4.3.2; R Foundation for Statistical Computing, Vienna, Austria) with a fixed seed (set.seed(20240103)), using simple randomization, and implemented with sequentially numbered, opaque, sealed envelopes (SNOSE) to ensure allocation concealment; outcome assessors were blinded to group assignment. No blocking or stratification was used.

Both groups trained with identical exercise selections, session frequency (three sessions per week), and coaching oversight; the only difference was the load-prescription method (VBT vs. PBT) and the presence/absence of real-time velocity feedback. Because both groups trained concurrently under the same coach, potential contamination was minimized by using dedicated lifting stations/lanes and orienting the Vitruve tablet toward the lifter so that velocity data were visible only to VBT athletes; the coach delivered standardized technical cues without discussing velocity metrics or load targets with the PBT group. No crossover between protocols occurred.

Primary outcomes were 30-m sprint time and vertical jump height; secondary outcomes were standing long jump and standing triple jump. Participants were instructed to avoid additional resistance training outside the intervention and to maintain their habitual diet and sleep. Session attendance and adverse events were prospectively recorded. A representative weekly microcycle is summarized in Table 2, and the 8-week periodization is shown in Table 3.

Table 2. Standardized microcycle implemented identically across groups in the general preparation phase.

Day	Training Contents	Focus/Intensity
Sunday	Active rest	Recovery
Monday	Warm-up; Sprint drills (10–40 m); Acceleration runs; Approach velocity runs; Technical triple jump drills (short approach); Bounding (hops, skips); Medicine ball throws	Technical/Speed (Hard)
Tuesday	Weight training (Back squat, Power clean, Romanian deadlift); Jump squat; Core stabilization; Cool-down	Strength (VBT/PBT) (Hard)
Wednesday	Warm-up; Hurdle mobility; Balance drills; Circuit/core strength; Pool/bike training (low intensity, medium volume)	Conditioning/Recovery (Easy)
Thursday	Weight training (Snatch, Jump squat, Leg extension and curl); Speed development drills; Functional movement and body awareness	Strength-Speed (VBT/PBT) (Medium)
Friday	Warm-up; Technical triple jump drills; Plyometric training (bounding, hurdle hops, depth jumps); Medicine ball throws; Contrast bath (hot/cold)	Explosive power/Technical (Hard)
Saturday	Weight training (Bench press, Lunge jump, Medicine ball slam); Controlled runs/striders; Cool-down	Strength-Endurance (VBT/PBT) (Medium)

Table 3. Periodized weight-training parameters across phases: intensity, sets × reps, and frequency.

Phase/Weeks	Primary Focus	%1RM	Sets × Reps	Frequency
Phase 1 (Weeks 1–2)	Strength foundation; technique	60–70%	3–4 × 8–10	3 times/week
Phase 2 (Weeks 3–5)	Strength development	70–80%	3–4 × 6–8	3 times/week
Phase 3 (Weeks 6–7)	Strength–power emphasis	80–90%	3–5 × 3–5	3 times/week
Phase 4 (Week 8, Post-testing)	Taper and assessment (recovery)	50–60%	2–3 × 8–10	2 times/week (recovery)

Note: In the VBT group, training load was adjusted based on barbell velocity thresholds (e.g., 15–20% velocity loss), whereas in the PBT group, training load was prescribed according to a fixed percentage of one-repetition maximum (%1RM). This distinction allowed direct comparison between velocity-based and percentage-based approaches within the same weekly training framework.

2.3. Velocity Monitoring and VBT Configuration

Velocity monitoring. Barbell velocity was measured with a linear position transducer (Vitruve, Madrid, Spain). The tether was attached to the barbell sleeve on the dominant side; the device was zeroed and calibrated before each session according to the manufacturer’s instructions. The system computed mean concentric velocity in real time and displayed repetition-by-repetition values on a paired tablet application. Signals were processed by the device’s internal algorithms with manufacturer-default sampling and filtering parameters. A representative image of the VBT measurement interface during the squat exercise is shown in Figure 1, while the set-termination logic is described below.

Figure 1. Velocity and power assessment during squat exercise using a linear position transducer.

Autoregulation and set termination. Initial loads targeted a concentric velocity range of 0.60–0.80 m·s⁻¹, depending on the exercise. Within each set, velocity loss (VL) was calculated relative to the set’s fastest repetition using VL (%) = [(v_best − v_current)/v_best] × 100; sets were terminated at 15–20% VL. Inter-set rest was 3–4 min; repetitions were executed with maximal concentric intent while avoiding failure. Weekly progression followed 2–5% load adjustments when mean set velocity exceeded the target range. Total volume (sets × reps) and, where feasible, volume-load (kg) were matched as closely as practicable between groups.

Figure 1 illustrates the kinematic variables recorded during the squat exercise with a 60-kg load using a linear position transducer. The display presents range of motion (ROM), mean power, peak velocity, and fatigue index. These measures provide quantitative insights into neuromuscular performance and allow for the evaluation of power output and movement efficiency during resistance exercise. The linear position transducer quantified barbell displacement along the vertical axis and automatically derived velocity and power outputs in real time to regulate training load, without reference to any commercial brand.

2.4. Training Protocol

During the general preparation phase, male high school triple jumpers participated in an integrated training program combining common weight training and technical/physical exercises. Weight training sessions were scheduled three times per week (Tuesday, Thursday, Saturday) and were performed following either a velocity-based training (VBT) or a percentage-based training (PBT) approach. These sessions included fundamental strength and power exercises such as back squat, power clean, Romanian deadlift, snatch, jump squat, leg extension and curl, bench press, lunge jump, and medicine ball slam.

In addition, technical and plyometric training were performed on non-weight training days. On Mondays and Fridays, athletes focused on sprint drills, acceleration runs, bounding, plyometric exercises, and short-approach triple jump practice to enhance approach velocity and explosive takeoff power. Wednesday sessions emphasized active recovery and supplementary training, including hurdle mobility, core stabilization, circuit training, and low-intensity conditioning such as pool or bike work. Sundays were reserved for full recovery.

This weekly microcycle (Table 2) was designed to balance training intensity across the week, ensuring high-intensity sessions on alternate days while allowing sufficient recovery and technical refinement. The program was specifically structured to enhance explosive strength, power output, and technical efficiency, aligning with the goals of both VBT and PBT interventions. The program was based on a four-phase resistance training model originally proposed by Matveyev [9], ensuring progressive overload and recovery across the eight-week intervention.

The periodized training plan consisted of four phases: strength foundation and technique (Weeks 1–2), strength development (Weeks 3–5), strength–power emphasis (Weeks 6–7), and taper and assessment (Week 8), as outlined in Table 3.

This integrated design enabled a systematic comparison between VBT and PBT approaches within the same weekly framework. In the VBT group, training load was adjusted according to barbell velocity thresholds (e.g., 15–20% velocity loss), while in the PBT group, load was prescribed based on fixed percentages of one-repetition maximum (%1RM). Lower velocity-loss thresholds (≤20%) have been recommended to preserve neuromuscular performance and minimize excessive fatigue, particularly in explosive strength training for youth athletes [10]. In the VBT group, a linear position transducer was attached to the barbell during free-weight squat and power clean exercises to monitor movement velocity and regulate training load in real time [11]. By maintaining identical training content and schedule (Table 2 and Table 3), except for the method of load prescription, the program ensured that any observed differences in performance outcomes could be attributed to the distinct loading strategies rather than disparities in exercise selection or training frequency.

2.5. Testing Procedures

All testing used a standardized 15-min warm-up (dynamic mobility, sprint drills, submaximal jumps) and was scheduled between 10:00 and 12:00 (KST) to minimize diurnal variation; each athlete was tested within the same time window at pre- and post-testing. Each test was practiced twice in a familiarization session before the intervention and performed two trials at pre- and post-testing; the best post-test value (and best pre-test value for baseline) was used for analysis. Tests were performed under calm, dry outdoor conditions on the facility track and sand pit.

2.5.1. 30-m Sprint

Sprint time was recorded on a synthetic track using a Brower Timing System with Smart Start (Brower Timing, Draper, UT, USA). Timing was triggered by the Smart Start ground sensor and stopped by a dual-beam infrared timing gate positioned at 30 m at a height of 1.0 m. Athletes performed two maximal efforts from a standing start, separated by 3–5 min of rest; the faster time was retained for analysis.

2.5.2. Standing Long Jump (SLJ)

SLJ was performed in the long-jump sand pit. With toes behind a marked line, athletes executed a countermovement jump with free arm swing. Distance (start line to rearmost heel mark) was measured using a steel tape (precision ± 0.5 cm). Two trials, 3 min rest; best result analyzed.

2.5.3. Standing Triple Jump (STJ)

From a stationary start, athletes performed a hop–step–jump sequence into the sand pit with free arm swing. Distance was measured as in SLJ. Two trials, 3 min rest; best result analyzed. One to two practice trials were allowed immediately before testing.

2.5.4. Vertical Jump (VJ)

VJ was assessed on a Skyhook Contact Mat (RDM Innovation, Largo, FL, USA), a wireless Bluetooth jump-measurement system with a large 31” × 31” contact area and mobile app (Skyhook App, Version 2.5.1 iOS/Android). The system provides real-time analysis of jump height (flight time), ground contact time, reactive strength index (RSI), and protocol-based testing with onboard/cloud data management. The vertical jump test was performed as a countermovement jump with free arm swing (Abalakov technique) to reflect sport-specific jumping mechanics.

2.6. Standardization, Blinding, and Compliance

All tests were conducted at the same venue and time window to minimize diurnal effects; test order was counterbalanced with ≥3 min between different tests. Outcome assessors were not involved in the training sessions and were blinded to group allocation. Training attendance was logged for every session, and no serious adverse events were reported (occasional delayed-onset muscle soreness only).

2.7. Reliability and Validity

The Vitruve linear position transducer, the Brower timing system, and the Skyhook Contact Mat are widely used for field-based assessment of sprint and jump performance and have demonstrated good-to-excellent test–retest reliability and validity [12] in applied sport settings. In the present study, a familiarization session was conducted, and duplicate trials obtained at the pre-test were used to estimate within-session reliability. Intraclass correlation coefficients ICC(2,1) (two-way random-effects, absolute agreement, single measures) with 95% confidence intervals were calculated from duplicate trials, and coefficients of variation (CV%) were derived from the typical error (TE = SD of the trial differences/√2) expressed relative to the mean of trial means. Reliability estimates were excellent to good across outcomes (n = 12): 30-m sprint ICC(2,1) = 0.975 (95% CI: 0.908–0.992), CV = 0.17%; SLJ ICC(2,1) = 0.938 (0.792–0.981), CV = 0.48%; STJ ICC(2,1) = 0.987 (0.951–0.996), CV = 0.34%; VJ ICC(2,1) = 0.880 (0.637–0.964), CV = 1.18%.

2.8. Statistical Analysis

In addition to the pre-specified ANCOVA (post-test score as the outcome, group as the factor, and baseline as the covariate) with Holm adjustment for the two primary outcomes, we performed small-sample–oriented estimation and robustness procedures. Standardized mean differences (SMD; Hedges’ g, corrected for small samples) with 95% CIs were calculated for between-group effects. Bias-corrected and accelerated (BCa) bootstrap confidence intervals for adjusted mean differences were computed using 10,000 resamples (random seed = 20240103). To account for potential violations of normality and homoscedasticity, a rank-based ANCOVA (Quade) and HC3-robust standard errors were additionally used. To reduce upward bias in partial η² with small samples, adjusted omega-squared (ω²) estimates with 95% confidence intervals were also reported. Model diagnostics examined homogeneity of regression slopes (group × baseline), residual normality (Shapiro–Wilk), and heteroscedasticity (Breusch–Pagan). Collectively, these robustness checks confirmed that the direction and magnitude of the effects remained consistent across analytical approaches.

3. Results

3.1. Baseline Equivalence

At baseline, no statistically significant differences were observed between groups for any outcome. Baseline SMDs ranged from small to large (e.g., long jump SMD ≈ +1.05), underscoring the importance of baseline-adjusted analyses and the inclusion of robustness checks. Accordingly, baseline values were controlled using ANCOVA, and findings were verified through bootstrap and rank-based sensitivity analyses.

3.2. Within-Group Training Effects

Both groups completed the 8-week intervention. Paired-sample tests indicated that the VBT group improved significantly across all outcomes, with effect sizes ranging from moderate to large. The PBT group also showed significant within-group improvements in SLJ and STJ, whereas changes in 30-m sprint and VJ were smaller. Detailed descriptive statistics and within-group pre–post changes are presented in Table 4.

Table 4. Detailed descriptive and within-group statistics for performance outcomes (pre–post).

Variable (Unit)	Group	Pre (Mean ± SD)	Post (Mean ± SD)	Δ% (pre→post)	Paired t p	Cohen’s d	N
30 m sprint (s)	VBT	4.131 ± 0.045	4.090 ± 0.050	−1.08%	<0.001	0.89	6
30 m sprint (s)	PBT	4.154 ± 0.042	4.140 ± 0.040	−0.34%	0.008	0.53	6
SLJ (m)	VBT	2.818 ± 0.052	2.880 ± 0.050	+2.07%	<0.001	1.02	6
SLJ (m)	PBT	2.810 ± 0.060	2.840 ± 0.060	+1.03%	<0.001	0.69	6
STJ (m)	VBT	8.865 ± 0.239	9.020 ± 0.220	+1.64%	<0.001	0.63	6
STJ (m)	PBT	8.780 ± 0.240	8.830 ± 0.240	+0.56%	0.0013	0.72	6
VJ (cm)	VBT	61.533 ± 1.827	65.230 ± 2.020	+6.01%	<0.001	1.39	6
VJ (cm)	PBT	59.580 ± 2.300	60.030 ± 1.390	+0.75%	0.0034	0.45	6

Notes. Pre/post means and SDs reflect values recorded in the Results. Δ% denotes the relative change from pre to post. p values are from paired-sample t-tests (within-group pre–post comparisons). Times are in seconds; distances in meters; VJ in centimeters.

3.3. Between-Group Effects (ANCOVA)

Between-group effects were evaluated using ANCOVA, with the post-test value as the dependent variable, group as the fixed factor, and the baseline value as a covariate. The family-wise error rate was controlled using the Holm procedure for the two pre-registered primary outcomes (30 m sprint and VJ); SLJ and STJ were analyzed as secondary outcomes without multiplicity adjustment.

Adjusted mean differences (AMD; VBT − PBT) favored VBT across all outcomes: 30 m sprint AMD = –0.030 s (95% CI –0.045 to –0.015, p = 0.0016, partial η² = 0.69); SLJ AMD = +0.034 m (95% CI 0.019 to 0.049, p = 0.000611, partial η² = 0.75); STJ AMD = +0.113 m (95% CI 0.080 to 0.146, p = 2.686 × 10⁻⁵, partial η² = 0.87); VJ AMD = +3.15 cm (95% CI 2.59 to 3.72, p = 4.93 × 10⁻⁷, partial η² = 0.95). For both primary outcomes, Holm-adjusted p values remained significant (30 m sprint: p = 0.0016; VJ: p = 9.86 × 10⁻⁷). The corresponding Hedges’ g (95% CIs) indicated moderate-to-large effects for both primary outcomes (Table 5).

Table 5. Compact summary of training effects and between-group differences (VBT − PBT).

Variable (unit)	VBT Δ% (pre→post)	PBT Δ% (pre→post)	ANCOVA AMD (VBT−PBT)	95% CI	Hedges’ g (95% CI)
30 m sprint (s)	−1.08%	−0.34%	−0.030 s	−0.045 to −0.015	−0.90 (−1.75 to −0.11)
Standing long jump (m)	+2.07%	+1.03%	+0.034 m	0.019 to 0.049	+0.48 (−0.22 to +1.31)
Standing triple jump (m)	+1.64%	+0.56%	+0.113 m	0.080 to 0.146	+0.69 (−0.12 to +1.50)
Vertical jump (cm)	+6.01%	+0.75%	+3.15 cm	2.59 to 3.72	+1.17 (+0.19 to +2.08)

Notes. AMD = adjusted mean difference from ANCOVA adjusted for baseline; Δ% = relative change from pre to post. Hedges’ g was computed by dividing ANCOVA-adjusted mean differences by the pooled post-test SD and applying a small-sample correction; 95% CIs were obtained by transforming AMD CIs using the same scale. Primary outcomes are 30 m sprint and vertical jump; Holm-adjusted p values are presented for primary outcomes only.

Robustness and Model Diagnostics

Model assumptions were verified, and results proved robust. There was no evidence of violation of the homogeneity of regression slopes (group × baseline interaction: all p > 0.10). Shapiro–Wilk tests showed residuals were approximately normally distributed (p > 0.10), and Breusch–Pagan tests indicated no serious heteroscedasticity concerns (p > 0.10). In addition to the pre-specified ANCOVA, effect estimates remained consistent in sensitivity analyses using HC3-robust standard errors, bias-corrected bootstrap CIs (10,000 resamples), and a rank-based ANCOVA (Quade). Across all robustness procedures, the direction and magnitude of between-group differences were unchanged, supporting the stability of the findings despite the modest sample size.

A compact summary of between-group effects is provided in Table 5.

4. Discussion

This study investigated the effects of velocity-based training (VBT) compared with percentage-based training (PBT) on sprint and jump performance in high school triple jump athletes. The main findings can be summarized as follows: (1) the VBT group achieved significant improvements in all four performance variables, including a reduction in 30 m sprint time (−1.08%), and increases in SLJ (+2.07%), STJ (+1.64%), and VJ (+6.01%); (2) the PBT group showed a significant improvement only in SLJ (+1.03%), while the other variables remained non-significant; and (3) VBT elicited medium-to-large effect sizes, particularly in VJ (d = 1.39), suggesting meaningful neuromuscular adaptations.

Effect-size estimates (Hedges’ g) supported moderate-to-large practical improvements in both sprint and jump outcomes, although precision was limited by the modest sample size.

The superior outcomes observed in the VBT group may be partly explained by real-time velocity feedback [13,14], potentially enhancing intent to move explosively and neural drive. Such autoregulation may be particularly advantageous in youth athletes, who are generally more responsive to neural adaptations than hypertrophic gains [15]. In line with this interpretation, meta-analytic evidence suggests that VBT may offer small performance advantages over PBT in jump, sprint, and power outcomes [16]. However, these potential advantages should be interpreted cautiously, given the low statistical power of the present trial.

The pronounced improvement in VJ (+6.01%, d = 1.39) may indicate VBT’s potential to optimize the rate of force development (RFD), a key indicator of explosive neuromuscular capacity, which is critical for explosive jump performance [17,18]. Improvements in VJ may contribute to competitive benefits by enhancing vertical power during the final jump phase of the triple jump. Conversely, the PBT group’s reliance on fixed %1RM loads may have induced unnecessary fatigue and insufficient high-velocity neural stimulation, thereby limiting explosive performance gains [19].

The observed effects are generally consistent with prior findings indicating that VBT may equal or surpass PBT in improving explosive athletic performance [14,20,21]. Interestingly, the magnitude of improvement in VJ observed in this study exceeded typical values reported in adult athletes (~4–5%) [22], possibly reflecting the heightened neuromuscular plasticity of adolescents [23]. Improvements in SLJ and STJ further align with evidence that approach velocity and elastic strength contribute substantially to triple jump performance [24].

From a practical perspective, these findings provide initial guidance for coaches when considering training strategies in youth jump athletes. First, the reduction in 30 m sprint time reflects enhanced approach acceleration, which strongly influences jump distance. Second, the observed increases in SLJ and STJ suggest better maintenance of elastic strength across the hop–step–jump sequence. Third, the technology required for VBT (e.g., linear encoders) is increasingly accessible and may be feasible to implement in school or club settings [25]. Nevertheless, future adoption should account for resource availability and athlete readiness.

In summary, this study provides preliminary evidence that VBT may be more effective than PBT in enhancing acceleration, jump performance, and explosive power in high school triple jump athletes. The substantial VJ gains, together with meaningful improvements in SLJ and STJ, highlight the potential of VBT to enhance neuromuscular function during adolescence. However, these conclusions should be considered tentative and require confirmation in larger and more diverse samples.

Several limitations warrant consideration. First, the small sample size (n = 6 per group) restricts generalizability and reduces precision. The observed effect sizes suggest non-trivial improvements, although the possibility of random variation cannot be fully excluded. Second, the eight-week intervention prevents inferences regarding long-term adaptation or translation to competition performance. Third, participants were limited to Korean male high school triple jumpers, potentially limiting external validity. Fourth, lifestyle factors (e.g., sleep, nutrition, psychological state) were not controlled and may have influenced performance responses.

Future studies should therefore include larger cohorts, longer intervention periods, and sport-specific outcomes such as official jump distances, approach velocity, and landing mechanics. Additionally, comparing different velocity-loss thresholds or hybrid VBT–PBT models may help refine training prescriptions for youth athletes. These performance gains may reflect enhanced stretch–shortening cycle efficiency, which is widely considered an important contributor to effective hop–step–jump transitions in triple jump.

5. Conclusions

This study provides preliminary evidence that velocity-based training (VBT) may offer greater improvements in acceleration and explosive jump performance than percentage-based training (PBT) in high school triple jump athletes. The observed gains, particularly in vertical jump performance, suggest that VBT can potentially support neuromuscular development during adolescence.

From a practical standpoint, VBT may serve as a feasible and informative method for guiding training load and intent in youth athletes, especially when velocity-loss thresholds are appropriately monitored. However, given the modest sample size and short intervention period, these findings should be interpreted cautiously and not generalized beyond similar populations.

Future research with larger and more diverse cohorts is needed to validate these promising initial findings and to determine optimal implementation strategies for VBT in athletic development programs.