Stephen Francis - interview/training

Would Stephen’s approach be considered concurrent? It seems to me that it resembles Dan Pfaff’s set up, with 2 accel, 2 special endurance/ week all year long.

Its odd how he says he avoids doing anything over 40m because thats when people tend to pull up with injury. Isn’t this a little counter intuitive?

You are what you train, how do these guys get so fast by avoiding true speed work?

Yes it is concurrent, this has been covered once before. Where are the two special endurance days, I see two ext tempo days?

They are fast because:

1: Genetics

2: They are fit, which helps your overall top speed and your ability to finish races.

3: Yams+Yohimbine.

nothing to think about really. It works for them and we don’t know the extent of the challenge

I am a believer also, my gpp will look very similar to there setup. I am debating if I should have long hills on wed or a controlled speed session. Tue will be tempo 300-400’s and Thur tempo 200’s.

I guess I was thinking that those ext tempo days could have been special endurance if they’re running them fast enough. I thought form runs and slow endurance would have been considered the tempo.

I thought those 300’s were done in 38-43sec = ext tempo for those guys.

How valid is the " it’s okay to train at a high intensity on consecutive days( 30 accel / SE )because we’re training different energy systems?" Might be emphasizing different energy system but it’s with the same CNS and hamstrings.

I understand the risk/reward about max testing with weights but wouldn’t watching the athlete just train give you an idea? Or even do a 3rm? We all know the role max strength plays in rep tests, but is it fair to use it as a predictor?

I a bit confounded that you do not view split squats or step-ups as single leg work :confused:

Split squats use both legs. The hip extensors of the front leg and the hip flexors of the rear leg.

Step ups also involve propulsion from both legs, and aren’t limited by balance due to their dynamic nature.

Neither is really a true unilateral lift.

Ive read his answer to that question before. I believe he has said he doesn’t consider those single leg movements because both feet are supporting the weights. He views One legged squats, single leg dead lifts and such as single leg movements because only one leg is supporting the weight. Correct me if I mistaken.

Split squats I can see, but step-ups, probably not as the majority of the lift is single support. I do not like single leg squats and have never found value in them. I wouldn’t even consider a single leg deadlift.

Without wishing to be overly provocative, what is sprinting other than a series of alternative single-leg impulses, so you can see why some coaches think derivative squats such as single-leg movements are appropriate.

Personally I think single-leg unloaded squats are probably related more to core-strength and stability than actual leg strength.

I coached a young woman who could do 10x single-leg deep squats off each leg (with the other leg triple extended in front but not touching the ground) but she was hopeless doing double-foot squat with a barbell.

On the other hand, I had the chance to work briefly with a footballer who squatted 246kg but could not do one rep unloaded single-leg squat. He could go down, but could not stand back up.


I coached a young woman who could do 10x single-leg deep squats off each leg (with the other leg triple extended in front but not touching the ground) but she was hopeless doing double-foot squat with a barbell.

I share the same opinion. We do the Borzov squats twice a week, 6-10reps of 2-4 sets.

Not provocative at all. I like split squats, lunges and step ups. Mennea did tons of split squats, albeit with relatively light weights.

I agree to an extent regarding the core strength relationship of single leg squats. All squatting movements require an element of core strength though and this needs to be accounted for in the planning of training. How should the ab work change as the squat max increases? An athlete squatting big weights should have incredible core strength, at least in theory. I’m sure that for that football player it was an issue of joint mobility or an inability to activate the appropriate muscle groups either individually or in sequence.

An athlete squatting big weights should have incredible core strength, at least in theory.

Perhaps it also depends on the built of an athlete. Picking between a tall 192-195cm athlete with long legs (hip height 118-121cm) and strong core area (no. of weekly sit ups - 2000) and a shorter 175-180cm athlete of medium built and relatively conditioned core area (who does only up to 150 sit ups a week), IMO squats should suit more the shorter athlete i.e. the shorter athlete would be able to squat more, in this instance even though his core may be slightly less strong. Then you take Timmermann or Günter Werner for example …and they blow this theory out of water.

A prominent US jumps coaches pointed this out during a Q+A a few years back (2002 I think, so maybe more than a few). He felt that he was overloading the core too much during max. strength if he kept the ab volumes high. Seems to make sense.

I do think that you premise that squatting big weights is easier for compact athlete is sound. This is even more the case for longer limbed females in my experience.

Is anyone aware of studies that compare the two and how improvment in one affects the other?

I have something of similar nature handy…

Effect of the movement speed of resistance training
exercises on sprint and strength performance in
concurrently training elite junior sprinters


The aim of this study was to determine the eþ ects of 7 weeks of high- and low-velocity resistance training on
strength and sprint running performance in nine male elite junior sprint runners (age 19.0 ± 1.4 years, best 100
m times 10.89 ± 0.21 s; mean ± s). The athletes continued their sprint training throughout the study, but their
resistance training programme was replaced by one in which the movement velocities of hip extension and
¯ exion, knee extension and ¯ exion and squat exercises varied according to the loads lifted (i.e. 30± 50% and
70± 90% of 1-RM in the high- and low-velocity training groups, respectively). There were no between-group
diþ erences in hip ¯ exion or extension torque produced at 1.05, 4.74 or 8.42 rad´ s-1, 20 m acceleration or 20 m
`¯ ying’ running times, or 1-RM squat lift strength either before or after training. This was despite signi® cant
improvements in 20 m acceleration time (P < 0.01), squat strength (P < 0.05), isokinetic hip ¯ exion torque at
4.74 rad ´ s-1 and hip extension torque at 1.05 and 4.74 rad´ s-1 for the athletes as a whole over the training
period. Although velocity-speci® c strength adaptations have been shown to occur rapidly in untrained and nonconcurrently
training individuals, the present results suggest a lack of velocity-speci® c performance changes in
elite concurrently training sprint runners performing a combination of traditional and semi-specific resistance
training exercises.

Let me know if you want it.

No, but the correlation is high for me. I’ve always found single leg bodyweight squats easier (relative to how difficult other people find it) than regular squats.

The relationship between maximal jump-squat power and sprint acceleration in athletes

Gordon Sleivert1 Contact Information and Matiu Taingahue2
(1) Faculty of Kinesiology, The University of New Brunswick, Fredericton, NB, E3B 5A3, Canada
(2) School of Physical Education, The University of Otago, Dunedin, New Zealand

Abstract This study investigated the relationship between sprint start performance (5-m time) and strength and power variables. Thirty male athletes [height: 183.8 (6.8) cm, and mass: 90.6 (9.3) kg; mean (SD)] each completed six 10-m sprints from a standing start. Sprint times were recorded using a tethered running system and the force-time characteristics of the first ground contact were recorded using a recessed force plate. Three to six days later subjects completed three concentric jump squats, using a traditional and split technique, at a range of external loads from 30–70% of one repetition maximum (1RM). Mean (SD) braking impulse during acceleration was negligible [0.009 (0.007) N/s/kg) and showed no relationship with 5 m time; however, propulsive impulse was substantial [0.928 (0.102) N/s/kg] and significantly related to 5-m time (r=–0.64, P<0.001). Average and peak power were similar during the split squat [7.32 (1.34) and 17.10 (3.15) W/kg] and the traditional squat [7.07 (1.25) and 17.58 (2.85) W/kg], and both were significantly related to 5-m time (r=–0.64 to –0.68, P<0.001). Average power was maximal at all loads between 30% and 60% of 1RM for both squats. Split squat peak power was also maximal between 30% and 60% of 1RM; however, traditional squat peak power was maximal between 50% and 70% of 1RM. Concentric force development is critical to sprint start performance and accordingly maximal concentric jump power is related to sprint acceleration.


In many team sports, sprints most frequently occur over very short distances from both standing and rolling starts. This has been demonstrated in various Rugby codes (Allen 1989; Deutsch et al. 1998; Keane et al. 1993) and in court sports such as basketball (McInnes et al. 1995). It has also recently been shown in Rugby Union that these short sprints typically (~80% of the time) do not involve a change in direction and about 85% of the time begin from standing, walking or jogging starts (Duthie 2003). Therefore, the acceleration phase and predominantly the initial acceleration phase (~0–10 m) are of major importance to athletes, but there is little research on the most appropriate form of resistance training to enhance acceleration. Research on track sprinters starting in blocks has identified that the first few ground contact phases of a short sprint are dominated by propulsive forces when compared to braking forces (Mero 1988), and by concentric muscle actions (Mann and Sprague 1980). The average horizontal impulse of track sprinters in the blocks and during the propulsive phase of the first ground contact have also shown significant correlations with initial running velocity when they are expressed relative to body weight (Mero 1988; Mero et al. 1983). These findings emphasize the dominance of the propulsive phase during initial acceleration, and the importance of propulsive force developed during the first few foot contacts of the sprint in maximizing initial running velocity.

It is generally accepted that for optimum transfer to dynamic movement the characteristics of the resistance-training stimulus should be specific to the activity in terms of muscles used, muscle action type, loading characteristics and range of movement (Sale 1992). Nevertheless, there does not appear to be any consensus on the appropriate method of resistance training to utilize when training to enhance acceleration, and no clear method of resistance training has been shown to enhance acceleration in comparison to other methods. Since explosive concentric muscle actions dominate sprint starts it seems logical that similar resistance training movements might be suitable for testing and training these neuromuscular qualities. Therefore, the relationship was examined between the kinetics and kinematics of two explosive jump squat exercises performed at different relative loads and sprint start performance, with a view to assessing whether or not such exercises should be recommended for individuals wishing to improve sprint acceleration. In this study, concentric jump power was measured in a traditional parallel squat and a split-squat exercise that was designed to mimic the knee and hip angles encountered during standing sprint starts. It was hypothesized that explosive concentric jump squat power for both exercises would be significantly related to sprint acceleration, but that split-squat power would show a stronger relationship to sprint performance than traditional squat power because it was a more sprint specific exercise.



The institutional ethics review board approved this study. After signing informed consent, 30 male athletes of mean (SD) age 20 (2.2) years, 184 (7) cm height and 90.6 (9.3) kg body mass, participated in this study. Subjects were selected from power-type sports and were actively competing in competitive rugby union (n=22), rugby league (n=5), or basketball (n=3). They had at least 1 year of weight training experience [mean 2.4 (1.4) years]; however, none had previously performed the split-squat exercise used in this study. All subjects were injury free at the time of testing.

Experimental design

The testing of each participant was administered over five separate sessions—two familiarisation sessions, two strength/power testing sessions, and one sprint testing session. All familiarization and testing sessions were completed on separate days with a minimum of 1 day and a maximum of 4 days between sessions. All subjects first participated in two separate familiarization sessions where they performed maximal sprinting and maximal effort jump squats using both parallel and split-squat techniques. Subsequently they were tested for maximal 1 repetition maximum (1RM) parallel and split-squat strength (random order), power at different percentages of 1RM load for both split and parallel jump squats, and sprint performance. The sub-1RM jump-squat trials were also performed in random order so as to avoid any systematic sequence effects on jump-squat kinematics and kinetics.
Sprint performance

Each subject participated in a standardized warm-up protocol consisting of 5 min cycling at a self-selected pace, 2 min of static lower-limb stretching and five 15-m sprints, the first three at 70–90% maximum effort and the last two at maximal effort. After 5 min rest they then performed six maximal 10-m sprints from a standing two-point start (straddle position) and data from the fastest sprint were used in all subsequent analyses. A minimum work to rest ratio of 1:40 was used to ensure recovery between each sprint. Sprint time data were recorded at 5-cm intervals from a hand-held wire velocimeter. Briefly, this system consists of wire on a spool attached to the middle-back of the athlete’s waist. As the athlete sprints, the spool unwinds and a magnetic encoder sends a pulse (time-stamp) to the computer every 5 cm. Pilot work indicated that this device provided reliable data over a short distance sprint, e.g., coefficient of variation (CV) for time to 5 m=2.1%, but 5-m times are typically slower (by ~0.25 s) than when measured using timing lights. This is because with the velocimeter timing begins as soon as the subject initiates movement, but with timing lights, timing does not begin until the light beam is broken by the subject’s body. Thus, with timing lights a flying start is used and sprint times are shorter than those reported in this study. The horizontal ground reaction force of the first foot contact was also recorded via a calibrated force plate (AMTI, Mass., USA) recessed in the laboratory floor. Force data was amplified 2,000× and low-pass filtered (10.5 Hz). The ground phase was divided into braking and propulsive phases according to negative and positive horizontal forces. The horizontal propulsive impulse was calculated by multiplying the average horizontal force during propulsion by the duration of the propulsive phase. Pilot work found that the reliability of measuring propulsive impulse via the above method was very good (CV=2.2%); however, the same was not true for braking impulse (CV=46.6%). The low measurement reliability for braking impulse was likely due to the minimal nature of the braking impulse during early acceleration and a subsequently very high signal-to-noise ratio.

Range of motion at the hip and knee joint during ground contact was also measured from captured video using qualitative analysis software (Video Expert II, Sport & PE Technology, New Zealand). The first 4 m of the sprint were filmed with a video camera (Panasonic M40, Japan) using a shutter speed of 0.002 s and a frame rate of 25 Hz. The camera was positioned at a 90° angle to the force plate, i.e., first foot contact. Hip and knee angles were measured at the time of initial ground contact, then 0.04 s and 0.08 s after this initial contact. The velocimeter, force plate, and video were all synchronized using an event synchronization unit (in-house developed), which provided a reference point on the video that corresponded to the start of data collection from the force plate and wire velocimeter.
Strength testing:

On separate days (1–4 days apart), 1RM was determined for each of two squatting techniques on the Smith Squat machine. The two squat techniques tested were: (1) traditional squat—feet shoulder width apart and directly under the bar, bar held across the top of the shoulders and upper back, bottom point of the lift defined as knee angle equal to 90° (measured manually with a goniometer), no limits on hip angle but attempt was made to keep torso as vertical as possible; and (2) split squat—a modification of one commonly used in jump plyometric training and has been described elsewhere (Allerheiligen 1994). Briefly, feet were placed in astride position with front foot toes directly under the bar, bar held across the top of the shoulders and upper back, bottom point of the lift defined as that which causes the knee [102 (10)°] and hip angles [110 (11)°] of the front leg to be equal to those measured during the first foot contact of the sprint. These angles were measured manually using a goniometer. Back foot toes were positioned directly behind the front foot heel.

For each exercise, subjects performed a warm-up of two sets of five repetitions at 50% of their perceived 1RM. They then performed a concentric-only squat at approximately 80% of their self estimated 1RM with knee angle beginning at 90°. Following successful completion of that lift, the load was increased to approximately 90% of estimated 1RM and the squat repeated. The load continued to be increased in 5- to 10-kg steps until the squat could no longer be completed successfully. The last successfully completed squat was recorded as 1RM. Fifteen minutes rest was provided between successive lifts. If 1RM was not determined within five lifts, then subjects returned on another day for retesting.
Power testing

Force exerted into the bar, velocity of the bar, and mechanical power exerted into the bar were measured at loads ranging from 30% to 70% of 1RM (10% 1RM increments) for both the traditional squat and split-squat technique. Participants performed three explosive jumps at each load (random order) in an attempt to maximize their power output and data from the most powerful repetition at each load was used in all subsequent analyses. A standardized warm-up protocol was followed before testing began and consisted of 5 min of cycling at a self-selected intensity, 2 min of lower-limb static stretching, two sets of five squats at 40% and 60% 1RM followed by a set of maximal effort jump squats with a load of 40% 1RM. A minimum of 3 min was allowed for recovery between successive lifts. Velocity-time, force-time, and power-time curves were collected for the concentric phase of each lift.

For all jumps the participants were harnessed to the bar via an upper torso harness to minimize movement of the bar off the shoulders. Knee and hip angles were measured with a manual goniometer while the subject was supporting an unloaded bar, and mechanical stops were positioned under the bar to form the start position. All lifts were performed from this dead stop position with no counter-movement, thus concentric force development was measured with no eccentric pre-load. Participants were instructed to explode powerfully on the “go” signal in an attempt to move the load as fast as possible through the entire range of motion. Two spotters then caught the weight at the top of the lift.
Force exerted into the bar by the subject was calculated by multiplying the mass of the loaded bar by the acceleration of the bar and then adding this to the weight of the loaded bar:

$$ {{\rm{Force}}{{{\rm{exerted}},{\rm{by}},{\rm{subject}}}} {\rm{ }}{\left( {\rm{N}} \right)}{\rm{ = Weight}}{{{\rm{bar}}}} {\left( {\rm{N}} \right)}{\rm{ + }}{\left( {{\rm{mass}}{{{\rm{bar}}}} ,{\left( {{\rm{kg}}} \right)}{\rm{ imes acceleration}}{{{\rm{bar}}}} ,{\rm{by}},{\rm{subject}},{\left( {{\rm{m/s}}^{{\rm{2}}} } \right)}} \right)}} $$

Acceleration of the bar was recorded with an accelerometer (Analogue Devices ADXL150EM1, Mass., USA) fixed to the Smith Press bar assembly. The accelerometer had a sensitivity of 200 mV/g and used a bandwidth of 0–100 Hz. Peak force and peak rate of force development was calculated for the concentric phase of the force-time curves of each squat. The bar velocity of each 10-ms segment of the lift was obtained by an integration of the acceleration signal from that segment. These velocities were then summed consecutively to measure bar velocity throughout the lift (Thompson and Bemben 1999):

$$ {{\rm{Velocity}}{{{\rm{t2}}}} ,{\left( {{\rm{m/s}}} \right)}{\rm{ = Velocity}}{{{\rm{t1}}}} ,{\left( {{\rm{m/s}}} \right)}{\rm{ + }}{\int\limits_{{\rm{t1}}}^{{\rm{t2}}} {{\rm{acceleration}},{\left( {{\rm{m/s}}^{{\rm{2}}} } \right)},{\rm{dt}}} }{\rm{ }}} $$

Pilot data suggested that velocity calculated from the accelerometer had less noise and gave similar values to those obtained from a velocity encoder. Average and peak bar velocity were obtained from the concentric phase of the velocity-time curves of each squat. Power measurements were obtained by multiplying the force exerted into the bar by the velocity of the bar:

$$ {{\rm{Net}},{\rm{power}},{\left( {\rm{W}} \right)}{\rm{ = Force}},{\rm{into}},{\rm{bar}},{\left( {\rm{N}} \right)}{\rm{ imes velocity}},{\rm{of}},{\rm{bar}},{\left( {{\rm{m/s}}} \right)}} $$

It should be indicated that the force and power calculations are therefore not representative of the force and power output of the subject against the ground, but rather representative of the subjects force and power against the bar. All force and power measurements are presented relative to body mass.

Means and standard deviations were used to describe all variables. Paired t-tests were used to identify significant differences between: propulsive and braking variables, split and traditional squat force, velocity, power, and rate of force development variables. Repeated measures one-way ANOVA were used to determine the effect of relative load on force, velocity, and power during both squats. Scheffé post-hoc analyses were then used to identify the loads or loading ranges at which maximum force, velocity, and power occurred. Pearson product-moment correlations were also calculated to determine the individual relationships between 5-m sprint time and force, velocity, power, and rate of force development for both squatting techniques. A stepwise multiple linear regression was then used to determine if the combination of any of these variables improved prediction of 5-m sprint time. A type I error rate of 5% was accepted for significance in all statistical analyses.


The major finding of this study was that maximal concentric power generated during both squat exercises was significantly related to 5-m sprint time. The correlation coefficients for these relationships (r=–0.64 to –0.68) were similar to that which Young et al. (1995) reported between average power in a concentric squat jump with an external load of 19 kg and 2.5-m sprint time (r=–0.74). There appeared to be no major advantage of using the split or traditional squat, as both produced similar power measures and had similar relationships with 5-m sprint time, even though the split-squat was a more mechanically specific exercise to the standing sprint start. The smaller, more specific range of motion in the split squat did, however, allow a larger load (but lower velocity) to be applied to the specific range of motion used during sprinting. Additionally, the split squat measurements had higher coefficients of variation, which could have reduced the Pearson R values.

Peak force and bar velocity were also related to 5-m sprint time although less substantially than power. Stepwise multiple linear regressions reinforced the fact that the power variables of both squat exercises accounted for the most variance in 5-m sprint time. Maximum power was observed over a wide range of external loads and thus, peak force and bar velocity was often measured within this range. As both force and velocity contributed to the development of power throughout this range, it was understandable that they also showed some relationship with 5-m sprint time. However, if peak force and bar velocity had been measured at both extremes of the force-velocity spectrum (i.e., unloaded and maximally loaded), the role of each in relation to sprint start performance may have been presented more clearly.

In the traditional squat, peak force (r=0.59) was more substantially related to 5-m sprint time than peak bar velocity (r=0.40). During the sprint start, the body had to be accelerated rapidly from stationary and the propulsive impulse was large, so there may have been a greater reliance on high force production as opposed to high movement velocity. Young et al. (1995) reported a high correlation coefficient (r=–0.86) for the relationship between 2.5-m sprint time and peak force in a concentric jump squat, with an external load of 19 kg. However, as such a small external load was used in the study by Young et al. (1995), it is difficult to make a comparison with the present findings because our peak force was determined using near maximal loads. In general, very few studies have investigated the initial acceleration phase and its relationship to peak force, using near maximal loads. Additionally, no previous studies were found that reported relationships between bar velocity and sprint start performance. As running velocity approaches maximum, those strength measures that require force to be produced at high velocities have been reported to be significantly related to sprint performance (Alexander 1989; Nesser et al. 1996; Wilson et al. 1995;Young et al. 1995). Wilson et al. (1995) reported a significant relationship between force at 30 ms in a concentric jump squat and 30-m sprint time (r=–0.62). Nesser et al. (1996) reported significant correlations between 40-m sprint time and peak isokinetic torque at a speed of 7.85 rad/s for the hip and knee extensors and knee flexors (r=–0.54 to –0.61). The present results therefore support what little has been reported in the literature and indicate that the development of peak force plays a larger role in sprint start performance than movement velocity.

A secondary purpose of this study was to determine the influence that relative load had on squat kinetics and kinematics. Split-squat peak and average power were maximal when using external loads equal to 40% and 50% of 1RM, respectively. Traditional squat average and peak power were maximal when using external loads equal to 40% and 60% of 1RM, respectively. There was however, no significant difference in split-squat peak or average power, and traditional squat average power when using external loads between 30% and 60% of 1RM. There was also no significant difference in traditional squat peak power when loads between 50% and 70% of 1RM were used. These loads at which maximum power were generated, are similar to those reported in the existing literature for athletic subjects, although several different loading ranges have been identified: between 15% and 45% of 1RM (Cronin et al. 2000; Newton et al. 1997); between 45% and 80% of 1RM (Izquierdo et al. 2002); and approximately equal to 30% of maximum isometric force (Kaneko et al. 1983; Wilson et al. 1993). The reasons for these discrepancies can largely be explained by the mode of exercise used and the methods of calculating power. When the power movements are in the vertical plane (e.g., squat), force calculations must be adjusted to include the effects of gravity on the load, and this has the effect of increasing the relative load at which peak power occurs. Additionally, some studies have also incorporated body mass into their power calculations when exercises are performed in the vertical plane, on the assumption that it too is being accelerated. This has the effect of markedly increasing absolute power (to the order of 5,000 W) and concomitantly reducing the relative load at which peak power occurs to values of 15–20% of 1RM, instead of 50–60% of 1RM. Given that not all of the body mass is usually accelerated and that the velocity of the bar (not the center of mass) has been measured and used in these calculations, many of these power derivations are not strictly mechanically correct. Clearly, a standard method for calculating power in resistance training movements needs to be agreed upon. In the meantime, researchers and practitioners should be aware of the implications resulting from including or excluding body mass in power calculations for exercises occurring in the vertical plane.

The range of relative loads at which peak power occurred in the traditional squat (50–70% of 1RM) was slightly higher than for traditional squat mean power and mean and peak power in the split squat (30–60% of 1RM). This may have been explained by the lower starting position for the lift (knee=90°, compared with knee ~110° for the split-squat). Because of the low start position the 1RM for the traditional squat [149.5 (22.6) kg] was significantly less than that for the split-squat [206.6 (34.4) kg]. In the traditional squat, loads were difficult to move initially, but later in the lift could be moved at high velocities and similar to those observed in the split-squat. Through the range of movement in which peak power occurred, both squats had similar bar velocities and absolute loads, although relatively the loads were very different. Therefore, traditional squat peak power was maximal at higher relative loads than that observed for the other power variables. Traditional squat average power was not affected in the same way, as the averaging process combined the initial low velocity as well as the latter high velocity movement. It is therefore suggested that when prescribing maximal power training using different exercises and ranges of motion, the loading guidelines should be specific to each individual exercise. Additionally, determination of the external training load that maximizes power for each exercise should be based on a maximal strength performance that involves the same range of motion as will be used in training. Related to these points is the selection of a resistance training load for optimizing sprint acceleration. In this respect, training at peak power output is commonly prescribed, but whether training at a load eliciting peak power is better than training at other loads for improving sprint acceleration is unknown. This should be investigated in a longitudinal training study.

Maximum power output in both squat exercises was strongly related to peak force and peak rate of force development but was not related to peak bar velocity. Kaneko et al. (1983) support this finding since their results indicate that in a training condition requiring maximal attempts to accelerate the training load, significantly larger improvements in maximal power were reported when using maximal loads as opposed to unloaded conditions. In the present study, the correlation coefficients for the relationships between peak force and average and peak power were similar for both squat exercises (r=0.78–0.84). Thomas et al. (1996) also reported a strong correlation between 1RM and maximum power output of the leg extensors (r=0.76). It was expected that power would be more closely related to peak rate of force development than peak force. However, the strength of the relationships between peak rate of force development and average and peak power varied considerably for both squats (r=0.62–0.80). Thus, peak force appeared to be more strongly related to power in general terms. These variations in correlation coefficients may have been explained by the differences in reliability for measuring peak rate of force development (CV=3.7–6.8%) when compared to measuring peak force (CV=2.1–3.0%). As maximum power was significantly related to peak force but not peak bar velocity, it appears that the ability to generate peak forces with heavy loads is a more important attribute for maximizing power output, than the ability to generate peak velocities with light loads.

In summary, maximum concentric power in parallel and split-squat exercises are moderately related to sprint start performance, and the mechanical specificity of the resistance exercise with respect to sprinting appears to be of little importance. Maximum concentric power can be generated using a range of external loads between 30% and 60% of 1RM when the effort to maximize bar acceleration throughout the entire range of movement is made. The specific load at which maximum power occurs within this range is highly dependent on the individual involved. Therefore, if a resistance training goal is to train at maximum concentric power, then this loading range can be used, with an emphasis on either high force-low velocity or high velocity-low force, depending perhaps on the specific training requirements of the individual. A training study is required to determine the effects of these types of training on sprinting performance.