What is velocity-based training?

Content of essentials article

• Article summary
• What is velocity-based training?
• The traditional approach to programming
• Why velocity-based training?
• Velocity-based technology
• Velocity-based autoregulation
• The load-velocity profile (LVP)
• Velocity-based autoregulation in action
• Benefits and challenges of velocity-based training

Article summary

• Velocity-based training (VBT) is a set of guiding principles and methods, not something “you do”
• VBT involves tracking movement velocity, typically a barbell’s, to inform diagnostic and programming decisions
• A fundamental purpose of VBT is autoregulation, with the ultimate goal of providing a complimentary programming strategy to traditional methods
• VBT can be used to predict 1RM but it is not always accurate
• VBT can be used for training volume through velocity loss and velocity stops

What is velocity-based training (VBT)?

Coaches should view velocity-based training (VBT) as a set of guiding principles and methods that you can use to enhance a training environment. At the heart of those principles is an objective strategy for quantifying physiological status, training effort and performance improvements. Coaches should not think of VBT as something “you do.”

At the simplest level, VBT involves tracking movement velocity, typically a barbell’s, to inform diagnostic and programming decisions. Coaches can choose from a continuum of practical, applied options that vary in immersive nature, complexity and the variable they target (Figure 1). Four distinct categories emerge from this framework: feedback, autoregulation, testing and volume control.

Before going into “why VBT,” we must consider which velocity metric to use.

Mean, mean propulsive and peak velocities are all options, and a coach’s decision to use one over the other is a function of training goals, personal preference or equipment availability, with each one potentially being more suitable for specific phases of training (Table 3).

For example, peak velocity ought to befit explosive or ballistic-type exercises as it reflects an individual’s maximum capabilities. However, the data are conflicting. Peak velocity appears more reliable than its mean counterparts for loaded jumping,^[14] but not for bench press throwing,^[15] creating confusion for practitioners. Additionally, mean propulsive velocity might provide a more accurate reflection of an individual’s neuromuscular capabilities, whereas mean velocity might better reflect athletic performance (Table 3).

Practitioners should seek to understand the underpinning mechanics of the different metrics and determine the most appropriate fit for their training needs.

The traditional approach to programming

The interaction between two programming variables, volume and load, is essential for maximising physiological adaptations. Volume is typically prescribed via sets and repetitions, using specific ranges based on physical components or physiological mechanisms (Table 2).^[1,2] The most common method for prescribing load is percentage of one repetition maximum (% 1RM), which entails a two-stage process: 1) perform a baseline incremental 1RM protocol; and 2) prescribe submaximal 1RM percentages corresponding to desired physiological adaptations, e.g., 90% 1RM to increase motor unit recruitment, rate coding and intra- and intermuscular coordination.^[3,4]

This traditional approach to programming, however, presents challenges. Coaches often implement simple load-rep continuums to programme in accordance with physical performance goals: maximal strength, power, hypertrophy and so on (Figure 2).^[5] Similarly, they prescribe rep ranges with respect to specific % 1RM, e.g., 3 reps at 90% 1RM for maximal strength improvements.

This approach relies on the premise that all individuals respond to the same stimulus in the same manner. But different athletes, like weightlifters vs. endurance athletes, perform a statistically significantly different maximal number of repetitions for a given % 1RM. The results from one study showed differences at 70% [17.9 vs. 39.9], 80% [11.8 vs. 19.8] and 90% [7.0 vs. 10.8].^[6]

Therefore, blanket prescriptions across athletes and squads could alter targeted physiological adaptations or provide a sub-optimal training stimulus.

Prescribing load via % 1RM across the course of an intervention can also limit athlete outcomes. An individual’s 1RM is fluid, and can change with their physiological or psychological status.^[7,8] Maximum strength can fluctuate as a result of fatigue, sleep, stress or nutrition.^[9–12] Alternatively, strength adaptations will likely occur across the course of an intervention. This iterative nature of strength can therefore reduce the efficacy of more traditional programming.

Combining traditional methods of prescription with contemporary autoregulatory tools such as velocity-based training (VBT), therefore, may permit more flexible programming while accounting for fluctuations in strength and fatigue.

Why velocity-based training?

The load-velocity relationship, similar to the force-velocity relationship A.V. Hill first theorised in 1938, underpins VBT.^[16] There is a finite amount of time available for force production due to the underpinning physiological mechanisms, so as the velocity of muscle shortening (i.e., velocity of movement) increases, force production decreases, and vice versa (Figure 4).

Force (N) underpins movement through acts of pushing or pulling (Force_net = mass x acceleration). Mass is constant during resistance training, so increases in force can only occur with increases in acceleration. Acceleration is a derivative of velocity, situating velocity at the heart of how we can quantify responses to resistance training.

Velocity-based technology

Technological advancements (see Weakley et al.^[17] for a comprehensive overview of VBT devices) have made measuring velocity live in session very simple. Most devices offer sophisticated interfaces, attractive online portals and plug-and-play hardware that are fast becoming a one stop shop for all programming, testing and monitoring needs.

Linear position transducers (LPTs) such as Gymaware are the gold standard, with low systematic error and high repeatability. However, they often come with higher price tags (~£2000). Inertial measurement units (IMUs) such as PUSH and Output Sports offer a much more attractive price tag and more versatility because they are wearables.^[18] However, IMUs’ data quality can sometimes suffer.

An exciting technology comes in the form of smart-device applications such as MyLift. These apps offer a cheap (£9.99) and tech savvy alternative, but can be limited in their functions, particularly if you want to use them live.^[18]

The portable, cost-effective and adaptable nature of VBT technology has enabled gyms, clubs and organisations all over the world to implement velocity-based training, accentuating the need for coaches to understand the most suitable device for their environment.

Velocity-based feedback

A basic principle of VBT is the simplicity of providing external sources of information to athletes, i.e., augmented feedback.^[21]

Live visual, verbal or auditory feedback can enhance acute and chronic physical performance,^[22,23] and is a major benefit of VBT (data taken from Thompson et al.^[24] under review). By providing visual or verbal feedback to athletes, these technologies help performance practitioners create a competitive environment, increase athlete engagement, raise motivation and spur athlete intent. VBT devices can provide specific kinematic targets, which can be more effective than simple verbal encouragement or instructions (e.g., “lift as fast as possible”).^[25,26] This simple application can be highly effective and is fundamental when implementing VBT.

Velocity-based autoregulation

A fundamental purpose of VBT is autoregulation, with the ultimate goal of providing a complimentary programming strategy to traditional methods. This has produced a continuum of practices designed to optimise load through sessional tracking of velocity (Figure 5).

Simple velocity-based autoregulation (VBA) methods might involve performing key movements (e.g., squat or jump) against light loads during warm-up sets to estimate readiness to train. By consistently performing the same protocol, coaches can roughly identify athletes who are particularly fatigued, and can then adjust their prescription accordingly. This, however, might not account for progressive increases in strength and power; and would only estimate whether to alter a program, but not how or by how much.

If coaches want to provide targets to athletes based on specific training parameters or adaptations, they can employ generic velocity zones (Figure 6). Coaches can then manipulate the absolute load the athlete lifts to maintain velocity within these pre-determined zones. This is a quick and efficient strategy, one that is very common in practice.^[24] However, these zones can be quite wide, potentially resulting in an inadequate stimulus that could impact the desired physiological adaptation.

An alternative strategy that would individualise the process is simply tracking an athlete’s historical training data to determine “baseline” velocities for specific load-set-rep combinations, and identify whether adjustments are required based on what they have previously done. Without up-to-date maximum strength data, however, it would be difficult to associate velocities performed with specific % 1RM, again impacting physiological adaptations.

A more immersive and effective way to manipulate and optimise load, therefore, might be with the introduction of the load-velocity profile (LVP).

The load-velocity profile (LVP)

How to construct one

An LVP represents the load-velocity relationship, and is the result of an incremental protocol similar to that for the 1RM but that measures velocity throughout. Absolute or relative (% 1RM) load can be plotted against the velocity metric of choice – mean, mean propulsive, or peak – to identify load-velocity characteristics (figure 7). Then we can apply a linear regression to generate a predictive equation to help determine load or velocity training targets.

Reliability vs. validity vs. stability

When basing training off of LVPs, practitioners need to consider their reliability, validity and stability.  

Reliability is the reproducibility of the profile: if we collect the same data on the same person under the same physiological conditions, will we get the same profile? Of course, like all tests, there will be associated error to factor in. Exercises such as bench press, back squat, deadlift, prone bench pull and power clean show excellent levels of reliability (intraclass correlation coefficient > 0.9 and coefficient of variation < 10%) .^[28–33] Coaches should carry out simple, in-house reliability protocols and statistics to ensure the data they collect is reproducible.

Validity refers to the strength of the load-velocity relationship, expressed statistically as R² or r, and is greater than 0.95 across many exercises.^[28–33] Again, when plotting the data, coaches should check the R² value to ensure the data is of acceptable validity.

Finally, the stability of an LVP will dictate its effectiveness for programming, monitoring and autoregulation. Despite limited research, velocities at specific % 1RM seem consistent, irrespective of fluctuations in strength or fatigue, baseline strength levels or lifting experience (Figure 8).^[34–38]

Ballistic vs. non-ballistic

LVPs rely on two fundamental principles. Athletes must perform each repetition with maximal velocity and intent to maximise reliability of the data, and coaches must consider the full load spectrum from body weight to 100% 1RM.

LVPs are commonly developed using non-ballistic exercises like the back squat, bench press or deadlift. However, performing a back squat with 20% 1RM at maximal velocity can be challenging. Additionally, there is a period of deceleration (negative acceleration, to be precise) at the end of the concentric phase, which can artificially lower outputs such as mean velocity (table 4).^[39,40] Ballistic equivalents like a loaded jump squat, bench press throw or trap bar jumps do not contain this period of deceleration. The athlete projects themselves or the object into space, resulting in a longer period of acceleration and superior mechanical output.^[41,42] Therefore, ballistic exercises with light to moderate loads could provide a more accurate and comprehensives reflection of an individual’s L-v capabilities.

Autoregulation

Early advocates proposed that LVPs could replace traditional strength testing by deploying normative data and generalised equations from the literature. Researchers suggested that any athlete could determine appropriate loading without the need for load-velocity data collection.^[36,44,46] While this would be a time efficient strategy for coaches to undertake, and in some cases the only possible option, L-v data is typically exercise and athlete specific (Figure 9, Table 5).^[28–30,32]

The individualised nature of load-velocity data suggests that coaches need to establish a baseline profile in order to fully maximise its potential. Coaches can then prescribe specific individualised zones from LVPs according to the desired physiological adaptations and relative loads, much like the generalised zones. For example, an athlete could be performing a maximal strength block of training with loads at 90% 1RM. If this athlete performs 90% 1RM at 0.4 m/s, a sensible zone of, for example, 0.37-0.43 m/s could ensure the absolute load hits the desired targets.^[38,53]

The difficulty with such a method is determining the necessary load adjustments to meet the specific physiological status of the individual for that day. Research has suggested ±0.06 m/s =  5% 1RM, ^[38,53] based on a smallest detectable difference. Smallest detectable difference is the smallest change in a measurement that, statistically, is not due to error.^[29] This is an effective method for autoregulating load, but there is still a generalised element to it.

1RM prediction

Plotting load against velocity allows practitioners to calculate a predictive equation for this relationship. One of the main uses for this equation is to estimate 1RM. Then coaches can individualise load manipulations on a session-by-session or even set-by-set basis, precisely prescribing the magnitude of the required change.

1RM prediction has had varying success in the research literature. Very accurate estimates of maximum strength are possible, particularly in upper body exercises such as bench press and prone bench pull.^[44,54–56] Importantly, the majority of this research used fixed path Smith machines. In lower body free weight exercises, 1RM prediction is not always as accurate, with mean differences going up to 30kg in some cases.^[57–59]

There are two potential reasons for this discrepancy between upper and lower body predictive validity:

1. Poor reliability of the V_1RM (velocity at 1RM), sometimes referred to as the minimum velocity threshold (MVT).

Predictive equations require a point of extrapolation, that is, a value that tells the equation where to take the predictive line. If you want to predict 1RM, the V_1RM seems the obvious choice. Poor reliability ^[28,29,59] and association with the line of best fit could impact the predictive validity of the model.

2. The load-velocity relationship for lower body exercise is not linear.

LVPs are often the result of linear regression. However, if the relationship is not truly linear, this mathematical model will not produce an accurate equation, since it would violate one of its main assumptions. That directly impacts predictive validity.

These explanations point to two possible solutions:

1. Pick a more reliable load as the point of extrapolation.

If the V_1RM isn’t reliable, use a lighter load as the point of prediction. 80% 1RM is reliable across lower-body exercises, so a simple extrapolation from predicted 80% 1RM to 100% 1RM can then be applied to estimate maximum strength.

2. Apply a curvilinear regression model (e.g., second-order polynomial) to the data.

Applying an extension of the linear regression model to lower body free weight LVP data could provide a better fit and a more accurate prediction. This is easily applied in most mathematical software, including Microsoft Excel.

The research literature supports both of these solutions.^[42,60] To ensure the effectiveness of their protocol, coaches would typically administer a submaximal LVP (20-80% 1RM in 100% increments) during the warm-up at the start of the session. Then they would recalculate all loads for that session based on the new predicted 1RM. However, getting full squads through a 4-6 load protocol can be difficult under the constraints of time and space. A more time-efficient protocol is the two point method.^[61,62]

The two point method follows the same principles as a traditional LVP, but only utilises two loads to create the predictive model (Figure 10). This model has  high levels of predictive validity and reliability in upper body exercises.^[55,62,63]

There are a few methodological and practical considerations when implementing the two-point method. First is the assumption that the relationship is truly linear (which, for upper-body exercise, seems to be the case), as only plotting two points will manufacture a perfect relationship (R² = 1). Second, the most reliable method requires the two loads be from reasonably far ends of the scale, e.g., 20% and 70% 1RM. Such large jumps in load could cause practical issues in the gym.

Despite limited evidence, VBT appears to be an effective autoregulatory and programming tool, particularly compared to traditional, non-flexible programming methods such as % 1RM. VBT groups across multiple studies produced superior strength, countermovement, sprint and change of direction performances than their traditional counterparts.^[38,53,64] This was also true when VBT was compared to the subjective autoregulatory method of repetitions in reserve (RIR).^[65] This research provides confidence in the effectiveness of VBT to objectively autoregulate load on a sessional and set-by-set basis.

In addition to superior strength and performance gains, VBT can also regulate volume more effectively than traditional methods. A crucial finding of recent intervention research was the ability to elicit greater improvements from significantly lower volume, time under tension and session RPE,^[38,53,64] a crucial discovery for practitioners. After all, who wouldn’t want greater gains for less work?

Testing

LVPs can be a performance diagnostic tool. Bookending a training intervention with test-retest assessments will detail specific force-velocity adaptations.

For example, if an athlete performs 100kg (~80% 1RM) at 0.6 m/s during the baseline profile and then post-training performs the same absolute load at 0.68 m/s, the athlete has increased their strength, provided we account for technological and procedural measurement error has been accounted.

Remember, F_net = ma. Applying this relationship across the full load-velocity spectrum will indicate how force-velocity characteristics changed as a result of training, providing more detailed information than perhaps the 1RM alone.

Volume control – Velocity loss and stops

One half of the battle in optimising programming is appropriate loading strategies. The other half is controlling and regulating fatigue. Inappropriate volume is typically the biggest root cause for excessive fatigue and muscular breakdown. Individuals come in all shapes and sizes with varying physical capacities, and blanket prescriptions of set-rep-load combinations might not be the most effective method to programme. For example, two athletes performing the same rep-set-load scheme (e.g., 5 x 5 at 85% 1RM) could have very different physiological responses. One athlete could be working to relative failure, while the other completes the task relatively comfortably. Cue VBT!

We previously described the inverse relationship between load and velocity. This is also true of velocity and repetitions.^[66] <ean and peak velocity will reduce across a set until the athlete reaches failure and an MVT is determined. This MVT, often referred to as “fail speed” can be an important piece of knowledge towards minimising fatigue, particularly around competition.

Velocity loss thresholds (VL) are a simple way of autoregulating volume in accordance with an individual’s physiological status. VL are quantifiable drops in velocity that dictate when to terminate a set. These markers are typically implemented as a percentage (e.g., 10-50% VL) of the first repetition or best repetition’s velocity, but can also be an absolute value VL (m/s).

Velocity stops (VS) work under the same principles as VL, however, provide an athlete with a specific velocity to terminate a set at. Typically, this method will require an LVP, but it could also derive from generalised, normative data. By applying a specific velocity stop, a coach has an appropriate proximity to failure to prescribe, ensuring minimal residual fatigue build-up.

The greater the VL, the more repetitions the athlete performs, resulting in a closer approach to failure.^[67–69] Similarly, greater metabolic, mechanical, perceptual and neuromuscular responses result from performing a set with a higher VL.^[67,70] VL are also effective in maintaining kinetic and kinematic output across the course of a set and session compared to traditional prescriptions.^[13,71]

Most importantly, VL are reliable across multiple sessions and weeks, ensuring effectiveness over time. This is useful for S&C coaches who might want to programme VL into a training phase to regulate metabolic and muscular fatigue, while maximising mechanical and neuromuscular output.

Benefits of velocity-based training

VBT is a very versatile tool that can compliment traditional methods of testing and programming. The benefits of VBT include:

• Underpinned by clear physiological and biomechanical theories and principles
• Quick and easy implementation using modern technology
• Fully Individualised programming
• Regular assessments of responses to training stressors, with prescriptions autoregulated on a sessional basis
• LVP can assess alterations across the full F-v curve, identifying areas of strength and areas for improvement, allowing for athlete classification (e.g., force or velocity dominant)
• Continuum of autoregulatory options is available to coaches, providing flexibility to suit any training environment
• Feedback can enhance intent and buy-in within a training environment
• Regulate volume using VL or VS

Challenges of velocity-based training

As with all S&C tools, VBT presents some challenges for coaches to consider. It is heavily reliant on technology, which can be difficult to implement in some environments or with large squads or groups of athletes. Additionally, the cost of equipment, connectivity troubleshooting, measurement error and erroneous repetitions or data can present problems for S&C coaches.

Practitioners can sometimes fall into the trap of “iPad coaching,” where they might find themselves too invested in the data than being present on the floor with the athletes. Additionally, some VBT practices can be time consuming, potentially eating into valuable training time.

Finally, without appropriate education and building the right habits, athletes can sometimes “chase velocities,” where technique suffers in pursuit of higher velocity.

Conclusions

VBT is an extremely versatile training tool that encompasses many practices across all S&C prescription, programming, and coaching. By utilising new and exciting technology, practitioners can objectively quantify improvements in force-velocity characteristics, autoregulate volume and load in accordance with physiological status, and create a competitive environment while generating buy-in and driving intent.

Even with all these benefits, VBT is not the be all and end all in S&C testing and programming. Coaches should implement it often as an aid to traditional methods such as 1RM testing and periodisation.