Velocity Based Strength Training - PUSH Band Trial by Eamonn Flanagan

There have recently been a raft of online articles on velocity based strength training and its potential application in optimizing and refining the strength training process. Some time ago, myself and Mladen Jovanovic of Complementary Training wrote a review article for the Journal of Australian Strength and Conditioning. That article covers some of the key scientific research validating velocity based strength training and outlines some potential practical applications for athletes and coaches. That article can be viewed here.

The area of velocity based strength training continues to develop and there are a number of commercially available devices. Carl Valle in particular has done a great job in outlining the range of monitoring devices that are commercially available. His “complete guide to bar speed trackers” is well worth a read. These devices typically have been linear position transducers such as the Tendo unit and the GymAware device. Linear position transducers consist of a central processing unit that attaches to the resistance training equipment (such as a barbell) via a retractable, measuring cable and via a data display unit (or smartphone) to measure bar speed velocity. More recently PUSH have launched an accelerometer based alternative to the linear position transducer. PUSH is a wearable device which is worn high on the forearm via an armband.

I’ve been lucky enough to have been trialling the PUSH band in recent months. I’ve used it extensively in my own training and intermittently with some athletes in an attempt to get a feel for the device and to try and assess its validity, reliability and its practical usefulness in strength and conditioning. The aim of this article is to outline how the PUSH band measures up against some of the key pillars of what the scientific literature has established with respect to velocity based training. This article will hopefully outline what are the known load-velocity relationships and trends we should see in training. And, whether the PUSH band measures and presents these relationships in a manner we would expect.

This article won’t attempt to outline the “whys” and “whats” of velocity based strength training. Smarter people than I have already done that. Recently, Carl Valle, Martin Bissinger and Bryan Mann have produced some great online insights into the usefulness and practicalities of using velocity based strength training in high level strength and conditioning (Link 1; Link 2; Link 3). Nor will this article attempt to comprehensively validate the PUSH device against other products. But hopefully by describing the process by which I have tried to apply and test the PUSH band in a practical setting, it will give readers some insight into some of the applications of velocity based training and the usefulness of the PUSH band.

The Load-Velocity Relationship

In key strength training exercises such as the squat and the bench press, there exists a very strong relationship between load and average concentric velocity of lifting (note: for the purposes of this article, all mentions of “velocity” refer to the average velocity in the concentric phase unless otherwise indicated). Load and velocity are very strongly, inversely correlated. Add more weight, and lifting tempo slows down. This isn’t news to anyone who’s spent any time under the bar but the strength of the relationship continues to surprise. The robustness of the relationship is up there with exercise intensity vs heart rate in terms of its strength and measurability. Whether measured in a laboratory setting or in the gym with a commercially available LPT device, correlation coefficients of > 0.95 are commonly observed (Figure 1). 

Figure 1

pic 1.JPG

Establishing a load-velocity profile is one of the cornerstones of velocity monitoring during strength training. Our article in JASC outlined the following guidelines for establishing the profile in exercises such as the bench press and the squat (Figure 2)

Figure 2

Figure 2

For me, establishing a load-velocity profile is the first test of a new velocity measuring device in a practical setting. Does it show us the same, well established relationship, and the same strength of relationship that we would expect? When testing the PUSH band during the bench press and the squat on myself and a number of athletes, it did pretty well in this area. Correlation coefficients of > 0.95 were commonly observed across 4-6 load data points (when intensity of effort and recovery times were closely controlled. Very rarely an “outlier” datapoint at one load would crop up that could slightly throw off a profile (and reduce correlation coefficient to 0.8) but this was rare and probably only happened once or twice in dozens of measurements. A good coach can recognise this “noise” when it happens and not allow it to interfere too much with the process. Overall, the PUSH band gives a very nice, reliable load-velocity profile (Figure 3). 

Figure 3

In these exercises, bench press and squat, once we develop the load-velocity profile, it can be used to predict 1RM. This can allow coaches to potentially assess 1RM via sub-maximal loads lifted with maximal intensity of effort. Maximal load attempts are associated with a specific velocity that is termed the “minimal velocity threshold (MVT)”. The MVT is the mean concentric velocity produced on the last successful repetition of a 1RM or a set to failure performed with maximal lifting effort. It appears that when lifting with maximum intensity of effort, there is a minimal velocity that needs to be achieved for the repetition to be successful. MVT has been shown to be exercise specific with approximate mean concentric velocities of 0.15 m/s for bench press and 0.30 m/s for back squat being reported. Interestingly this MVT is consistent within subjects regardless of relative load (Figure 4). 

Figure 4

Research from Izquierdo et al. showed no statistically significant difference in “Minimum Velocity Threshold” in a reps to failure protocol across a range of sub-maximal intensities and measured actual 1RM.

Similarly, with the PUSH band, I found this MVT to be similar to the published data of approximately 0.3m/s for the back squat and to be stable across a range of loads and rep ranges when lifted towards fatigue. The chart below (Figure 5) is for a single subject only and shows a number of different loads lifted “towards” failure in different training days. The reported terminal velocity is a fraction higher than has been shown in the research, but this is probably accounted for by the fact that the loads weren’t lifted to absolute failure. Regardless, the trends observed and the data recorded is very similar to what research has established and what we’d expect to see.

Figure 5

As this terminal velocity is stable. It can be used to predict 1RM and track changes in maximal velocity and maximal strength over time (Figure 6 & 7).

“Creating a load/velocity profile in an identified key training exercise allows coaches the opportunity to track an athlete’s progress, over time, across a spectrum of velocity demands. This is particularly applicable for coaches who are interested in velocity specific adaptation to training and not solely focused on maximal strength development. Creating a load/velocity profile also allows coaches to compare athletes against each other across the velocity spectrum. Load/velocity profiling can also be used to predict 1RM values using sub-maximal loads. This may be of great interest to coaches and athletes if true maximal testing is not viable or appropriate. Measuring velocity with sub-maximal loads during sessions allows coaches to estimate daily 1RM values which can be used to assess the efficacy of training programs and the training status of the athlete.” (From Jovanovic & Flanagan, 2014).

Figure 6 & 7

The research of Jidovtseff et al. (2011) is worth noting here. These authors found that predicted 1RM based on the load-velocity profile was highly correlated with tested 1RM. But that the absolute 1RM values were different. This suggests that predicting 1RM based on the load-velocity profile is highly appropriate to measure training induced adaptation but values shouldn’t be used interchangeably with absolute tested true 1RMs (Figure 8). 

Figure 8

One of the attractions of the PUSH device is undoubtedly its price point. It costs significantly less than established brands such as GymAware and the Tendo unit. While these devices have shown good validity and reliability in the published research, PUSH has not yet been subject to published peer-reviewed assessment of its validity and reliability during barbell exercises (although it has been validated with dumbbell exercises versus a motion capture system - here).But I took the opportunity to get a glimpse at its validity versus the a linear position transducer unit. Ultimately to establish the reliability and validity of a device such as PUSH, it needs to be tested against known gold standard of velocity measurement such as 2D or 3D biomechanical video analysis. But in the absence of that, a comparison versus an established measuring device in the marketplace could give some insight. I used both devices for three repetitions over 4 ascending loads in the back squat (12 paired datapoints). The measurement with both devices were very highly correlated (R2 = 0.95) but the LPT device tended to record higher absolute values across all intensities (a small effect size in difference; ES = 0.36).  

Figure 9

This is a pretty crude way to examine validity and it highlights the need for new devices on the market to be put through rigorous testing. But with such a big price point difference between PUSH and LPT devices, it’s good to see the more affordable model giving very similar results. The difference in absolute velocity between an LPT and PUSH gives pause for thought however. As a coach, this shows me that I can’t take all my historical data from LPT devices and assume it will “transfer” or be interchangeable with the PUSH band. While their captured data may be highly correlated, in absolute terms it is different. So caution is needed here and for accelerometer based devices there may be a need to establish their own load-velocity profiles and minimum threshold velocities in the scientific literature.


As an alternative to LPT velocity monitoring the PUSH band has real promise. It appears to consistently track the trends we would expect to see in velocity based strength training. Load-velocity profiles are robust, minimum velocity thresholds are relatively consistent with the literature and the device, at first glance, appears to correlate well with other LPTs on the market. But while this is a very limited case study and much more stringent lab based testing is needed to really validate the PUSH band, my initial trials with the device were promising, especially considering the price point difference between PUSH and other on the market LPTs. One limitation of the PUSH band currently is that feedback isn’t quite “live”. Scores are presented at the termination of a set, rather than instantaneously rep-by-rep during a set. This is a significant limitation, as augmented feedback of training (especially training with a velocity or power-output focused training goal) is one of the major advantages of velocity based training.  But I’ve been assured by the good folks at PUSH that live feedback is in development and on the way. An exciting, possible advantage for PUSH over existing LPTs is the fact that the device is wearable which opens up the possibility for measuring velocities in ballistic training exercises using implements. Being able to accurately measure outputs in explosive medicine ball throws, slams and shot-puts alongside heavy resistance training with the same device is an exciting possibility and to my knowledge, not something that is available elsewhere. This is an area where accelerometer based devices could add real value to the realm of velocity based monitoring of strength training.


Izquierdo et al. Int J Sports Med. 27: 718–724. 2006

Jidovtseff et al. Journal Str. Cond. Res. 25: 267-270. 2011

Jovanovic & Flanagan, Journal Aus. Str. Con. 21: 58-69. 2014


Eamonn is a strength & conditioning coach for Irish Rugby, a sport scientist (with numerous publications) and a pretty solid Olympic weightlifter. Follow Eamonn on twitter - @eamonnflanagan.