Some great findings and ideas from velocity-based strength training
I have been thinking more and more about velocity-based lifting recently as a method of prescribing load and volume for an individual. I have wrote couple of blog posts and article on this topic that you might want to read first:
The mentioned links should cover the bases. What I want to do now is to discuss possible applications and some interesting findings.
I have re-reading great paper by Izquierdo et al. Effect of loading on unintentional lifting velocity declines during single sets of repetitions to failure during upper and lower extremity muscle actions and I came to couple ideas.
What the researchers did is to perform Bench Press and Parallel Squat exercises to failure with 60%, 65%, 70% and 75% of 1RM, while trying to perform each rep as fast as possible.
These graphs are very interesting and I will get back to them, but first I want to convey some of my ‘insights’ using velocity measurement in the gym over the last year and something.
Here is the load-velocity table for my ATG pause back squat and pause bench press I did somewhere in February this year. This is based on the first rep for each load which is usually the fastest.
I we visualize %1RM used in bench press and squat and mean velocity reached, we get the following graph:
What is immediately apparent is that the slope of the curve for bench press is steeper. We can calculate those and we get -1.46 for bench press and -0.89 for squat. In plain language, one tend to lose more speed with increasing loads in bench press than in squat.
Another apparent feature is that velocity at 1RM (we are going to call it MVT – minimal velocity threshold) is lower for the bench press (~0.1 m/s) than for squat (~0.3m/s). We can talk all that why this might be the case – but at the end of day it is not really important. What is important is to remember that there is different MVT for every exercise and that they differ from individual to individual (not much thought).
What might be interesting to find out is weather MVT changes when some improves his 1RM? According to study by González-Badillo and Sánchez-Medina (Movement Velocity as a Measure of Loading Intensity in Resistance Training) it is not changing over time, at least for bench press
This is important for couple of reasons – we can use velocity to prescribe intensity. Even more important is that 1RM might vary from day to day due readiness or normal variability, thus using %1RM might be misleading and not taking into account improvements or decrements in strength. Using velocity might solve this issues, along with being auto-regulatory in nature. Long story short, instead of prescribing 5 reps with 80% 1RM, one might prescribe 5 reps with 0.5m/s starting speed. What is also interesting is that providing immediate feedback might increase motivation, competition, stability of performance and higher improvements – based on the studies by Randell et al. (study1, study2), at least for jump squats – but I believe that this might be true for non-ballistic exercises as well.
Ok – this is what happens to the velocity of the first (best) rep across loads. But what happens to velocity when we repeat sub-max sets to failure? That is also what the study by Izquierdo et al ought to find out.
What we can see from the graphs (see at the beginning of this article) is that the velocity across reps is falling down quicker in the bench press than in squat. I will get back to this soon and why is this important.
What is VERY INTERESTING is the finding that mean velocity in 1RM load is pretty much the same as Mean velocity in the last rep of nRM test. What does this mean is that my last rep in 5RM is probably going to be very close to 0.3m/s for squat and 0.1m/s for bench press. How can we apply this to practical settings? Well, the closer we come to our MVT in multiple-rep sets, the closer we are to failure. According to a study I blogged about here, the closer we are to failure (indicated by loss of velocity) the higher the neuro-muscular fatigue. Hence, by monitoring last rep velocity we might produce different levels of fatigue in a given set.
I am interested to see if this prediction holds true across loads (or reps-per-set), or if I have 2 reps in the tank with 12RM load would the velocity be same as when I have 2 reps in the tank with 5RM load? Since I don’t have data for this, I ought to digitalize the data point from the graphs in this study. Here is what I got for bench press
And for squat
What is interesting to note here is that number of reps done with same %1RM is higher in squat than in bench press.
As you can see from the graphs this relationship between reps left in tank and velocity is sound (it is beyond me how to quantify magnitudes for this – my statistic knowledge is medium); average SD for velocity across reps left in the tank is 0.02 and average %CV is around 5% for both squat and bench press.
What this means, and is VERY INTERESTING, is that we can estimate proximity of failure based on rep speed (taking into account that the effort to lift fast should be 100%). Not with 100% accuracy, but pretty close.
I have mentioned that velocity drops a lot faster across reps for bench press than for squat. This is important since certain authors in velocity-based strength training circuits recommend using % drop as a threshold to stop a set. For example, they prescribe starting velocity (e.g. 0.5 m/s) and velocity drop of 10% (stop doing set when velocity drop for more than 10%, and in this case that is 0.45 m/s) for both bench press and squat. Based on the data I have presented I think this is not that smart (I have tried it also, with me and with some players once). It works for bench press, but for squat you might end up doing a lot more reps (especially if there is a bit of bouncing in the hole). Over aprox. 75% 1RM velocities for squat are faster (see bench vs squat load-velocity graph), his velocity drop across reps is slower and hence using % is not the way to go.
The solution might be prescribing absolute velocity stop. For example starting with 0.45 and finishing when velocity reach 0.4 m/s.
There is still a lot of practical trial-and-error left to be done, but here are some recommendations
Estimate load-velocity curve for each individual and each core lift when you do 1RM testing
Estimate MVT (or velocity at 1RM or last rep) for each individual and each core list
If you perform 3-5RM test, estimate the slope of the velocity curve – this might be used later to predict proximity of failure and velocity stops (when to stop the set).
Use associated velocity with certain nRM and/or %1RM (e.g. 0.5m/s for 5RM or 85% 1RM; 0.3 for 1RM or 100% 1RM) and use the velocity to prescribe instead of nRM or %1RM because it is more reliable plus you get a lot more other benefits
To manage fatigue in the set, prescribe velocities stops (should I put the trade mark on this one?) for certain reps-left-in-the-tank.
Manage volume (number of sets) by time allotment (e.g. 20min for squats); or by comparing the average velocity of the sets across reps (need to ‘research’ this approach); or by prescribing number of sets based on cycle; or by allowing certain drop in reps from set to set.
When doing multiple sets, one could stick to the weight associated with certain velocity intensity and velocity stops (in this case that might result in drop in reps across sets) or one might decrease weight to maintain starting velocity. I am clueless what method to use – might depend on the cycle or be rotated from time to time.
So the prescription might be something along these lines:
20min time allotment slot for squats (or prescribed number of sets)
Starting velocity 0.4m/s
Keep that weight across sets
Velocity stop 0.35 m/s
If your first rep is equal to or less than 0.35, stop completely even if time is still available
Damn, I am becoming to sound like DB Hammer
Anyway, I urge coaches to try velocity-based approach. Make sure to check the best LPT system on the market today – GymAware. Data collection and analysis is walk in the park, along with the setup.