Sunday, July 22, 2012

Couple of insights from explosive lower body training with PowerTool/GymAware

As I have posted in the previous blog post I am currently playing with PowerTool/GymAware and aiming to increase my power output in squat jump, VJ and bench throw, along with trying to improve strength in squat and bench press using (a form of) dynamic effort method, during which I cut my set where my power drops below 90% of initial rep.  

Also, I am putting back Olympic lifts (clean, snatch and hang clean and hang snatch) into the program after a very long break. I am starting very conservatively and as you will notice my technique is not so ‘smooth’ and ‘quick’. 

On the bottom of the page you can find embed video. Please note that I am also starting video editing, so please bear with my crappy transitions between clips and no sound (at least no cheesy music background). 

I have started the training with the snatch. My technique (especially the catch) was ‘unconfident’ because I haven’t done snatches for a long time, even longer than I skipped cleans. On the second set on the video (actually a third or fourth in training) I kind of improved the catch (except the first rep where I sumo-ed it) and I finished with couple of overhead squats with a pause in the bottom. 

On the next clip I did some cleans just to finish up. Again, technique doesn’t look ‘snappy’ and I tend to ‘bounce it’ off my quads, as can be seen by the drift of the bar forward. I guess this will get a lot better with more practice and heavier weight. I have also finished with some front squat holds.

After the cleans, I wanted to test  my VJ using broomstick. The GymAware needs to be started from the set-up position (standing tall) so it can recognize this as the ‘zero position’ and calculate difference between this point and highest point. I have reached 49.1cm in later set (you need to trust me because I didn’t tape it). This is without arm swing so I guess it is a decent score. We have problems with the celling height and we don’t test tallest athletes because of the fear of hitting it, or at least fear influencing the score.

When it comes to countermovement jump with 20kg (empty barbell) I wanted to see how my depth will impact the power readings. Theoretically, going deeper will increase the work done (Force X distance) and if the time of movement is the same power will be higher. Unfortunately, it seems that going deeper affects eccentric-concentric transition (and ‘energy return’) and increase the movement time, thus decreasing power output. As can be seen on the video, on the 3rd and 4th rep and purposively went deeper in the squat and scored less on power output. On 5th and 6th rep I went pretty ‘shallower’ which resulted in more ‘snappier’ lift and thus higher power output. 

In the following table you can find the analysis for the each rep (the full GymAware app with web based software does this automatically - I don't have that one so I copy-pasted the values manually) 

Total system

Mean Power
Peak Power
Mean Power
Peak Power
Mean Velocity
Peak Velocity
Concentric time

As can be seen, I have changed the jumping style (dip) and scored different of the Height and Power measures. 

On the following table I did correlations for the variables:

If we take out fatigue out of equation, we can clearly see negative correlations between the DIP and the Power (both external, total mean and peak) and perfect correlation with CONCENTRIC TIME.

Height has pretty high correlation to power measures (0.70 - 0.72), and mean velocity (0,78). 

All power measures have almost the perfect correlation between themselves (as can be seen by gray area). BTW, Total system is when I take my BW into equation.

This creates some problems in measurement and evaluation of the players.

First, taller players will score higher on the power readings (anyone have correlates between height and VJ or power output?) since they do more work.

Second, it is hard to do inter-individual comparison without controlling the depth in some way. Depending on the individual style of jump, some players will produce more power than others. With the data I have at the moment, I don’t see any correlation between P20kg (external power and NOT total system power – which brings me to the topic of different power calculations) and VJ performance (and not even correlates with 10m, 505 or any other performance test; but let’s wait for more samples). This might mean that the individual might change the jumping style to reach highest VJ or to reach highest power output readings. Then it might be important to think what kind of feedback are we giving to the players – power readings or vertical height reached (it might be smart to tell them to jump highest without reporting power reading since this might influence the power score for the guys who have some physics background by adapting their jumping style:) )

It is probably ok to compare the intra-individual performance (for example changes in my power readings over time) because of the same jumping style, but I question its use to compare individuals (in terms of power) WITHOUT some way of controlling the depth (dip). I would really love to hear reader’s experience with this, along with possible studies on this (VJ and P20kg correlates). It might be interesting to see correlates between power output using different methods (external power, total power, relative external, relative total, etc) during free countermovement jump, regulated countermovement jump (certain depth/angle) to VJ (and 5-10m sprint as well),  and to power outputs within different ways of measurement. If someone know a study like this, or studies, please let me know. Dan Baker sent me his study on correlates between Power and 10m-40m sprints. I will try to find and read some studies on these issues.

When we did testing at the beginning of the pre-season we used Smith machine and the players jumped from pre-calculated height where they take 90 degrees in the knee. But that means concentric only. If I remember correctly the researchers used total body power (they used 90% of BW plus 100% BW on the bar) dived my body mass. Yes, power measurement is confusing (I suggest the readers to check the papers by Cormie et al.). So, I suggest when someone discuss power training, correlation or anything one needs to ask the following questions:

1. How was the power measured – using 1-LPT, 2-LPT, force plate, force plate and LPT(s), COM analysis using kinematic data

2. How was the power calculated – using external work only, or taking BW into account, was is it peak or mean

3. What type of jumping was done – static, countermovement, arm-swings, with regulated starting position, with unregulated starting position, what kind of feedback was given and what was the goal of the jump (height or power output)

4. How much resistance is used and how it is calculated (as % of 1RMs, as % of BW, etc)

For example, with total system power measurement, peak power is reached with 0% 1RM external resistance. With external power measurement, peak power is reached with 40-50% 1RM (if I remember correctly).  All these makes discussion on power confusing.

We tend to focus on numbers and forger to ask about how did we (or others) actually got them - what type of measures we did, how did we impact those measures with tasks, contexts, feedback and biases. Sport performance, be it power or anything else is not easily represented by number.

I am also posting with permission the email response by Dan Baker when we exchanged couple of emails on the topic of power measurement. I wish coaches can hear more real-life training wisdom from Dan and I wish one day he write down a very practical book. Whole this year I was bothering Dan with questions and advice requests and I must admit that I have learned more by reading his email responses than from the most books. I want to thank Dan one more time. 

The jump squat question has been raised and is raised all the time.  Technically you should add body mass  (or some portion of it, again another debate) to the equation and I used to, but I don't now as I will explain.

Technically, yes body mass + jump squat mass.  I will attach some of Cormie's work with force plate + 2 LPT shows that peak power occurs with 0 extra load. during JS,  - but even she doesn't add all the body mass (minus 12% for weight of shins and lower body "not moving" when the calculation is made….so confusing!

Also Peak power v Average (mean) power.  I always use mean power, peak power will always occur with the lowest resistance, even in upper body studies.  I used just 1 LPT in my studies, so in Pru's work you can see that she gets similar results when using only 1 LPT, but gets different results based on LPT + force plate etc.

So I now just use the power generated externally, as the athlete always has to move his body anyway ~ thus I don't include body mass anymore for monitoring training and just use the 1 LPT.

If we train athletes over time and their body mass changes, am I getting real changes in leg power or changes due to heavier athletes?  I was getting this problem a lot as some young players can change 10 kg of body mass in a year, from 17 to 18 or 19 yrs.  

So now I just weekly monitor with 20 kg and train and test with 60 kg and 100 kg (but we train the whole spectrum, but mainly 60-100 kg).  But we don't add body mass to the calculations.  

So I am looking at External Power output.

I think of it like this, as a spectrum 

20 kg JS
60 kg JS
100kg JS
1RM Power clean (hang)
mean of 125 kg
1RM full squat
mean 177 kg

So by looking at these simple tests, I know about unloaded leg power (VJ),  and strength (1RM squat) and everything in between.  If someone has good VJ, but poor power with 60 kg and 100 kg JS, then they will be weak in the squat (and also in the PC)

Conversely, strong in the squat and poor in the 60 and 100 kg JS power, they will be slow in VJ and 20 kg JS.  You can do the VJ as a CMJ with a broom handle and the GymAware so you take the confounding effect of arm swing out.

By leaving body mass out of it, I take away that confounding variable in their JS power.  Just helps with truing and monitoring long-term effects, but for uni. studies, you may have to include the portion of body mass like Cormie did.

For soccer, it may be CMJ, 20 kg, 40 and 60 kg.  Certainly the Aust. Instit. of Sport have found that JS with 40 kg most useful for basketball and volleyball players, but for rugby league, resistances that equal our opponents body mass (100kg) are a better indicator.  So just use absolute external loads for testing and monitoring and any loads for training.  You will work it out pretty quickly.  

My general rule of thumb is use 40-60% of the weight you used for strength training that week.  So if your athlete used 80 kg x 5 reps for squat atg, then do jump squat with anywhere from 32 to 48 kg .  But if you have a group of players of similar strength, say squatting 80 to 100 kg x 5, then it is easier and more competitive to have them all do 45 kg for jump squats x 5.

Back to video – after countermovent jump with 20kg, I did planned squat jumps from static position. I have included the one with 40kg. As the last set I did continuous (rhythmical) jump squats with 30kg (to also utilize some PAP/contrast after jumps with 40). Interestingly, when I did static squat jumps with 30kg I created 686W (external), but when I did rhythmical ones I created 857W, or 24% more, which might be accounted to eccentric/elastic energy return. 

After all those jumps, I proceeded to do squats at 75%, 80%, and 85% of 1RM. I have repeated this twice, plus I did one more set with 75% for 4, equaling to 7 sets of squats. Theoretically, I tried to squat down slowly, hold the bottom position for a second and then try to lift it up as fast as possible. I think I lack the ‘groove’ at the moment and I cannot do it with the ‘snap’ (talking about doing nothing for a week or two). I wanted to do each set until my speed dropped below 90% of initial rep. The problem I had was that GymAware cannot set alarms and thresholds based on percentages (or at least I don’t know how to do it), but rather absolute numbers (for example, it warns me when my velocity drops below 0,5 m/s), so I was playing to see how much reps are actually above 90% and they are 3, 2, 1 for 75%, 80% and 85% (anyone see similarities with Prilepin chart? ). I guess I need bigger increments for squats compared to bench press (where I managed to do 4, 3, 2 with 75, 80 and 85%), something like (70, 77.5, 85). Trial and error people. You learn from the mistakes :).

Funny thing is that I felt more fresh with 7 fast sets than with 2-3 sets close to failure I usually do (my joints don’t seem to like ‘grinding’ sets, and all those incremental programs approaching failure  over weeks usually result in minor injury in my case). I hope those fast, large in number technical sets will yield dividends at the end in terms of improved power and strength.


I want to thank my friend Joel Grandell from Hammarby Boxning for helping me make these videos.

Hope this was insightful to you as it was to me. Thanks for reading.


  1. Dear Mladen,
    Have you thing to apply to your strength training schedule the specificity of the soccer inttermitence? This means a short of cluster method, with very few maximal power reps but with few pauses (10"-20").
    Kind regards

  2. You should check out this paper:

    Lake JP, Lauder MA, and Smith NA. Barbell Kinematics should not be used to estimate power output applied to the Barbell-and-body system center of mass during lower-body resistance exercise. J Strength Cond Res. 2012;26(5):1302-7.


  3. Hi Manuel,

    Although clusters are interesting method, I don't think they are needed for average soccer player, especially not to improve 'specific intermittent endurance'. Clusters are used to improve strength/power performance in more advanced strength & power athletes. There might be a short period where we might use them with team sports, but I wouldn't put any special emphasis on them.

    When it comes to combining strength and endurance, to create some form of 'sport specific' strength training I am more than against of such practices.

  4. Thanks Naoki. I will check that one out!