Just in order to clarify concepts, let me show you a little piece of my thesis (and the way I'm showing all this work done with GPS).

My first idea before having the possibility of using GPS technology (I'm talking around 1998) was to be able to detect how much power (force and velocity) was applying a player in order to relate it to his own thresholds (individualize), and to know what the exertion of the player was, comparing training and competition.

Then it appeared the possibility to work with GPS tech, so the analysis I did was implemented in two ways (2006-07): the usual till today, based on distances, time in HR zones, speed and number of maximum sprints detected by speed and by acceleration (this depending on the speed shown when starting the acceleration), impacts, etc, and the second approach that was introducing the number of "actions", what in fact are number of accelerations of low threshold, arbitrarily putted in 1m/s/s (after some simple studies).

After that point, I tried to reorganize (2008-09) the knowledge I got about the way we were interpreting soccer conditioning loads (and other sports), which is based in physiological concepts, absolutely necessary for understanding how our body reacts to different stimuli (training), but in some way it is based in a concept that means intensity and exertion are the same (related between them), but this, in soccer, is not completely true!!!

The fact that intensity and exertion are synonymous comes from other sports with a different motor pattern: the continuous one. In this case, it's clear that more intensity (speed of displacement usually) is absolutely related with more exertion.

And this has led to many people to confound that when a player was not showing higher levels of speed, this meant that his exertion was "bad" or "less" ... and this is not always the case.

My personal approach is that as soccer players show an intermittent activity pattern, is it possible to analyze it depending on the characteristics of the two phases: action and pause.

Both of them (action and pause) can be defined by it's intensity and it's duration.

And this is the key of the specific approach, which help us understand not only what are the cause of the transitional phases of higher conditional exertion during the games (without the problem of the time where the ball is not in play), but also to understand the specific characteristics that differ between the specific soccer drills, and also help us to a better comprehension of all the methodology and also the possibility of implementing different periodization models based (apart from the coordinative and cognitive level) on the conditional exertion.

Arrived to this point, I have been using the acceleration value in order to identify each "action" (traditionally sprint), but this is not the "complete solution" as you know, done that sometimes in a game, the player don't show a significant acceleration but he gains speed, arriving to a high level sprint speed.

So from my point of view, based on the way I work, there exist two ways (at first) to better understand our knowledge about the specificity of our methodology in soccer training system: the first is related with the reference framework of analysis, based on the possibility of detect the intensity of actions and pause, and the duration of those actions and pause. This could be done implementing the GPS or video based software for this analysis.

The second is related with the way we identify an action (or a period of action), that actually is just based on speed or acceleration separately. Here is where the proposal of Di Prampero and col. (2005, 2009) shown in Osgnach and col. (2010) and developed by Colli and col. (they are working on this in a simple way but with a low cost GPS ... that will break the market!) appears as a "not perfect but a big solution" for this issue.

In this way, there won't be the necessity to analyze speed and acceleration separately, but just look for metabolic power threshold to identify periods of action and pause.

Enjoy