In a previous post, we discussed the importance of running speed to almost every team-sport athlete. Of course the ability to generate high top speeds is great in some instances, but when considering longer values of time (i.e. more than a one-out effort), the measurement of speed becomes a bit more difficult. When the activity involves constant stopping and starting, the average speed can give some general information regarding the intensity of movement, but is quite limited. Therefore, it is the ability to change speed rapidly (i.e. acceleration or deceleration) that is of far more importance to team-sport athletes. The figure below illustrates an example of the time course of a sprint effort from a rugby league player. It can be seen that during this maximal effort, a distance of around 20 m was required to reach what is typically considered high-speed running (5.5 m∙s-1 – yes we talk in metres per second, humans travel in metres per second, race cars travel in kilometres per hour!). Players are not often afforded such space in a game of rugby league, and therefore the ability to generate speed quickly (i.e. acceleration) may be more important.
Example of the time course of one maximal sprint in one team-sport athlete. HS = high-speed, 5.5 m∙s-1; VHS = very-high-speed, 7 m∙s-1.
Several studies have shown that acceleration efforts are energetically demanding, while the eccentric contractions that occur during deceleration may result in significant muscle damage. Knowledge of the true activity profile of matches can assist strength and conditioning coaches in designing conditioning drills and sessions, whilst this information can also assist nutrition staff plan their post-session recovery protocols. Global positioning systems (GPS) and accelerometers offer a simple and practical way to measure these actions during training and matches – provided the limitations of these technology are kept in mind. The market is now flooded with device manufacturers, and a mind-boggling number of acceleration-based variables that can be exported. Choosing which variables to use when prescribing and monitoring training drills can be intimidating, and our recent study in JSCR  attempted to clear the air a little bit around what’s actually useful and what’s not.
Let’s start with acceleration/deceleration counts. Before I start telling you how bad they are, let me begin by saying that they are really simple and an easy concept to convey to coaches, but unfortunately that’s about where it ends. The act of classifying continuous data into discrete bands is inappropriate and incorrect. Consider an example where one athlete wore two GPS devices simultaneously during a sprint. In this particular example, one device has measured a peak acceleration of slightly over the predefined threshold (3 m∙s-2 in this case), whilst the other measured slightly below the threshold, despite both units attempting to measure the same activity. In Unit #1 an acceleration event would be triggered, where Unit #2 would not, despite a non-substantial difference between the two units (~0.02 m∙s-2). The same could be said for two repeated efforts using the same device, because we know the reliability of GPS for measuring acceleration decreases with intensity [2, 3].
Further to this, the criteria that determines the classification of an acceleration or deceleration effort varies drastically between manufacturers, and often these companies refuse to be upfront about what needs to occur for an event to be triggered. A simple example of such criteria would be the minimum duration required over threshold to trigger an event. Let’s say the software requires a minimum duration of 0.6 seconds, such as the effort below. If the minimum effort duration was set at 0.8 seconds, this event would not have been triggered. Furthermore, if the athlete’s effort slipped below threshold for just a single data point, the effort could also be ignored.
This might not necessarily be an issue for day-to-day use, provided software/firmware updates aren’t performed that alters these criteria mid-season . However, if a new study is published where an acceleration profile for the sport I’m currently working in is presented, I need to be sure that the same units are used with the same software/firmware versions just for me to use these data in my day-to-day role, which is rarely the case.
As an alternative to acceleration/deceleration counts, several studies have suggested measuring time spent or distance covered during acceleration/deceleration bands. This technique definitely helps attenuate some of the problems that we see with counts, but definitely not all of them. We no longer have the issues associated with event “triggers”, because all data is considered in these cases. As such, we see a vast improvement in reliability using these techniques, but still poor reliability for high-intensity activity. In an attempt to improve the stability of these measures, practitioners may decide to select conservative intensity thresholds, such as 1 or 2 m∙s-2. Whilst this might increase the measure’s reliability, this come at the cost of being able to differentiate between activity that occurs at a high intensity, and the activity that occurs just slightly above threshold.
To add another complexity, let’s consider a centrally located player in a game of rugby league – a hooker for example. This player might not accelerate or decelerate maximally regularly if ever throughout a match, and will barely ever reach what is typically referred to as high-speed running. Does this mean his running is worth nothing? Definitely not. This position is required to perform constant accelerations and decelerations throughout a match, moving up and back off the defensive line, and accelerating out of dummy half to engage the defenders. Between my supervisor Grant Duthie and I, we decided that it is probably the density of this activity that is important.
Both Grant and I like to keep things very simple, so we decided to simplify this whole concept – what’s the point of putting all these rules in place and confusing the whole thing? We came up with the “average acceleration” metric, which was simply an average of the absolute of all acceleration data for a selected period. This measure is simple – it tells us how much a player’s speed has changed during a specific drill, period or match. This measure encompasses all accelerations and decelerations that occur, regardless of magnitude, duration or direction. Although this method classes all acceleration and deceleration activity together, which could be considered a weakness, having one simple acceleration measure to convey to coaches is always good.
Take the below example, where I have pulled some GPS for one player during two separate training drills. In the first case, the player completed a traditional conditioning session, which involved some high-intensity shuttle running, over a distance of 30 m. From the player’s speed we can calculate acceleration, and the output from the GPS provider tells us that this player completed 12 acceleration efforts over a pre-defined high-intensity threshold (3 m∙s-2), and 16 deceleration efforts of similar magnitude. It is reasonable to conclude that these numbers are around about the mark, as the acceleration intensity seems to fall away towards the back end of each of the 6 reps (but were still recorded as efforts but within a lower band). Using our average acceleration metric, we can see that on during this period, this athlete changed his speed by around 0.84 m∙s-2 on average throughout the drill.
Speed and acceleration during a traditional conditioning drill (6 x 30 s shuttles).
The second half of this figure shows an instance of a small-sided game, where the same player was only able to generate 3 high-intensity accelerations, and did not reach the required deceleration threshold once. It would not be unreasonable to assume that the first drill exhibited a greater acceleration requirement than the second, given only the “count” data. However, due to the density of acceleration and deceleration work in the second drill, the player changed his speed on average by 0.93 m∙s-2, substantially more than the first.
Speed and acceleration during a small-sided game (7 v 7).
In fairness to the manufacturers, GPS devices do a decent job of detecting acceleration/deceleration events when they are clear-cut and preceded by and followed by a bout of straight-line running, such as that occurring in traditional conditioning drills such as shuttle running. However, without even watching the player complete the drill, we already know how many acceleration and deceleration efforts will occur, based on the number of turns prescribed – the GPS device adds little to what we already know in this case. Alternatively, it is the stochastic, chaotic movements prevalent in team sports that we don’t know about, and need to quantify accurately. When we don’t know the prescribed distance or time, GPS are a great practical tool to do that – and the average acceleration technique seems best for measuring the stop/start nature of running within team sports.
As discussed in my previous post, the ability to accelerate and decelerate efficiently and effectively is extremely important to team sports. As such, the measurement of the magnitude of one-out efforts may be useful for quantifying abilities pre- and post-training program. Unfortunately GPS do not possess the accuracy to do so, so other technology are required for this reason. Instead, focusing on what GPS can measure allows us to more effectively prepare our athletes for the demands of competition.
- Delaney, J.A., et al., Importance, reliability and usefulness of acceleration measures in team sports. J Strength Cond Res, 2017.
- Akenhead, R., et al., The acceleration dependent validity and reliability of 10 Hz GPS. J Sci Med Sport, 2014. 17(5): p. 562-6.
- Buchheit, M., et al., Monitoring accelerations with GPS in football: time to slow down? Int J Sports Physiol Perform, 2014. 9(3): p. 442-5.