From fixed thresholds to individual acceleration-speed profiles
A prop and a winger may finish a match with a surprisingly similar number of accelerations on the report. That number, however, describes two very different games.
The count itself is accurate. What it cannot tell you is what those accelerations cost, because a single fixed threshold is being asked to describe efforts that begin from completely different starting conditions. A recently published study in Frontiers in Sports and Active Living examined 51 official matches of a professional United Rugby Championship side across two seasons, and put the question directly: how much of the mechanical load are we actually seeing when we count accelerations?
The same threshold, two different efforts
One of the most commonly used acceleration thresholds in team sports is 2.5 m⋅s⁻². For a given player, at a given moment, crossing it can mean three entirely different things: a small fraction of what they could produce from a standing start, close to the peak they can reach when already running at moderate speed, or something physically unattainable because they are already running fast.
What separates the three cases is the speed at which the acceleration begins. That relationship — between running speed and the acceleration a player can still achieve at that speed — is the individual acceleration-speed profile (ASP): from zero up to top speed, how much acceleration remains available. It is specific to each athlete, and it can be derived in situ, from the training sessions and matches players already complete, without adding a testing day to the week.
Once the ASP is known, a second reading of the same action becomes available: its intensity can be judged against what that particular player could have produced at that particular speed, rather than against a fixed value.
From profile to power: what a burst is
The approach works in two steps.
First, the ASP is converted into external power across the whole speed spectrum, and 80% of each player’s maximum is taken as the limit — a relative cut-off, in line with previous work on mechanical power in team sports, that isolates the upper tail of locomotor intensity.
Second, everything above that limit is measured: the work done, called high external work positive (hEW⁺, J⋅kg⁻¹), and the number of events, called bursts. hEW can also be broken down into force-, power- and speed-dominant components, providing additional context on how high-intensity work was accumulated.
FIGURE 1 – Left: the individual acceleration-speed profile (red line), together with acceleration-speed data obtained from maximal efforts initiated at different running speeds. The dashed horizontal line shows the conventional 2.5 m·s⁻² threshold, illustrating how the same absolute threshold represents different relative demands depending on running speed. Right: the corresponding external power profile; the dashed red curve represents 80% of maximal external power, used to identify bursts.
From Osgnach et al. (2026), Front Sports Act Living, CC BY 4.0.
What the data showed
The dataset covered 51 matches and 64 players, giving 1,172 individual player-match records, alongside 414 monthly in-situ profiles that met the inclusion criteria of at least nine sessions and 480 working minutes within the month.
The profiles themselves are instructive. Forwards and backs differ far more in maximum theoretical speed (7.94 vs. 9.17 m⋅s⁻¹) than in maximum theoretical acceleration (7.14 vs. 7.31 m⋅s⁻²). The heavier players accelerate nearly as hard; they simply do it over a shorter speed range.
Then the central finding. Read through three different lenses, the same matches give three different pictures of how far apart the two groups are:
| Forwards | Backs | |
| Accelerations >2.5 m⋅s⁻² | 0.44 / min | 0.50 / min |
| Bursts | 0.16 / min | 0.25 / min |
| hEW | 1.26 J⋅kg⁻¹⋅min⁻¹ | 2.94 J⋅kg⁻¹⋅min⁻¹ |
Counting accelerations above a fixed threshold, the groups look relatively similar — forwards produce 12% fewer events per minute. Counting individualised bursts, the difference becomes larger. Measuring the work performed above the power limit, it becomes much larger: more than double.
Same match, same players, three different pictures of positional demand.
Why rugby is not soccer
Comparing these figures with soccer data from the same research group produces an inversion worth pausing on. In soccer, bursts outnumber accelerations above threshold — a ratio of 1.64. In rugby, the ratio falls to 0.48 for backs and 0.36 for forwards.
FIGURE 2 – Bursts plotted against the number of accelerations above 2.5 m⋅s⁻², for each positional group. The black diagonal is the line of identity: below it, a player records fewer bursts than accelerations. Rugby forwards and backs (red and blue diamonds) sit consistently below the line; soccer defenders, midfielders and attackers (purple circles, from Osgnach et al. 2024) sit above it.
From Osgnach et al. (2026), Front Sports Act Living, CC BY 4.0.
The likely explanation is where the movement starts. Average speed across a rugby match is 69.2 m⋅min⁻¹ for forwards and 74.1 for backs, against 111.3 in soccer. Tackles, rucks and mauls interrupt play constantly, and set pieces require players to start moving repeatedly from near-zero speed. Starting from a standstill makes the 2.5 m⋅s⁻² threshold easy to cross without the effort being especially demanding. Cross-sport comparisons of this kind should be read with care — playing tempo, tactical organisation and match dynamics all differ — but the direction of the difference is consistent with how the two games are played.
What the approach cannot see
The framework describes running. It does not account for collisions, scrums, rucks, mauls or other set-piece actions, which represent a substantial share of total match load in rugby — an omission that matters most for forwards, whose overall load is therefore underestimated by locomotor metrics of any kind, external power included. The analysis is also descriptive rather than inferential, and deliberately concentrates on the positive phases of running, leaving decelerations outside its scope.
The external power approach is complementary to established GPS indicators, not a replacement for them.
What changes in practice
For a performance team already collecting GPS data, the practical implications are concrete:
- do not expect burst counts and acceleration counts to move together — in rugby they do not;
- judge intensity against each player’s own profile rather than a squad-wide threshold;
- update profiles from sessions already in the system, without adding testing;
- for forwards, treat contact load as a separate account that locomotor metrics will not close.
The value of the external power approach is not that it produces a better number. It is that it makes the question harder to avoid: when two players finish a match with the same count, what exactly is it that they did the same?
Author: gpexe educational team
References:
1. Osgnach C, di Prampero PE, Botter A, Maniera M, Morin J-B, Pavei G. Mechanical load monitoring in rugby: limitations and future perspectives. Front Sports Act Living. (2026) 8:1840846. doi: 10.3389/fspor.2026.1840846
2. Osgnach C, di Prampero PE, Zamparo P, Morin JB, Pavei G. Mechanical and metabolic power in accelerated running — part II: team sports. Eur J Appl Physiol. (2024) 124(2):417–31
3. Morin JB, Le Mat Y, Osgnach C, Barnabò A, Pilati A, Samozino P, et al. Individual acceleration-speed profile in-situ: a proof of concept in professional football players. J Biomech. (2021) 123:110524
4. Roberti V, Accelerations: what are you really measuring?, gpexe