Part of the Society for Transparency,

Openness and Replication in

Preprint
not peer reviewed
Kinesiology (STORK)

Exposures to near- For correspondence:

[email protected]

to-maximal speed
running bouts
during different
turnarounds in
elite football:
association with
match hamstring
injuries
Martin Buchheit1,2,3, Maxime Settembre1, Karim Hader1 and Derek McHugh1

1
Kitman Labs, Dublin, Ireland;
2
Lille OSC, Lille, France;
3
HIITscience, Revelstoke, Canada

All authors have read and approved this version of the Author Martin Buchheit
manuscript. This article was last modified on July 17, 2022. @mart1buch can be reached on
Twitter.
Please cite as: Buchheit M, Settembre M, Hader K and McHugh D. (2022). Exposures to near-to-
maximal speed running bouts during different turnarounds in elite football: association with match
hamstring injuries SportRχiv.

ABSTRACT
Purpose: To examine the association between the occurrence of near-to-maximal

sprinting speed (near-to-MSS) running bouts during training and hamstring injury

rate during the consecutive match of the same turnaround in elite football (soccer).

Methods : Retrospective data from 36 team-seasons (16 elite teams performing in

top European leagues) were analysed (627 players, 96 non-contact time loss match

hamstring injuries).

We examined 1) the association between match hamstring injury rate and the

occurrence of >85%, >90% or >95% MSS exposures or not during training within

each turnaround and 2) whether the above-mentioned associations differed

depending on the day(s) of the turnarounds when these exposures occurred.

Results: The longer the length of the turnarounds and the lower the speed

thresholds, the greater the number (and proportion) of near-to-MSS exposures. For

some turnarounds, there were no match hamstring injuries when players were

exposed to running bouts >95% MSS during training vs. when there were no or

lower relative speed exposures (i.e., >85 or >90%). Finally, irrespective of the

turnaround length, there were no match hamstring injuries when >95% MSS

exposures occurred at D-2, while in contrast, injuries still happened when players

were not exposed at all, or when these exposures occurred at D-3 and/or earlier

within the turnaround.

1
Conclusion: While the present observational study design precludes the

examination of causal relationships, programming >95% MSS exposures at D-2

during 3-d to 8-d turnarounds may be a relevant strategy to decrease the incidence

of match hamstring injuries in elite football.

INTRODUCTION

Hamstring strain injuries remain the most prevalent time loss injuries in

professional football.1 While their relative occurrence may have slightly decreased

in relation to the likely increased match demands over the past decade,2

practitioners are still seeking mitigation strategies both in the gym and on the

pitch.3 Among the different recommended strategies, the use of eccentric-based

exercises4 and exposures to near-to-maximal sprinting speed (near-to-MSS) running

bouts (either with or without the ball) are now the most recommended.5,6 Sprinting

is indeed both complex and unique at many levels (e.g., legs interaction, elastic

energy transfer, reflexes, kinematics, kinetics)6 and a similar recruitment intensity of

the hamstring muscles (i.e., electromyographic activity) cannot be reached with

isolated gym exercises.7

In practice, recent studies have shown relationships between hamstring strain

injuries and near-to-MSS exposures both in Australian Rules Football8,9 and Gaelic

Football10 players. More precisely, both under- and over near-to-MSS exposures

were associated with the higher injury rates, suggesting the existence of an optimal

chronic “dose” i.e. number of weekly exposures 8,10 and/or monthly cumulative

distance.9 This optimal chronic dose is likely specific to each population and

2
context, and it is therefore difficult to provide guidelines for all practitioners on the

back of those two studies. More importantly, those studies do not provide clear

guidelines on how and when to program these near-to-MSS exposures during

turnarounds of different lengths. How fast football players should run is also still

unclear, since large variations in relative velocities have been reported, ranging

from ≥80,9 to 858 or even 95% of MSS.10

In fact, the question of the optimal intensity (i.e., >85% vs >90% or 95% MSS) and on

which day to program these near-to-MSS exposures is something that has not been

examined scientifically despite its immense importance in terms of match

performance and hamstring injury management.11 The only partial answer to this

question that is available to us today comes from the 100 elite football (soccer)

practitioners that we surveyed in 2021.12 While the large majority of the responders

confirmed the need to regularly expose players to these high-speed running bouts,

there was a lack of agreement as to when MSS work should be programmed,

especially whether it should 2 or 3 days before the match i.e. D-3 vs D-2. This was

likely due to the lack of robust evidence, and this programming practice was rather

based on experience and/or adherence to typical periodization paradigms and

models (e.g., tactical periodization, R. Verheijen, or el modelo estructurado of FC

Barcelona).

In order to shed light on this important topic, we examined retrospective data from

19 elite teams performing in top football leagues across the globe. We first

examined the association between match hamstring injury rate and the occurrence

of >85%, >90% or >95% MSS exposures or not during training within different

turnaround length. The second aim was to examine whether the above-mentioned

3
associations would differ depending on the day(s) of turnarounds when these

exposures occurred.

We more precisely also looked at the timing of these exposures within turnarounds

of varying length. We believe that the information provided will help performance

staff support managers to optimise the programming of their microcycles within

their own context.

Methods

Study design and procedures

The overall research was based on retrospective analyses of both match hamstring

injury occurrences and players’ training locomotor (running) activities collected via

an online database (i.e., Kitman Labs platform, Dublin, Ireland) commonly used by

all the football (soccer) teams involved in the study.

Data extraction and anonymity

Each player and club is provided with an ID number on the platform. The researchers

in charge of the analysis could only pull and analyze data associated with their IDs -

no names included. Then, data was transformed and coded for injury occurrence

(dates only used for assessing occurrences, such as during a match vs during training

and when in relation from/to the previous match) and type (contact or non-contact

injury, without any more details), to provide a final dataset. The medical staff of each

team registers injury details in the platform as a part of their daily player care

4
management, including variables such as date of injury, type and location of injury,

as well as severity (days lost). Similarly, player game and training session

participation are recorded as part of the team staff’s daily monitoring. Additionally,

the measures of training and competitive load are also added to the platform. The

fact that all clubs used the same platform ensured the standardisation and the

reliability of all types of entries, from medical information to exposure measures

(e.g., session duration and GPS data attached to the system calendar). We

nevertheless ran a thorough data health check to ensure that all data retained for

analysis met the same standard. In addition to all the steps above that guaranteed

high levels of both data security and anonymity

(https://www.kitmanlabs.com/privacy-security-and-compliance/), permission was

granted by the teams for their inclusion in this research study, therefore ethics

committee clearance was not required.

Population

The data belonged to 19 different teams competing in the EPL, the Italian Serie A,

the French Ligue 1, the Bundesliga, the Scottish Premiership, the MLS and the

Dutch Eredivisie from January 2018 to December 2021. Team-seasons for which

injury information was not accessible were not used for analysis. Likewise, when

there was not enough information about players in the platform (e.g. exposure for

less than 15 players over the entire season) or insufficient locomotor tracking data

(e.g., daily maximal speed not defined or provided), the team season was not

included. This initial sample represented 84 team-seasons.

5
Injuries

Injury is often defined as an occurrence sustained during either training or match-

play which prevents a player from taking part in training or match-play for 1 or

more days following the occurrence.13 More precisely, in this study we focused on

non-contact hamstring injuries (i.e., a substantial strain of either the biceps femoris,

semitendinosus, or semimembranosus muscle), as registered by the medical staff

of each club, using the Orchard Sports Injury and Illness Classification System

(OSIICS) offered by the online platform. In this study we wanted to focus on

hamstring injuries that substantially impact training and match participation and so

only considered injuries that caused a minimum of 3 days of training/playing

interruption i.e. ≥3-day time loss. In fact, we excluded all mild injuries (<2 days lost)

because injuries in this category could conceivably not have an impact on the next

game availability or training dynamic within the same turnaround. In addition, this

choice has allowed us to exclude days lost due to potential training removal as a

result of player management, as it sometimes happens in clubs. If the medical staff

registered injuries from the start to the end of the season, we assumed that they

strictly adhered to this practice throughout the whole season, and that there was

no missing data for this metric in this situation.

Turnarounds

A n-d turnaround was defined as a microcycle with n days between the first and

second match, where n is the count of days from the first day after a match up to

and including the following match day. The shortest observed turnaround was 3

days (3-d) e.g. playing a match on Sunday and again the following Wednesday, while

6
the longest was 8 days (8-d) e.g. playing on Saturday and again the following

Sunday. The longer and less common turnarounds were excluded from analysis. In

total, 1358 turnarounds (at the team level) were extracted and were grouped by

their respective length.

Turnaround participation and injuries

A player was considered as involved in a turnaround if one or more of the following

three conditions held:

1. An injury was registered for the player at any point during the microcycle,

2. He played the match that ended the current turnaround or,

3. He had participated in at least one field-based training session, in the case of

3- to 5-d turnarounds, or two field-based training sessions, in the case of 6-

and 8-d turnarounds, within that turnaround, where field-based training

sessions were based on whether GPS readings were recorded or not.

Each individual player's involvement in a turnaround was considered as a player-

turnaround. Player-turnarounds in which any injuries occurred, other than non-

contact time loss hamstring injuries, were removed from the analysis. At the level of

data preparation, the final data set included 36 team-seasons, including a total of

667 players, 1581 injuries including 229 non-contact time loss (> 3 days missed)

hamstring injuries, 1495 non-international matches and 6698 training session days.

Near-to-maximal sprinting speed exposures

The maximal sprinting speed of a player was calculated based on the available data.

The ideal scenario was when a club was actually testing for MSS, and in this case

7
the resulting MSS was used for analysis. When proper testing data was not

available, we used the average of the three highest speeds reached in the entire

GPS data set of each player (after having manually removed all possible erroneous

data >37 km/h).11 Following both the research literature8-10 and actual sport science

practice in the field, we marked when near-to-MSS exposures occurred using

individual speed thresholds at >85%, >90% and >95% of each player’s MSS.11

Near-to-MSS exposures during turnarounds

To understand the frequency of near-to-MSS events, we labeled all individual player

training sequences leading to a match by the turnover value and added an indicator

as to whether or not a neat-to-MSS event had occurred. We coded the entire

individual player training sequences leading to the match (as a block of 2 to 7 days

for 3- to 8-d turnarounds) as including (true) and not (false) one or more near-to-

MSS exposure using >85%, >90% and >95% of each player’s MSS.11

Timing of near-to-MSS exposures during the turnarounds

To understand the actual programming of near-to-MSS exposures, the individual

player near-to-MSS exposure distribution patterns were coded within each

microcycle. Each day was labeled as to whether a near-to-MSS exposure occurred,

say ‘x’ for a day without and ‘o’ for a day with exposure(s), and the frequency of

each of the possible combinations e.g. x/x/x, o/x/x, x/o/x, x/x/o, o/o/x, o/x/o, x/o/o,

o/o/o for 4-d turnarounds assessed for each turnaround. We then decided to group

together the first training days of each turnaround up to and including D-3 e.g. for a

7-d turnaround we grouped D-6 to D-3 together as D-3, and D-2 and D-1 were

8
considered as unique days. We did this for two main reasons: 1) to simplify the

analysis given the large amount of combinations, especially for the longest

turnarounds and 2) coaches generally split the between-match training cycle into

two phases (recovery/compensation from D+1 until D-3, and match preparation D-2

and D-1). This grouping allowed us to approximate all turnarounds as a standard 3-

day turnaround. Since 3-d turnarounds do not include a D-3, this microcycle was

only included in a part of this analysis. Finally, only 3-day combinations containing

≥200 player-sequences, i.e. at least 10 turnarounds of 20 players, were retained.

Data analysis

Considering all the above, the final analysis was run on a total of 627 players

participating in 5052 training session days and 1358 non-international matches for

a total of 24486 player-turnarounds (3 to 8 days), and 152 hamstring injuries, with

96 of those occurring during matches, as part of the 36 team-seasons.

Since preliminary analysis did not show any trends suggestive of differences

between the different leagues or continents, all data were pooled together to

increase sample size.

We first looked at the associations between match hamstring injury rate within

each turnaround and whether >85%, >90% or >95% MSS running bout exposures

occurred during the overall block of training sessions leading to those matches.

Next we examined match hamstring injuries in relation to the day when these

exposures occurred (e.g., near-to-MSS exposure at D-2 vs D-1, respectively).

9
Statistical analysis

Results are presented as a mean and 95% confidence intervals (using the exact

binomial approach). Substantial differences were assumed when the CIs did not

overlap.14

RESULTS

Near-to-MSS running bouts occurrences

The number of player-sequences within each turnaround examined where >85%,

>90% and >95% MSS exposures occurred (true) or not (false) during the training

session days leading to the match are shown in Table 1, 2 and 3, respectively.

Overall, the proportion of training sequences with at least one near-to-MSS

exposure was clearly lower than without, i.e., 40%, 24%, 10% for >85, >90 and >90%

MSS, respectively.

When looking at the turnaround level, the longer the length of the turnarounds and

the lower the speed thresholds, the greater the number (and proportion) of near-

to-MSS exposures (Figure 1 and Tables 1-3). For example, when considering >85%

MSS exposures, there were 2 (5-d turnarounds) to 5 x (3-d turnarounds) more

sequences without exposures than with. For the longest turnarounds however,

sequences with near-to-MSS exposures were 2 (7-d turnaround) to 3 x (8-d

turnaround) greater than those without.

Occurrence of Number of Number of hamstring Injury rate/1000

>85% MSS player- match injuries sequences (95% CL)

10
 running bouts sequences

during the

training session

days leading to

the match

False 7389 33 4.5 (2.9-6)

3d
True 1626 9 5.5 (1.9-9.1)

False 5663 17 3.0 (1.6-4.4)

4d
True 2750 9 3.3 (1.1-5.4)

False 1461 2 1.4 (0-3.3)

5d
True 1177 2 1.7 (0-4.1)

False 1037 1 1.0 (0-2.9)

6d
True 1580 4 2.5 (0.1-5)

False 1430 4 2.8 (0.1-5.5)

7d
True 3677 11 3.0 (1.2-4.8)

False 439 0 0.0 (0-8.4)

8d
True 1207 4 3.3 (0.1-6.6)

Total False 17419 57 3.3 (2.4-4.1)

True 12017 39 3.2 (2.2-4.3)

Table 1. Number of player-sequences within each turnaround examined where >85% MSS

11
exposures occurred (true) or not (false) during the training session days leading to the match,

number and rate (95% confidence limits, CL) of hamstring injuries during matches.

Occurrence of

>90% MSS running

bouts during the

training session Number of

days leading to the player- Number of hamstring Injury rate/1000

match sequences match injuries sequences (95% CL)

False 8259 38 4.6 (3.1-6.1)

3d True 756 4 5.3 (0.1-10.5)

False 6933 22 3.2 (1.8-4.5)

4d True 1480 4 2.7 (0.1-5.3)

False 1992 4 2 (0-4)

5d True 646 0 0 (0-5.7)

False 1724 4 2.3 (0-4.6)

6d True 893 1 1.1 (0-3.3)

False 2564 7 2.7 (0.7-4.7)

7d True 2543 8 3.1 (1-5.3)

8d False 843 2 2.4 (0-5.7)

12
 True 803 2 2.5 (0-5.9)

Total False 22315 77 3.5 (2.7-4.2)

True 7121 19 2.7 (1.5-3.9)

Table 2. Number of player-sequences within each turnaround examined where >90% MSS

exposures occurred (true) or not (false) during the training session days leading to the match,

number and rate (95% confidence limits, CL) of hamstring injuries during matches.

Occurrence of

>95% MSS running

bouts during the

training session Number of

days leading to the player- Number of hamstring Injury rate/1000

match sequences match injuries sequences (95% CL)

False 8774 42 4.8 (3.3-6.2)

3d True 241 0 0 (0.0-15.2)

False 7837 23 2.9 (1.7-4.1)

4d True 576 3 5.2 (0-11.1)

False 2416 4 1.7 (0-3.3)

5d True 222 0 0 (0.0-16.2)

6d False 2254 5 2.2 (0.3-4.2)

13
 True 363 0 0 (0.0-10.1)

False 3902 12 3.1 (1.3-4.8)

7d True 1205 3 2.5 (0-5.3)

False 1310 2 1.5 (0-3.6)

8d True 336 2 6 (0-14.2)

Total False 26493 88 3.3 (2.6-4)

True 2943 8 2.7 (0.8-4.6)

Table 3. Number of player-sequences within each turnaround examined where >95% MSS

exposures occurred (true) or not (false) during the training session days leading to the match,

number and rate (95% confidence limits, CL) of hamstring injuries during matches.

Overall occurrence of near-to-MSS running bouts and match hamstring injuries

When looking at all turnarounds pooled, there was no difference in injury rate

between training sequences including vs. not including near-to-MSS running

exposures, irrespective of the speed threshold considered (Table 1-3). However,

when looking within each turnaround, there were no match hamstring injuries

when players were exposed to running bouts >90% MSS (i.e., 5-d turnaround) and

>95% MSS (i.e., 3-, 5- and 6-d turnaround) during the training sessions days leading

to matches (Table 3 and Figure 2 lower panel).

In contrast, injury rate was still substantial when considering running bouts >85%,

and when looking at the majority of turnarounds with >90% MSS exposures (Table

1 and 2, Figure 2 upper and middle panel).

14
Figure 1. Proportion of (training) player-sequences including at least one >85%, >90% and
>95% MSS running bouts occurrence as a function ofthe length of the turnaround.

Daily programming of near-to-MSS exposures and match hamstring injuries during 4- to

8-d turnarounds pooled

When looking specifically at the day(s) when >95% MSS was reached within an

average turnaround (i.e., 4- to 8-d turnarounds pooled), there were four main

patterns with large sample sizes (n >200): near-to-MSS occurrence at D-3 and

before, n = 990 player-turnarounds and 6 injuries; at D-2, n= 480 and 0 injuries; at

D-1, n = 215 and 2 injuries, and no exposure throughout the turnaround, n = 11168

and 46 injuries. The other day-combinations (e.g. occurrences both at D-2 and D-1)

15
had all very low sample sizes (n < 50), for a total of 126 player-turnarounds in total

and no injuries; these later combinations were not used for analysis.

When examining the pooled data and comparing these main four patterns, there

was no observation of match hamstring injury when >95% MSS was reached at D-2

- and only for that day (Figure 3). In contrast, injuries still happened when players

were not exposed at all, or when these exposures occurred at D-3 and/or earlier

within the turnaround. The difference in injury rate between exposures at D-2 vs D-

1 was unclear, likely due to the very low number of injuries for the latter (n=2!).

16
Figure 2. Match hamstring injury rate (with 95% confidence intervals, and per 1000

turnarounds participation) in players achieving (true) or not (false) >85% (upper panel), >90%

(middle panel) or >95% (lower panel) of their maximal sprinting speed (MSS) during the training

session days leading to the match, for the different turnarounds examined.

17
Daily programming of near-to-MSS exposures and match hamstring injuries during all

turnarounds pooled

During 3-d turnarounds (excluded from the above analysis since not including D-3

data), 97% of the player-sequences (n = 5854 player-turnarounds and 42 injuries)

did not include 95% MSS exposures, with an hamstring injury rate of 7.1 (6.1-8.3).

The number of other player-sequences were all below 80, with no injury when

>95%MSS exposures occurred at D-2 or D-1.

When adding the 3-d turnarounds to the previous analysis to increase the number

of injuries up to 96 in total (Figure 4, but then removing the D-3 aggregation to be

consistent), the trends were similar than in Figure 3, but there was almost no

overlap anymore between the D-2 vs non-exposure alternatives (Figure 4).

18
Figure 3. Match hamstring injury rate (with 95% confidence intervals, and per 1000 player-

turnarounds) in relation to the training session day(s) of the turnaround when running bouts

>95% MSS occurred. *Note that D-3 is an aggregation of all training session days of the

turnaround before D-3 included (e.g., D-3 summarizes occurrences from D-6 to D-3 for a 7-d

turnaround, see methods). Data presented here are from 4- to 8-d turnarounds pooled

together; since there is no D-3 data during 3-d turnarounds, data from the entire 3-d

turnarounds is excluded from this analysis.

19
Figure 4. Match hamstring injury rate (with 95% confidence intervals, and per 1000 player-

turnarounds) in relation to near-to-MSS exposure over the last two training day(s) of the

turnaround - 3-d to 8-d turnarounds pooled.

DISCUSSION

This is to our knowledge the first study to examine both the occurrence of near-to-

MSS running bouts within typical turnaround in elite football, and the association

between the programming of near-to-maximal speed exposures and match non-

contact time loss hamstring injury rates.

20
 The main findings were the following: 1) the large majority of players arrived

to the match without having been exposed to near-to-MSS running bouts during

the training days of the current turnaround (60% for >85% MSS, 76% for >90% MSS

and 90% for >95% MSS), 2) the longer the length of the turnarounds and the lower

the speed thresholds, the greater the number (and proportion) of near-to-MSS

exposures, 3) for some of the turnarounds, there were no match hamstring injuries

when players were exposed to running bouts >95% MSS during the training session

days leading to matches, vs. when there was no exposures. In contrast, this was not

apparent when considering running bouts only >85% or >90% MSS, and finally, 4)

there were no hamstring injuries when >95% MSS exposures occurred at D-2, while

in contrast, injuries still happened when players were not exposed at all, or when

these exposures occurred at D-3 and/or earlier within the turnaround.

Near-to-MSS running bouts occurrences

The most common practice was not to touch near-to-MSS running speeds during

training. On average, there were 3 to 10 x more player-turnarounds without near-

to-MSS exposures than with, and it was only during the longest turnarounds that

these higher running speeds were reached (Figure 1).

The first part of these findings is not surprising, and is likely related to the type of

drills programmed by most coaches, which do not allow players to reach high

speeds.15 It is now well established that small-side games over small spaces are

insufficient with this regard (since players may need to maximally sprint over at

least 30 m to reach near-to-MSS speeds16), and that often, the only way to get

players exposed to near-to-MSS exposures is to either program finishing and

21
transition drills with enough depth17 and/or individual sprinting drills with or

without the ball.18 The current results (discussed below) lend support to this latter

practice.

The influence of the turnaround length on near-to-MSS running bout occurrence is

also consistent with the results of our recent survey,12 where the most important

drivers for the programming of almost all training contents, and especially those

demanding either at the neuromuscular or metabolic level, was reported to be the

distance from and to the next match. With not enough time between matches, the

emphasis is put on recovery, and practitioners likely consider maximal sprint work

too demanding to be performed close to the previous match (the residual fatigue

from the previous match may increase injury risk during sprint training itself). In

fact, during periods of match congestion, the typical training programming

(recovery and easy sessions) does not allow near-to-MSS exposures for starters;

those higher-speed exposures may only be possible (and required, see below) for

substitutes. When to program those high-speed exposures for subs is another

great question for practitioners, and a tentative answer to this will be provided in

the last part of the discussion. Finally, these results are also consistent with the

common trend found both in the scientific literature19 and the coaching

community, suggesting that 48h of recovery are generally needed between sprint

training sessions/events.

How fast is enough?

While researches have shown associations between high-speed exposures and

injury risk,8-10 there was still a lack of evidence about the minimal intensity required

22
for those runs to be protective. In fact, when it comes to selecting the minimal

running velocity that may be associated with reduced hamstring injuries, large

variations in relative velocities have been reported, ranging from ≥80,9 to 858 or

even 95% of MSS.10 While the present observational study design precludes the

examination of causal relationships, our results show for the first time in a very

large sample of elite football players (627 players for a total of 24486 player-

turnarounds), that near-to-MSS exposures may need to be performed >95% MSS

during training to be associated with reduced match hamstring injuries (Figure 2).

While limited with the present data, the fact that >95% MSS exposures may be

associated with lower injury rates than when only reaching lower relative speeds

(85% and 90% MSS), may be related to both higher levels of movements specificity

(e.g., leg interaction, elastic energy transfer, reflexes, kinematics, kinetics)6 and

hamstring muscles recruitment that increase in parallel to running speed.

Programming near-to-MSS running bouts during the training microcycle

Previous research had suggested the existence of an optimal chronic “dose” in

terms of near-to-MSS exposures (i.e., number of weekly exposure 8,10 and/or

monthly cumulative distance).9 However, this optimal chronic dose is likely specific

to each population and context, and it is therefore difficult to provide guidelines for

all practitioners on the back of those three studies. More importantly, those studies

did not provide clear guidelines on how and when to program these near-to-MSS

exposures during the weekly microcycle and during turnarounds of different

lengths. For these reasons, we believe that our results shed some light on the

potential (more) optimal practices in the field.

23
In this very large data set, there were no match hamstring injuries when near-to-

MSS exposures were programmed at D-2. Importantly, this was the case only when

near-to-MSS exposures were programmed on that day (Figure 3 and 4). Despite the

overlap of the CIs, and aknowledging that association does not imply causality, this

trend suggests that reaching near-to-MSS at D-2 may be the most advantageous

strategy with respect to match hamstring injury occurrence.

The actual programming of MSS exposures at D-2 vs D-3 was actually one the areas

the most debated among the practitioners we surveyed.12 In accordance with the

discussion around the alternance of moderate vs. light loads between D-2 and D-1,

the sequence order of high-speed running (HSR) and MSS work may have some

relevance in the context of injury risk. In fact, since high training loads including

HSR and playing over large spaces (which are mainly programmed on D-3,

irrespective of the periodization approach12) likely induce acute posterior chain

fatigue,20 the programming of MSS work the next day (D-2) could expose players to

a higher risk of injury during those sprints (assuming that increased neuromuscular

fatigue and the changes in mobility/pelvic control that follow such sessions increase

injury risk).21,22 For that reason probably, and in somewhat contradiction with the

orientation of the tactical periodization approach that advises to plan speed work

on D-2,12 75% of practitioners reported to program MSS on the same day as HSR (D-

3) for both 6- and 7-day turnovers (Figure 7). This is often achieved during game-

play sequences over large spaces17 and/or through specific speed top-ups post

session when speed targets are not reached.18 Albeit anecdotal, several

practitioners commented in their notes that while they had started to program MSS

work at D-2 in the line of the tactical periodization paradigm, they ended up

24
changing this specific programming aspect for the above-mentioned reasons 22,24

Another important comment in relation to this specific point, is that having ‘speed’

as the focus of the third acquisition day (following ‘strength’ and ‘endurance’12 have

been sometimes misunderstood: ‘speed’, as originally introduced, may not

necessarily involve MSS work, but could simply refer to speed of execution, which is

often implemented via short attacking transition work and finishing actions.

While the benefit of programming vs not programming near-to-MSS exposure is

straightforward (i.e., preparing muscles to match-specific demands),5,6 it remains

unclear why exposing players at D-2 may be more appropriate than at D-1 or D-3

and earlier (if this is that clear, considering the CIs overlaps, though). This may be

related to the recovery time course of the posterior chain muscles when running

near-to-MSS.25 Exposures at D-1 may not allow those muscles to be completely

recovered on match day, and the stimulus (short-term conditioning effect?) may

fade away when performed too early in the week (D-3 and earlier), losing it’s

‘protective effect’. Clearly, studies examining this recovery times course in

ecological conditions would help better understanding this programming aspect.

Practically, if D-2 was to be the most appropriate day for near-to-MSS exposures as

per the current results (Figure 3 and 4), the programming of the other days of the

week may need to be tailored accordingly (i.e., D-4 and D-3), so that players do not

reach D-2 with excessive levels of neuromuscular fatigue - not to be at higher risk of

hamstring injuries during the exposures themselves. Additionally, while those D-2

exposures may concern the entire squad for long turnarounds (i.e., 6- to 8-d), they

may only concern subs for 3- to 5-d turnarounds. In this latter scenario,

25
practitioners reported to program these exposures either on match day

immediately post-match, at D+1 or D+2 (in relation to potential days off).12 The

present D-2 practice is then straightforward when that day is either a D+1 (3-d

turnaround), or a D+3 (4-d turnaround). For 5-day turnaround, the option could be

to delay this exposure up to D+3/D-2, and/or spread it across multiple days (match

day and then again D+3/D-2). As always, players and practitioners' experience

would dictate the possible applications of the present findings in their own

context.26

The lack of clearer differences between the different exposures scenarios (CIs

overlap) - despite the very large data set - is likely related to the fact that other

factors than the programming of near-to-MSS exposures per se may have a greater

effect on injury rate, and, in turn, could have diluted/confounded the univariate

analysis. This is an important limitation of the present analysis. While we thought to

answer the simple question of the programming of near-to-MSS exposures, it is

clear that injuries are largely multifactorial in nature27 and that different chronic

training loads and match minutes prior to the turnarounds examined, may also

directly affect injury rates. However, we deliberately decided to zoom within each

turnaround, since this is the way the very large majority of practitioners operate in

the football field, taking and programming one turnaround after the other, with

each of them being almost independent of the previous.12 Additionally, the

simultaneous consideration of player profiles (e.g., age, injury history, strength,

mobility or flexibility) and other measures of internal training load and responses to

load should also improve the analysis - while making the current outputs less

straightforward for practitioners. There is in fact a trade-off between the desire for

26
simple questions to have simple answers (e.g, when is it best to sprint?) and more

sophisticated analytic approaches that may have more precision but require more

effort to interpret in order to provide direct applications (i.e. results of multivariate

analyses can be difficult to translate into simple yes/no answers).

Limitations

In the absence of consistent MSS testing practices across the different teams

examined, player’s MSS was determined from the available GPS data. While recent

results have shown that players may be able to reach their true MSS during

matches and some specific training sessions,15 we were not able to verify this at the

individual player level. It is therefore possible that inaccurate MSS were used in the

analysis, which may have added noise to the results. Also, the low number of

observations and injuries for some training sequences within some turnarounds

can sometimes increase injury rate beyond its actual magnitude, which should be

considered when interpreting the results. Finally, the injury records used for

analysis are as good as what practitioners may have registered. Relying on injuries

based on practitioners' entries is however common practice,1 and we believe that

the value of the information provided, derived from a very large sample size (n =

24486 player-turnarounds), partly outweighs those possible limitations.

Practical applications

While the present observational study design precludes the examination of causal

relationships, reaching >95% of MSS during training may be more protective

against non-contact time loss match hamstring injuries than only reaching >85% or

27
>90%. Additionally, programming 95% exposures at D-2 may be the more relevant

strategy to decrease the incidence of non-contact match hamstring injuries than no

programming exposures at all, or having those exposures at D-3 and/or earlier in

week. If D-2 was to be the most appropriate day for near-to-MSS exposures, the

programming of the other days of the week needs to be tailored accordingly (i.e., D-

4 and D-3), so that players do not reach D-2 with excessive levels of neuromuscular

fatigue - not to be at higher risk of hamstring injuries during the exposures

themselves.

CONCLUSION

Using a very large data set (for a total of 667 players across 38 team-seasons), we

showed for the first time that the large majority of players arrived at the match

without having been exposed to near-to-MSS running bouts during the training

days of the current turnaround. However, and while association does not imply

causation, match hamstring injuries in elite football were systematically lower when

>95% MSS exposures were programmed at D-2.

CONTRIBUTIONS
Contributed to conception and design: MB and MS
Contributed to acquisition of data: MS and MB
Contributed to analysis and interpretation of data: MB, MS, KH and DM
Drafted and/or revised the article: MB, MS, KH and DM
Approved the submitted version for publication: MB, MS, KH and DM

28
ACKNOWLEDGEMENTS
The authors thanks Stephen Smith and Darcy Norman for their comments on a
draft of the current manuscript.

REFERENCES

1. Ekstrand J., Spreco A., Bengtsson H., Bahr R. Injury rates decreased in men's
professional football: an 18-year prospective cohort study of almost 12,000 injuries
sustained during 1.8 million hours of play. Br J Sports Med. 2021 Oct; 55(19):1084-
1091. doi: 10.1136/bjsports-2020-103159. Epub 2021 Feb 5.
2. Buchheit M, Eirale C, Simpson BM, et al. Injury rate and prevention in elite football:
Let us first search within our own hearts. Br J Sports Med, 2018.
3. Buckthorpe M, Wright S, Bruce-Low S, et al. Recommendations for hamstring injury
prevention in elite football: Translating research into practice. Br J Sports Med,
2019;53(7):449-456.
4. Bourne MN, Timmins RG, Opar DA, et al. An evidence-based framework for
strengthening exercises to prevent hamstring injury. Sports Med, 2018;48(2):251-267.
5. Butler S, Running fast: The cause, the cure and a vaccine, in BJSM Blog,
https://blogs.bmj.com/bjsm/2019/03/19/running-fast-the-cause-the-cure-and-a-
vaccine/, Editor. 2019.
6. Edouard P, Mendiguchia J, Guex K, et al. Sprinting: A potential vaccine for hamstring
injury? Sport Performance & Science Reports, 2019;January(48):v1.
7. van den Tillaar R, Solheim JAB, and Bencke J. Comparison of hamstring muscle
activation during high-speed running and various hamstring strengthening exercises.
Int J Sports Phys Ther, 2017;12(5):718-727.
8. Colby MJ, Dawson B, Peeling P, Heasman J, Rogalski B, Drew MK, Stares J.
Improvement of Prediction of Noncontact Injury in Elite Australian Footballers With
Repeated Exposure to Established High-Risk Workload Scenarios. Int J Sports Physiol
Perform. 2018 Oct 1;13(9):1130-1135. doi: 10.1123/ijspp.2017-0696. Epub 2018 Oct
19.
9. O'Connor F, Thornton HR, Ritchie D, et al. Greater association of relative thresholds
than absolute thresholds with noncontact lower-body injury in professional
australian rules footballers: Implications for sprint monitoring. Int J Sports Physiol
Perform, 2019:1-9.

29
10. Malone S, Roe M, Doran DA, et al. High chronic training loads and exposure to bouts
of maximal velocity running reduce injury risk in elite gaelic football. J Sci Med Sport,
2017;20:250-4.
11. Buchheit, M., Simpson, BM, Hader, K., and Lacome, M. Occurrences of near-to-
maximal speed running bouts in elite soccer: insights for training prescription and
injury mitigation.Sci Med Footb. 2021 May;5(2):105-110. doi:
10.1080/24733938.2020.1802058.
12. Buchheit, M. Sandua M, Berndsen J, Shelton, Smith S, Norman D, McHugh D and
Hader K. Loading patterns and programming practices in elite football: insights from
100 elite practitioners. Sport Perf & Sci Research, Dec 2021, V1.
13. Bahr R, Clarsen B, Derman W, Dvorak J, Emery CA, Finch CF, Hägglund M, Junge A,
Kemp S, Khan KM, Marshall SW, Meeuwisse W, Mountjoy M, Orchard JW, Pluim B,
Quarrie KL, Reider B, Schwellnus M, Soligard T, Stokes KA, Timpka T, Verhagen E,
Bindra A, Budgett R, Engebretsen L, Erdener U, Chamari K. International Olympic
Committee consensus statement: methods for recording and reporting of
epidemiological data on injury and illness in sport 2020 (including STROBE Extension
for Sport Injury and Illness Surveillance (STROBE-SIIS)). Br J Sports Med. 2020
Apr;54(7):372-389. doi: 10.1136/bjsports-2019-101969.
14. Cumming, G. Understanding The New Statistics. Effect Sizes, Confidence Intervals,
and Meta-Analysis. 2011,Routledge, 2011, 536p. ISBN 9780415879682.
15. Kyprianou E, Di Salvo V, Lolli L, Al Haddad H, Villanueva AM, Gregson W, Weston M. To
Measure Peak Velocity in Soccer, Let the Players Sprint. J Strength Cond Res. 2022 Jan
1;36(1):273-276. doi: 10.1519/JSC.0000000000003406. PMID: 31800476
16. Buchheit M, Simpson BM, Peltola E, Mendez-Villanueva A. Assessing maximal
sprinting speed in highly trained young soccer players. Int J Sports Physiol Perform.
2012 Mar;7(1):76-8. doi: 10.1123/ijspp.7.1.76. Epub 2011 Oct 12. PMID: 22001861
17. Asian-Clemente J, Rabano-Munoz A, Requena B, Suarez-Arrones L. High-speed
training in a specific context in soccer: transition games. Int J Sports Med. 2022 Mar
10. doi: 10.1055/a-1794-9567. Online ahead of print. PMID: 35272387
18. Jeffreys I, Huggins S, and Davies N. Delivering a game speed-focused speed and agility
development program in an english premier league soccer academy. Strength and
Conditioning Journal, 2017;40(3):23-32.
19. Haugen T, Seiler S, Sandbakk Ø, Tønnessen E. The Training and Development of Elite
Sprint Performance: an Integration of Scientific and Best Practice Literature. Sports
Med Open. 2019 Nov 21;5(1):44. doi: 10.1186/s40798-019-0221-0.
20. Buchheit M, Lacome M, Cholley Y, Simpson BM. Neuromuscular Responses to
Conditioned Soccer Sessions Assessed via GPS-Embedded Accelerometers: Insights
Into Tactical Periodization. Int J Sports Physiol Perform. 2018 May 1;13(5):577-583. doi:
10.1123/ijspp.2017-0045. Epub 2018 May 22.

30
21. Verrall GM, Slavotinek JP, Barnes PG, Fon GT, Spriggins AJ. Clinical risk factors for
hamstring muscle strain injury: a prospective study with correlation of injury by
magnetic resonance imaging. Br J Sports Med. 2001 Dec;35(6):435-9; discussion 440.
doi: 10.1136/bjsm.35.6.435.
22. Hennessey L, Watson AW. Flexibility and posture assessment in relation to hamstring
injury. Br J Sports Med. 1993 Dec;27(4):243-6. doi: 10.1136/bjsm.27.4.243.
23. Kyprianou E. Let’s talk about weekly plans in soccer. Blog post June 2020, EK Sports
Performance.
24. Kyprianou E. Balancing Physical & Tactical Load in Soccer. A practical guide for soccer
practitioners. Complementary training. 2019
25. Rhodes D, McNaughton L, Greig M.. The temporal pattern of recovery in eccentric
hamstring strength post-soccer specific fatigue. Res Sports Med. 2019 Jul-
Sep;27(3):339-350. doi: 10.1080/15438627.2018.1523168. Epub 2018 Oct 8.
26. Buchheit M, Allen SV. To Optimize? First, Empathize.Int J Sports Physiol Perform. 2022
Apr 1;17(4):505-506. doi: 10.1123/ijspp.2022-0036. Epub 2022
27. McHugh D & Buchheit, M. Analytics For Sport Performance And Health: A Reality
Check. Kitman Labs Website. Last accessed April 4, 2022.

31