Torque stiffness of rugby shoes

1. Introduction

Rugby is a high impact sport but it is relatively unique due to the differing roles, and thereby different forces are encountered by players in different positions. Preatoni et al. [1] demonstrated that during scrimmaging these forces where mainly head on and lateralcompression forces. However rucking and mauling are inherentparts of the forwards’ game which will generate higher rotationalforces. It has previously been demonstrated that there are significant differences in rotational stiffness in different football bootdesign [2–6]. The effect of this is that differing designs of foot-ball boots have been shown to be associated with higher rates ofanterior cruciate ligament injury [7].The aim of this study is to measure the peak torque and rota-tional stiffness of different designs of rugby boots as this may besignificant in choosing the correct boot for players in differentpositions.

2. Materials and methods

This is a laboratory-based biomechanical study. Five different types of rugby shoes typically worn by scrum players were selected and used in the study. These shoes were all cleated with 8 screw-inmetal studs designed for soft ground.

2.1. Shoes design

The five shoes were:

  • Adidas AdiPure Regulate (A).
  • Canterbury Stampade Club 8 stud (C).
  • KooGa EVX II LCST Boot (K).
  • Mizuno Fortuna SI Rugby (M).
  • Puma Esito Finale H8 (P).

Each shoe had eight metal studs on the shoe with the layout ofthe studs being similar in all shoes with six studs on the forefootand two on the heel part of the sole (Fig. 1). The characteristics of the shoes are summarized in Table 1.

Rugby shoes models.

Rugby shoes models.

Schermata 2015-01-22 a 16.55.23

2.2. Surfaces description

A natural grass (Paradello Vivai, Italy, Brescia) made of a mixtureof Lolium perenne and Poa pratensis cultivated on a layer of sand andsoil (4 cm thickness), was provided in form of turfs (60 cm × 40 cm;mass: 19.23 ± 0.89 kg). In order to investigate the possible variabil-ity of composition and water content of each turf, prior to testing,each turf was weighted verifying that the weights were compara-ble among the different turfs. The mean weight was 17.7 kg with astandard deviation of 1.6 kg, thus confirming that the surfaces werenot a confounding variable.The turf was housed within a wooden box. In order not to allow any rotation of the turf within the box during the tests, two woodenplates were pressed on the turf and constrained to the inferior partof the box using grips.

2.3. Artificial foot

An artificial left foot (EU size 42–43) was prepared filling a sil-icon foot cosmetic cover (Road runner foot Engineering, Milano,Italy) with a silicone rubber in which an angled metallic core struc-ture was immersed (Fig. 2). The internal structure is composedby an iron pin with rectangular section rigidly bound throughout screwed connections to two angled plates. The distal part of the ironpin (11 cm in length), which was only partially inserted into thefoot, was used as the interface element for gripping to the loadingmachine.

2.4. Testing machine

Tests were performed on a MTS 858 Bionix servohydraulic test-ing machine (S/N 1015457, MTS, Minneapolis, MN) installed inthe Laboratory of Biological Structure Mechanics of the Politec-nico di Milano. The MTS testing machine was equipped by anaxial–torsional hydraulic actuator, with 25 kN axial capacity and250 Nm torsional capacity, a ±100 mm range LVDT displacementtransducer and a ±140◦range ADT angular transducer mounted onthe actuator. The loads applied were measured by means of a MTSaxial/torsional load cell (model 662.20D-05, S/N 1007099, ±25 kNmaximum axial load, ±250 Nm maximum torsional load). The testswere conducted in air at room temperature (24 ± 2◦C).

2.5. Test procedure

Prior to testing the wooden base housing different turfs wassecured to the inferior grip of the MTS through a T bar located atthe inferior side of the panel. The shoe was dressed on the synthetic foot, then it was assured to the superior grip of the MTS throughthe pin (Fig. 3).Tests were performed applying a quasi-static preload vertical force of 1000 N on the foot using the MTS testing machine, thanthe test was carried out under angular control at a rotation speedof 45◦/s, until a maximum angular rotation of 140◦was reached.

Artificial foot used in rotational tests.

Artificial foot used in rotational tests.

Experimental set-up used in rotational tests.

Experimental set-up used in rotational tests.

These values were defined to reproduce a condition similar to thoseexperienced by players involved in rugby scrums. Tests were per-formed applying both internal (IR) and external rotation (ER) withsupport on the tip (Fig. 3), to simulate wheeling of the scrum inboth directions. Foot orientation with respect to the ground wasadjusted using a 45◦angulated platform lying on the ground. Inorder to assess repeatability of the results, each test was repeatedsix times, paying attention not to repeat the test on a previouslyused part of the turf.Peak torque and rotational stiffness (slope of the torque–anglecurve between 10◦and 30◦) were obtained for each test to suf-ficiently describe the linear section of the shoe–surface interfaceresponse curve (Fig. 4).

2.6. Data analysis

The torque and the angle, measured by the load cell and theADT respectively, were zeroed before testing and then sampled atthe frequency of 500 Hz. For statistical analysis, two-way repeatedmeasures ANOVAs (≤0.05) were employed, using cleat design asfactor. If the ANOVA revealed significant differences, the corre-sponding data-sets were compared using an unpaired Student’st-test. Normality was tested with the Shapiro–Francia test, assum-ing a level of significance of 0.05. Effect size in ANOVA analysis is ameasure of the degree of association between the effect (e.g. typeFig

Measured parameters.

Measured parameters.

Schermata 2015-01-22 a 16.59.42

of shoe) and a dependent variable (e.g. peak torque and rotationalstiffness) across different testing conditions (i.e. internal and exter-nal rotation). In accordance with a previous publication (Ref. [5]),we calculated Eta-squared (2) to express the proportion of vari-ance in the dependent variable that is attributable to each effect:this parameter ranges between 0 and 1 and it indicates a variabledegree of strength (0–0.2 for a weak effect, 0.3–0.5 for a moderateeffect and values greater than 0.5 indicates a significant effect).

3. Results

Mean and standard deviation obtained for both rotational stiff-ness and peak torque are summarized in Table 2. Once the peaktorque is reached at 40◦, the turf was destroyed as a result andneeded to be replaced for subsequent tests.The internal rotation peak torque of the shoes ranged between51.78 ± 2.69 and 68.39 ± 12.74 N m, while the external rotationpeak torque ranged between 49.05 ± 5.28 and 65.61 ± 4.60 N m.Shoe C (Canterbury Stampade Club 8 stud) had the highest peaktorque in both internal and external rotations. Shoe M (Mizuno For-tuna Rugby) had the lowest peak torque in internal rotation and thiswas significantly lower than shoes A, C, and P while Shoe K (KooGaEVX II LCST Boot) had the lowest external rotation peak torque andthis was significantly lower than shoe A, C, and M (p < 0.05).With regards to rotational stiffness, the internal rotationstiffness ranged between 0.58 ± 0.04 and 0.82 ± 0.23 N m/degwhile external rotation stiffness ranged between 0.57 ± 0.03 and0.81 ± 0.07 N m/deg. Again shoe C (Canterbury Stampade Club 8stud) had the highest internal and external rotational stiffnesseswhile shoe K (KooGa EVX II LCST Boot) has the lowest rotationalstiffness in both internal and external rotation. Shoe K (KooGa EVXII LCST Boot) had statistically significant lower internal and externalrotational stiffnesses compared to shoes A, C, and shoe P (p < 0.05).

4. Discussion

Brooks and Kemp reported that rugby had a higher incidence ofinjuries than many other contact sports [8]. In general, rugby play-ers tend to sustain more upper limb and head and neck injuriescompared to football players, whom on the other hand sustainhigher rate of lower limb injuries compared to rugby players [9].Bathgate et al. found that the incidence of ankle injuries in rugbywas relatively low, at only 6% [10]. In a review of published litera-ture on rugby injuries, Kaplan et al. reported that 36–56% of injuriesoccur during the tackle phase with the incidence of injuries duringthe set pieces being 1–13% [11].
O’Connor and James provided a key study on the association oflower limb injuries and boot design in elite football players [2]. Oneof their key points is that in football, foot and ankle non-contactinjury is associated with the player–boot–surface interaction [2].They recommended that sports medical teams should ensure thatplayers boot selection is appropriate to the players physiology andposition to prevent this becoming a potential injury risk factor. Forthis reason a naturally simulated surface was chosen for our studyto avoid a potential confounding factor.Obviously scrums can wheel in either direction as well as pro-gressing forwards, backwards or laterally. It is not possible toproduce a laboratory-based scenario encompassing every force thattakes place. The test procedure is aimed to reproduce the forcesmost likely encountered during the scrum phase. This involvedapplying a quasi-static vertical load of 1000 N on the shoe and thena fast rotation of 40◦/s both internally and externally. These forcesrepresent a typical force that would be produced by in-line scrum-ming with the rotational forces representing forces that would beproduced with wheeling of the scrum. We accept, due to the inter-action between the facing teams pushing and rotating in differentways, tangential force components will also be produced. However,the multitude of all possible forces that could be produced are toonumerous to be tested in this study.All shoes used in the study had screw-in round metal studs withthe same number of studs in each shoe in a distribution of 6 studsin the forefoot area and 2 studs in the heel area. The overall peaktorques of the five rugby shoes used in our study showed that theinternal rotation peak torque was 57.75 ± 6.26 N m while that ofexternal rotation was 56.55 ± 4.36 N m. The Peak internal and exter-nal rotational stiffness were 0.696 ± 0.1 and 0.708 ± 0.06 N m/degrespectively. Model C (Canterbury Stampade) yielded the highestpeak torques and rotational stiffness in both internal and externalrotation while shoe model K (KooGa EVX II LCST Boot) showed thelowest rotational stiffness as well as lowest peak torque in externalrotation. The high rotational stiffness and peak torques recorded inshoe model C when compared to the other shoes could be due to thefact they had the longest studs (18 mm) with a small tip diametercompared to the other shoes which meant that these dimensionsallow more studs penetration in the ground surface thus more rota-tional stiffness and increased peak torque. All boots were testedwith their original supplied studs.Galbusera et al. performed a study on the rotational interactionin three types of soccer shoes on two types of playing surfaces usingvery similar testing setups [6]. The cleat design in their study wasa 6 detachable metal studs, 12 molded rubber studs and 13 bladescleat shoes. The tested shoes in their study were put on an artifi-cial foot which was connected to the test machine. They applieda static preload of 1000 N after which the tests were carried outunder angular control at a rotation speed of 45◦/s, until a maxi-mum angular rotation of 140◦was reached. They tested each shoeat tip and full sole loading support. We compared our results whichwere tip support to their tip support results, as the tip loadingsupport represents the scrum phase foot position. We comparedour results, which were tip support, to their tip support resultswhich is the scrum phase foot position. The peak torque of thethree soccer shoes in internal rotation ranged between 34.3 ± 2.6and 45.0 ± 5.3 N m while that of external rotation ranged between34.9 ± 3.7 and 39.8 ± 3.0 N m. The rotational stiffness in their studyranged between 0.42 ± 0.03 and 0.57 ± 0.07 N m/deg for internalrotation and 0.40 ± 0.06 and 0.48 ± 0.08 N m/deg for external rota-tion. These peak torques and rotational stiffnesses are lower than thanthe values we recorded from the rugby shoes indicating that therugby shoes used in our study showed higher stiffness and moreresistance to torque forces than the football shoes tested on naturalplaying surface.The marked difference in stiffness and peak torques betweenrugby shoes and football shoes confirm that good design practiceshave been used to provide the appropriate level of flexibility orstability necessary for each sport. The rugby shoes tested in ourstudy are commonly used by scrum forwards. However differentpositions in the scrum will have differing main functions, althoughobviously there will be some crossover. Front and second row play-ers will likely perform more with in-line forces whereas back rowforwards are likely to be subjected to more rotatory forces. It istherefore our opinion that if the stiffer boots are chosen then itmay hamper the more mobile players movement or not aid them byproviding stability. If there is correlation with results from footballstudies, there may be a correlation with sustaining injuries.We believe that the risk of non-contact injury in rugby is acombination of many factors of which the interaction between theshoe–surface interaction has a role in. This interaction depends onthe shoes design, cleat design, type of playing surface and posi-tion of the foot on the ground. It is therefore imperative that theboots with the correct mechanical properties are chosen for playersdependent upon their main function within the sport.

5. Conclusion

Rugby boots show higher stiffness and peak torques than pre-viously reported results from football boot studies. There is greatvariance between different designs. In our opinion, to maximizepotential performance and lower the potential of non-contactinjury, care should be taken in choosing boots with stiffness appro-priate to the players main playing role.

Conflict of interest

We would like to declare that there are no conflict of interest.

Download PDF


[1] Preatoni E, Stokes KA, England ME, Trewartha G. The influence of playing levelon the biomechanical demands experienced by rugby union forwards duringmachine scrummaging. Scand J Med Sci Sports 2013;23:178–84.
[2] O’Connor AM, James IT. Association of lower limb injury with boot cleat designand playing surface in elite soccer. Foot Ankle Clin 2013;18:369–80.
[3] Taylor SA, Fabricant PD, Khair MM, Haleem AM, Drakos MC. A review of syn-thetic playing surfaces, the shoe-surface interface, and lower extremity injuriesin athletes. Phys Sports Med 2012;40:66–72.
[4] Bentley JA, Ramanathan AK, Arnold GP, Wang W, Abboud RJ. Harmful cleats offootball boots: a biomechanical evaluation. Foot Ankle Surg 2011;17:140–4.
[5] Villwock MR, Meyer EG, Powell JW, Fouty AJ, Haut RC. Football playing surfaceand shoe design affect rotational traction. Am J Sports Med 2009;37:518–25.
[6] Galbusera F, Zai TOrnese D, Anasetti F, Bersini S, Volpi P, La Barbera L, et al.Does soccer cleat design influence the rotational interaction with the playingsurface? Sports Biomech 2013;12:293–301.
[7] Lambson RB, Barnhill BS, Higgins RW. Football cleat design and its effect onanterior cruciate ligament injuries. A three-year prospective study. Am J SportsMed 1996;24:155–9.
[8] Brooks JH, Kemp SP. Recent trends in rugby union injuries. Clin Sports Med2008;27:51–73.
[9] Junge A, Cheung K, Edwards T, Dvorak J. Injuries in youth amateur soccer andrugby players – comparison of incidence and characteristics. Br J Sports Med2004;28:168–72.
[10] Bathgate A, Best JP, Craig G, Jamieson M. A prospective study of injuries to eliteAustralian rugby union players. Br J Sports Med 2002;36:265–9.
[11] Kaplan KM, Goodwillie A, Strauss EJ, Rosen JE. Rugby injuries: a review ofconcepts and current literature. Bull NYU Hosp Jt Dis 2008;66:86–93.