Secondary Logo

Journal Logo

Original Research

Strength and Power Determinants of Grinding Performance in America's Cup Sailors

Pearson, Simon N1; Hume, Patria A1; Cronin, John B1,2; Slyfield, David3

Author Information
Journal of Strength and Conditioning Research: September 2009 - Volume 23 - Issue 6 - p 1883-1889
doi: 10.1519/JSC.0b013e3181b2ba55
  • Free



Muscular capability is a major factor in determining performance across almost all sporting and athletic events. Resistance training has long been established as playing an important role in the development of muscular capability (8); however, depending on the specific demands of the event in question, the muscular requirements for performance and therefore the type of training required for enhancement will differ (1,11). Understanding which muscular characteristics (e.g., strength or endurance; force, velocity, or power) are most important for the performance of a specific event is a key issue in maximizing the transfer of training to performance and therefore improving training efficiency (14). Such an approach has been used in a variety of sports such as soccer (12), rugby (6), and athletics (13); however, there is a general paucity of research into the strength and power requirements of yachting performance.

The America's Cup is widely regarded as the pinnacle of sailing competitions, and, as such, considerable resources are put into improving the performance of the America's Cup yachts (Figure 1) and the sailing teams. One important physical aspect of on-water sailing performance is grinding, a cyclic, high-load, high-intensity upper-body task that provides the power behind tacking and gibing and influences the efficiency of wind use through movement of the sails. The grinding handles are situated on top of an 87 cm pedestal and are at the end of 2 cranks that are orientated at 180° from each other, 1 on either side of the pedestal, making the overall set-up similar to an upper-limb bicycle. Sailors perform grinding in both a forward direction, pushing away from their body at the top of the rotation and backward direction, pulling toward their body at the top of the rotation, depending on the gear in use.

Figure 1
Figure 1:
An America's Cup yacht during test racing.

With regard to resistance training, the movement patterns used in forward and backward grinding correspond well with the bench press and bench pull (prone row) exercises, respectively. Previous research has found grinding performance to correlate strongly with maximal strength in these 2 exercises, with the relationship increasing when grinding is performed against a higher load (9). However, those results were based on 1 repetition maximum (1RM) scores predicted from training data and offer little in terms of whether certain kinetic or kinematic factors related to maximal strength may be more or less important to performance, which may in turn have implications for resistance training programming.


The purpose of this study was to examine the relationship between forward and backward grinding performance and the kinetic and kinematic characteristics of the bench press and bench pull exercises. This information can be used by physical conditioners to better advise best practice for conditioning for grinding performance.


Experimental Approach to the Problem

This study was designed to examine the relationship between grinding performance and the characteristics of the power-load relationship in the bench press and bench pull exercises. Forward and backward grinding performance was assessed using a custom-made grinding ergometer. A Smith machine was instrumented to allow measurement of the kinematics and kinetics of the bar and weights during performance of the bench press and bench pull exercises.


Eleven elite-level male sailors from the Emirates Team New Zealand syndicate (who competed in the final of the 2007 America's Cup) participated in this study. All sailors performed grinding as part of their on-board sailing role. All participants had an extensive strength-training background (minimum of 3 yr), and the bench press and bench pull exercises were commonly used as part of their training program. Sample size was small because of the elite nature of the participants and therefore restricts the level of statistical power.

The sailor's mean (±SD) age, body mass, and height were 33.9 ± 5.5 years, 97.8 ± 12.5 kg, and 186.0 ± 7.1 cm. All participants provided written, informed consent within the guidelines of the AUT University Ethics Committee (reference 04/221).


Grinding performance testing was conducted on a grinding ergometer (Dynapack, Wellington, NZ) with standard pedestal (870 mm vertical) and crank arm (250 mm) dimensions for a main sheet grinding pedestal on an America's Cup class yacht (Figure 2). Gearing for the ergometer was linked through a multiple-speed dynamometer set up to output a number of grinding performance measures. Power output was obtained from the grinding ergometer using a bidirectional oil hydraulic system custom designed to meet the tactile characteristics of the rigging at the grinding station. Speed was based on a 24-slot disc attached directly to the motor input shaft. Output was obtained by way of an analogue to digital converter using 8-bit resolution to a C++ customized data collection system sampling at 40 Hz. Mechanical load was varied using a custom-designed cog selector allowing 1:1 and 3:1 ratios driven by toothed belts. Hydraulic load was applied using a dynamic closed-loop controller modified to operate at low speed. Calibration was performed using calibrated masses and known lever lengths, and the machine was verified accurate to 0.5% or better throughout its range. Reliability of external work (SEM = 1.6-3.9%; intraclass coefficient [ICC] = 0.91-0.97) as a grinding performance measure using the ergometer and protocols outlined here has previously been reported (10).

Figure 2
Figure 2:
Sailor on the grinding ergometer during performance testing.

Biomechanical strength characteristics in the bench press and bench pull were tested using a modified Smith machine (Figure 3). A linear position transducer (Unimeasure, Corvallis, OR) was attached to the bar and measured bar displacement with an accuracy of 0.1 mm. These data were sampled at 500 Hz and relayed to a Labview (National Instruments, Austin, TX) based acquisition and analysis program. The data acquired from the linear transducer in this testing setup have been reported previously as reliable (ICC = 0.92-0.98) and valid across a range of movements and testing conditions (4).

Figure 3
Figure 3:
Testing set-up for the power-load spectrum of bench press exercise.


Before testing sessions, sailors had a minimum of 1 day of rest without any training. The usual team protocols before training for nutrition, hydration, and sleep were followed and were the same before both testing sessions. Grinding and strength performance testing were performed on separate days within 7 days of each other. Both forward and backward grinding performances were tested using 2 torque loading conditions: 48 N.m (moderate) and 68 N.m (heavy). Testing loads were selected to mimic moderate- and high-load conditions during on-water grinding manoeuvres based on rpm ranges for a primary grinder. All 11 sailors completed the moderate load but because of the physical requirements for grinding effectively at the heavy load, only the 6 sailors regarded as principal grinders (i.e., their main on-board responsibility is grinding) completed the heavy load condition. Sailors completed a self-determined warm-up on the grinding ergometer before testing, with testing consisting of 2 trials of each condition. Trials were maximal, of 8 seconds duration and separated by 5 minutes, and alternated between forward and backward grinding to reduce the possible influence of any fatigue effects.

For strength-power testing, each sailor completed a 60-minute testing session involving both the bench press and bench pull exercises. Familiarization was conducted through a self-determined, exercise-specific warm-up typically consisting of 3 to 4 warm-up sets of the particular exercise using progressively heavier loads. After the warm-up, the individuals' 1RM (Smith machine, concentric-only) was determined to the nearest 2.5 kg. The spectrum of loads for the power testing were then determined from 10-100% of 1RM at 10% intervals. Single repetitions of each load were performed in ascending order, with the instruction that each lift should be performed as explosively as possible without releasing the bar. All lifts were concentric-only, with the bench press initiated from mechanical stops positioned approximately 30 mm off the sailor's chest and the bench pull initiated from a supported supine position. Each lift was separated by a rest period of 1 to 2 minutes (increasing with load), which was considered sufficient to avoid or minimize any fatigue effects. Evidence that any possible fatigue effects were minimal was confirmed by the sailor's ability to repeat their determined 1RM effort from the start of the session in the last of their ordered lifts (100% of 1RM).

Data Analyses

Raw power values were calculated by the Dynapack ergometer software using the formula Power (W) = Torque × (2π*rpm)/60. The raw power curve was then smoothed using a second-order recursive Butterworth low-pass filter and peak power and external work calculated using a customized Labview program. For grinding performance, the measure of interest to this study was external work performed (kJ), the 5-second integral of the area under the power output curve, starting at the occurrence of peak power. External work has previously been shown to be a more reliable performance measure than peak power on this grinding ergometer (10).

Displacement-time data from the linear transducer used in strength-power assessment were filtered using a low-pass Butterworth filter with a cutoff frequency of 6 Hz and then differentiated to determine instantaneous velocity, acceleration, force (using additional load information), and power output data over the range of motion for each load condition. Measures of force determined using this methodology have been previously validated and found to correlate highly with force plate measures across a range of movements and testing conditions (2,4). Variables of interest from the strength-power testing were 1RM, maximum force, maximum velocity and maximum power across the spectrum of testing loads, and the load (%1RM) at which power was maximized (Pmax). Force, velocity, and power were all calculated as the mean concentric value during the concentric phase of the analysed repetition.

Statistical Analyses

Descriptive statistics for all variables are represented as mean and SD. For the power-load spectrum, Pmax values (load and power output) were calculated using the line estimation function (least squares method) in Microsoft Excel (Redmond, WA). Relationships between individual characteristics and grinding performance were analysed using Pearson correlation analysis and Cohen's magnitude of effect scale (3). The influence of individual characteristics on grinding performance was examined using stepwise linear regression in SPSS (v13.0, Chicago, IL) with grinding performance as the dependent variable. Default probability criteria for the model were retained for entry (F ≤ 0.050) and removal (F ≥ 0.100). Presence of significant systematic discrepancy between measures from the bench press and bench pull was determined using a two-tailed unpaired t-test (α level of p ≤ 0.05).


Grinding performance data for the 4 conditions examined are presented in Table 1. Tables 2 and 3 display the mean values for each of the strength-power variables of interest along with their correlation with grinding performance. Table 4 contains individual data for the 1RM and power variables.

Table 1
Table 1:
Performance results for forward and backward grinding at two loads.
Table 2
Table 2:
Relationship between strength-power measures from bench press and forward grinding performance at two different loads.
Table 3
Table 3:
Relationship between strength-power measures from bench pull and backward grinding performance at 2 different loads.
Table 4
Table 4:
Individual 1 repetition maximum and power measures from the bench press and bench pull, with correlations (r) of power variables with 1 repetition maximum for each exercise.

The results for forward grinding in Table 2 show very large, significant (p < 0.01) positive correlations between both 1RM and the force able to be produced for the bench press (r = 0.87-0.99), with a stronger relationship when grinding against a heavy load compared with grinding against a moderate load. Velocity capability showed no significant relationship to forward grinding performance. Power capability and the relative load at which maximum power (in the bench press) was generated both showed moderately large (r = 0.49-0.55) but nonsignificant relationships with grinding performance. Power capability was positively associated with grinding performance, whereas Pmax load was negatively associated.

The results from the bench pull testing and their relationship to backward grinding performance are detailed in Table 3. As with forward grinding, 1RM and force capability demonstrated the largest correlations with moderate load performance (r = 0.87-0.95); however, power capability also had a very large, significant correlation and showed the strongest relationship with heavy load performance (r = 0.85-0.98). Velocity capability showed a very large, positive correlation with heavy load backward grinding performance (r = 0.97) although only a moderately large, nonsignificant correlation with performance at lower (moderate) load (r = 0.56). Pmax load had a similar (moderate, negative) correlation with backward grinding performance as with forward grinding.

Stepwise regression did not add considerably to the results in Tables 2 and 3, with maximal strength (represented as 1RM) being the key predictor of forward grinding performance. Bench press 1RM and maximum force explained 87% of performance variation for moderate load forward grinding (1RM only: r2 = 0.753; 1RM + max force: r2 = 0.866), whereas 1RM alone explained 97% (r2 = 0.966) of heavy load forward grinding performance. For backward grinding, 1RM was still the key predictor in moderate load performance (r2 = 0.796); however, power capability showed the greatest common variance with grinding performance at the heavy testing load (r2 = 0.960).


Bench press 1RM and bench press maximum force were the 2 strongest predictors of forward grinding performance, with the relationship improving as grinding load increased. In the stepwise linear regression model used for this data, 1RM and force together explained 87% of the intersailor variation in forward moderate load grinding, whereas 1RM by itself explained 97% of the variation in forward heavy load grinding. Although power had moderate correlations (r = 0.49-0.55), the ability to generate velocity in the bench press had a trivial relationship with forward grinding performance. It appears that maximal strength, rather than power, is the muscular characteristic most related to forward grinding performance.

For backward grinding and bench pull, the maximal strength variables (1RM, force) still had very strong relationships with grinding performance; however, the ability to generate power and velocity appeared to be comparatively much more important than for forward grinding. Although 1RM was still the best predictor of medium load backward grinding performance (explaining 80% of intersailor variation), the relationship for power was similar (correlation of r = 0.85 for power vs. r = 0.90 for 1RM). Furthermore, for heavy load backward grinding, both power (r = 0.98) and velocity (r = 0.97) had slightly stronger relationships with performance than 1RM or force (r = 0.948 for both). This also showed in the regression model in which power capability explained 96% of intersailor performance variation by itself. It should be noted that there were only 6 sailors in the heavy load grinding group, and lower numbers tend to decrease the precision of statistical analyses and increase the likelihood of unusual findings. As such, although the prominence of power as a predictor of grinding performance is not entirely unexpected, the strong relationship between velocity and heavy load performance could well be questioned because it is slightly counterintuitive that speed on its own should be so crucial.

What cannot be discounted, however, is that power (and to a lesser extent velocity) have been shown to be an important factor in backward grinding performance, whereas it was fairly negligible in forward grinding. One possible explanation for this finding is the difference in muscle architecture (fiber length, type, arrangement, etc.) between the flexor/“pull” muscles used for the bench pull and the extensor/“push” muscles (bench press). The greater fiber lengths and longitudinal fiber arrangement of the primary movers in the bench pull exercise (latissimus dorsi, biceps brachii, brachialis) are characterized by faster shortening velocities, whereas the primary movers for the bench press (pectoralis major, triceps brachii) have shorter fiber lengths, greater pennation angles, and subsequently greater force capability (7). This theory would appear to be consistent with the higher power outputs and higher Pmax load (%1RM) for bench pull compared with the bench press observed here.

What these findings suggest is that there is probably merit in training the push and pull components of grinding differently. The resistance training regime for the sailors involved in this study was primarily focussed around improving maximal strength/force generation because this was understandably (and correctly) viewed as a key factor in grinding performance. Although the findings from this study fully support that approach for enhancing forward grinding, it may be that greater performance improvements could be made in backward grinding by incorporating a greater velocity or power component in the training program while maintaining the maximal strength base that is still crucial for backward grinding performance.

One additional factor that should be addressed with regard to possible training stimulus for improving power is the relevance of the load at which the mechanical power output of the muscle is maximized. There is a school of thought that this may be important in maximizing power and performance gains (5), and this has been shown to be the case in some instances (11). Although the findings from this study cannot contribute in any definitive manner to this particular question, it is worth noting that the relationship between power capability and Pmax load was moderate for the bench pull (r = −0.529, p = 0.094) and negligible for the bench press (r = −0.199, p = 0.558). When combined with the moderately negative correlations of Pmax load with performance (Tables 2 and 3) and 1RM (bench press: r = −0.603, p = 0.050; bench pull: r = −0.522, p = 0.099), this indicates that, as an individual's training status advances and they become stronger or more powerful, the relative load at which power is maximized decreases. It has been observed from previous research that relative Pmax load is transient (5), and although the findings from this study support that conclusion, they do suggest that it may in fact be more a case of strength and power capability being transient and changing with training while the absolute load that maximizes power remains constant, resulting in a lower relative load.

Practical Applications

It is apparent from this study that the strength-power characteristics that determine grinding performance differ between the push-based forward grinding and the pull-based backward grinding. Although maximal strength and force production capability are crucial attributes across all grinding conditions examined here, for backward grinding, the need to be able to produce this force at speed (power) at least matches, if not surpasses, maximal strength as the most important muscular function characteristic. It therefore appears possible that greater benefits for backward grinding performance can be derived by altering the training stimulus for upper-body pull-based resistance exercises to more efficiently stimulate the development of muscular power in addition to maximal strength. This information was later used by the Emirates Team New Zealand physical conditioner to develop a conditioning intervention to help improve grinding performance.


This research was funded by the Tertiary Education Commission (New Zealand) and Emirates Team New Zealand. Thanks are given to the sailors and support crew at Emirates Team New Zealand for their participation and assistance with this research. The authors have no professional relationships with any companies or manufacturers identified in this study.


1. Campos, GE, Luecke, TJ, Wendeln, HK, Toma, K, Hagerman, FC, Murray, TF, Ragg, KE, Ratamess, NA, Kraemer, WJ, and Staron, RS. Muscular adaptations in response to three different resistance-training regimens: specificity of repetition maximum training zones. Eur J Appl Physiol 88: 50-60, 2002.
2. Chiu, LZF, Schilling, BK, Fry, AC, and Weiss, LW. Measurement of resistance exercise force expression. J Appl Biomech 20: 204-212, 2004.
3. Cohen, J. Statistical power analysis for the behavioral sciences. Mahwah, NJ: Lawrence Erlbaum, 1988.
4. Cronin, JB, Hing, RD, and McNair, PJ. Reliability and validity of a linear position transducer for measuring jump performance. J Strength Cond Res 18: 590-593, 2004.
5. Cronin, J and Sleivert, G. Challenges in understanding the influence of maximal power training on improving athletic performance. Sports Med 35: 213-234, 2005.
6. Harris, NK, Cronin, JB, Hopkins, WG, and Hansen, KT. Relationship between sprint times and the strength/power outputs of a machine squat jump. J Strength Cond Res 22: 691-698, 2008.
7. Lieber, RL and Friden, J. Functional and clinical significance of skeletal muscle architecture. Muscle Nerve 23: 1647-1666, 2000.
8. Pearson, D, Faigenbaum, A, Conley, M, and Kraemer, WJ. The National Strength and Conditioning Association's basic guidelines for the resistance training of athletes. Strength Cond J 22: 14-27, 2000.
9. Pearson, S, Hume, P, Ackland, T, and Slyfield, D. Anthropometry and strength predictors of grinding performance in America's Cup sailors. In: Proceedings of XXII International Symposium on Biomechanics in Sports. Ottawa, Canada, 2004. pp. 407-410.
10. Pearson, SN, Hume, PA, Slyfield, D, and Cronin, JB. External work and peak power are reliable measures of ergometer grinding performance when tested under load, deck heel and grinding direction conditions. Sports Biomech 6: 71-80, 2007.
11. Wilson, GJ, Newton, RU, Murphy, RU, and Humphries, BJ. The optimal training load for the development of dynamic athletic performance. Med Sci Sports Exerc 25: 1279-1286, 1993.
12. Wisloff, U, Castagna, C, Helgerud, J, Jones, R, and Hoff, J. Strong correlation of maximal squat strength with sprint performance and vertical jump height in elite soccer players. Br J Sports Med 38: 285-288, 2004.
13. Young, W, McLean, B, and Ardagna, J. Relationship between strength qualities and sprinting performance. J Sports Med Phys Fit 35: 13-19, 1995.
14. Young, WB. Transfer of strength and power training to sports performance. Int J Sports Physiol Perform 1: 74-83, 2006.

upper body; directional difference; muscular performance

© 2009 National Strength and Conditioning Association