Recovery phase and boat speed efficiency

Recovery phase and boat speed efficiency

Discussions about the rhythm of the recovery phase are quite often seen in rowing literature and rowing forums. It was argued that “proper” movements of the rower during the recovery phase could make the boat velocity less variable and increase its efficiency and performance. A theory behind this idea originates from a 1986 article by Sanderson and Martindale (1), which concluded: “…the maximum kinetic energy of the rower and boat relative to the rower-boat centre of mass can be reduced by speeding up the initial part of the recovery, slowing the middle part of the recovery, and speeding up the final part of the recovery.” This hypothesis has not yet been proved experimentally.

Definitions of the shell velocity efficiency Es was discussed in RBN 2015/01. The analysis of a large data sample measured in singles (n=2760) revealed that Es is highly dependent on the stroke rate (y=-0.00028x+0.988, r=-0.63), which explains 40% of its variation. So, a simple correlation between variables may reflect their mutual dependence on the stroke rate, but not relationship between them. Therefore, linear trends were built on rate-dependant variables and their deviations from the trend were analysed, and were called normalized variables (e.g. nEs).

To follow the hypothesis above, the shapes of the handle and rower’s segment velocities during recovery should be more rectangular, i.e. the ratio of its average velocity to the maximal (R = Vav/Vmax, which was analysed in the previous Newsletter) should be higher. However, the correlations between nEs and R of velocities during the recovery were insignificant (r<0.1), which means various rhythms of recovery appear to be ineffective on the variation of the boat speed and shell efficiency and disproves the above hypothesis.

To find factors affecting the shell velocity efficiency, all 166 biomechanical variables were correlated with nEs and the factors were ranked. The highest magnitude of correlation was found in the length of the leg drive (r=-0.45, y=-0.0218x+0.011). This means, a longer seat displacement increases the boat velocity variation and energy losses, but the effect is very small: typically, every 1cm of longer seat travel decreases efficiency by 0.02% - less than 0.1s over 2km race.

To illustrate this finding, two single scullers of a similar level, but with different rowing styles were compared at 34spm (Fig.1). Sculler 1 (1.93m, 98kg) has a much longer seat travel of 56cm (Table 1) and his technique could be classified as the ‘Adam’ style (RBN 2006/03). Sculler 2 (1.96m, 95kg) represents an exaggerated ‘DDR’ style and has only a 35cm legs drive at a similar stroke length. The comparison of the main biomechanical variables (Fig.2) shows that Sculler 2 also has a better finish technique: he returns the trunk earlier (a), at the end of the drive (“Finish through the handle” RBN 2006/10), which creates less boat acceleration at the beginning of the recovery (b).

The highest peak of the boat acceleration (c) and velocity (d) during recovery were related to much higher legs (seat) velocity (e) in Sculler 1 due to his longer seat travel. This increases variation of the boat speed and reduces the shell efficiency. As a result, Sculler 2 had a 4.4% lower variation of boat velocity, so his shell efficiency was 1.2% higher, which means a gain of about 5s at 2km race. In fact, the speed of the sculler 2 was 18s faster, (albeit at different weather conditions.) Only the stroke rate and length of seat travel affect the boat speed variation and efficiency; other characteristics of the recovery phase appeared to have no effect on it.

The above finding confirms Volker Nolte’s (2) principle N.4: “The displacement of the centre of gravity (i.e., the seat displacement - VK) in the horizontal plane should be minimised without losing length in the stroke and there should be no lost time with stops or pauses”.

Does this mean that half slide rowing is preferable? Of course not, because this would reduce power production and average boat speed, but the gain is very small. However, the length of legs drive should be optimal and excessive seat travel could be counter- productive.


1.       Sanderson B., Martindale W. 1986. Towards optimizing rowing technique. Med. Sci- Sporls Exerc, Vol. 18, No. 4, pp. 454-468.

2.       Nolte V. 1991. Introduction to the Biomechanics of Rowing. In: FISA Coaching Development Programme Course - Level III. p. 89