Blade Attachment Boosts Rowing Propulsion by Reshaping Leading-Edge Vortices

In competitive rowing, where races are sometimes decided by fractions of a second, even small changes to equipment can matter enormously. A new laboratory study has now put one of the sport's most talked-about accessories under the microscope: the RANDALLfoil, a small wing-like attachment fixed along the top edge of a standard rowing blade. By driving scaled blades through realistic rowing motions in a water tank and simultaneously measuring both the forces on the blade and the swirling flow structures around it, researchers found that the attachment increased propulsive impulse by 9.01% and impulse efficiency by 4.91% for a representative scaled motion, offering a detailed hydrodynamic explanation for the performance gains that rowers have reported anecdotally.

The RANDALLfoil first entered the international spotlight at the 2019 World Championships, where Italian professional rower Martino Goretti used the attachment and won all of his qualifying races as well as the final. Rowers who adopted the device reported a distinctive set of sensations: less slip of the blade through the water during the drive phase, less unwanted pitch of the blade, and a so-called 'lock effect' felt at the catch, the moment the blade enters the water. Until now, however, most studies of the attachment have focused on the athlete rather than the fluid.

Tethered rowing experiments had reported higher force production and altered metabolic responses when rowers used Big Blades fitted with RANDALLfoils, and 500-meter on-water sprints had shown shorter completion times and higher mean boat velocity with rowing rate and physiological variables remaining similar between blade conditions, but these approaches could not isolate the underlying hydrodynamic mechanisms. To disentangle the physics from the athlete, the research team turned to a controlled laboratory reproduction of the rowing stroke. A rowing stroke consists of three phases: the catch, when the blade enters the water; the drive, during which the blade rotates about the oarlock and follows a curved path while the boat translates forward; and the release, when the blade exits.

The new study focused exclusively on the submerged drive phase, using 1:4 geometrically scaled blades modeled after the Concept2 Big Blade and Smoothie2 Plain Edge sweep designs, 3D-printed with an adjusted thickness of 3 millimeters to ensure stiffness, and mounted on a robotic rig that reproduced blade-tip paths and oar angles measured from elite men's coxless four race-pace rowing. The experimental apparatus combined a six-axis force and torque transducer sampled at 10 kilohertz with two linear stages and a stepper motor that drove the blade through its prescribed translation and rotation. Each run took place in a 610 by 610 millimeter glass water tank with a water depth of 420 millimeters, with the blade fully submerged 30 millimeters below the free surface to minimize surface reflections.

The team calculated a surface blockage ratio of 2.31%, small enough to have a negligible effect, and allowed the water to quiesce for 15 minutes between runs. A pitch angle of zero degrees was used, consistent with earlier work. Forces were low-pass filtered with cutoff frequencies of 20 hertz for the faster motion and 5 hertz for the slower one, ensemble-averaged over four runs, and decomposed into components normal and tangential to the blade, which were then resolved into the propulsive direction of boat motion and the non-propulsive lateral direction.

The kinematics were tested at two velocity scalings, a 1:4 motion with a path time of 0.8 seconds and a slower 1:8 motion with a path time of 1.6 seconds, based on race-pace blade kinematics from the literature. The hydrodynamic forces scaled approximately with the square of the velocity scaling factor, confirming that the laboratory results could be compared across conditions. The Froude numbers for both experiments and on-water rowing were subcritical, but the Reynolds numbers remained an order of magnitude below full-scale rowing, a limitation the team acknowledged explicitly.

Alongside the force measurements, the researchers used two-dimensional particle image velocimetry, illuminating a laser sheet at the mid-height of the blade and imaging the seeded flow from beneath the tank at 155 hertz. Eight fields of view were stitched together to follow the blade along its long path, and the team applied Graftieaux's criteria to identify the center and boundary of the leading-edge vortex, a coherent swirling structure that forms near the blade early in the drive. From these fields they extracted the vortex circulation magnitude, its growth rate, its area, and the distance between the vortex core and the blade surface, tracking how each quantity evolved through the stroke until the vortex left the measurement frame at 60% of the motion.

The force results revealed a clear temporal pattern. Early in the drive, the RANDALLfoil produced lower or comparable propulsive force than the standard blade, but from roughly 60% of the way through the motion onward it generated consistently higher propulsive force. For the representative 1:4 scaled motion, this late-drive advantage outweighed the early reduction, producing the 9.01% increase in propulsive impulse and the 4.91% increase in impulse efficiency, a measure of how well the generated impulse aligns with the direction of boat motion.

For the slower 1:8 motion, the same trend appeared with smaller gains of 4.10% in propulsive impulse and 2.09% in impulse efficiency. In the context of on-water rowing, these gains are consistent with greater momentum transfer in the propulsive direction and reduced blade slip through the water, the very effect that rowers have long described subjectively. The pitching torque measurements added a further layer of nuance.

Although the total integrated pitching torque over the full drive showed a negligible difference between the two blades, with overlapping error bounds, the RANDALLfoil produced consistently lower pitching-torque magnitude during the first half of the stroke. The authors suggest this early reduction may offer a physical explanation for the improved blade stability and reduced pitch sensation reported by rowers in previous biomechanical studies. The attachment's larger late-drive force, meanwhile, was accompanied by increased pitching torque late in the motion, and the researchers note that the higher late-drive loading is consistent with reports from rowers that the RANDALLfoil feels 'heavier' in the second half of the drive, a sensation that may reflect genuinely increased useful hydrodynamic loading rather than a drawback.

Perhaps the most scientifically interesting finding concerns the vortex dynamics. Changes in the blade-normal force could not be explained by leading-edge vortex circulation magnitude alone. Instead, the force response was best interpreted by considering circulation magnitude, circulation-growth rate, vortex area, and the distance between the vortex core and the blade together.

The early force drop after the start-up peak coincided with weakened circulation growth, while the later recovery of blade-normal force coincided with a rapid increase in vortex area and growth rate for the RANDALLfoil. At intermediate times, the attachment's vortex sat farther from the blade with lower circulation, matching a smaller force magnitude, evidence that the RANDALLfoil modifies blade loading by changing both vortex development and the coupling between the vortex and the blade surface. This interpretation builds on earlier work showing that rowing propulsion is not purely drag-based.

For a standard blade orientation, lift contributes roughly 40% of the propulsive impulse, with drag supplying the remaining 60%, and lift dominates until nearly halfway through the drive as the leading-edge vortex remains close to the blade surface. The present study's demonstration that vortex proximity, area, and growth rate jointly govern the force response aligns with a broader literature on accelerating and oscillating plates, in which leading-edge vortex loading depends on multiple flow parameters rather than circulation alone. The authors are careful to spell out the study's limitations: only the submerged drive phase was captured, the PIV measurements were two-dimensional and taken at the blade mid-height, so three-dimensional flow, tip effects, and spanwise vortex structures were not resolved, and Reynolds numbers were far below full scale, with the slower motion producing impulse efficiencies well below the value of approximately 0.84 considered representative of full-scale rowing.

Future work, the team writes, should examine the full stroke, three-dimensional flow structure, and higher-Reynolds-number conditions.

Subject of Research: Hydrodynamic effects of the RANDALLfoil rowing blade attachment during scaled rowing motions

Article Title: Effect of RANDALLfoils on the fluid dynamics of rowing

Article References: Cereck, W., Young, C., & Morris, S. (2026). Effect of RANDALLfoils on the fluid dynamics of rowing. Sports Engineering, 29(2), Article 39. https://doi.org/10.1007/s12283-026-00572-5

Image Credits: AI Generated

DOI: 10.1007/s12283-026-00572-5

Keywords: rowing, RANDALLfoil, hydrodynamics, leading-edge vortex, particle image velocimetry, fluid dynamics, sports engineering, blade slip, impulse efficiency, propulsive force, water tank experiment, Effect