Shaft straightness and arrow grouping: Park 2017 study

A team from four countries, coordinated by Park, used a shooting machine and a compound bow to measure how grouping at 25 metres changes when shafts are less straight, when the nock changes and when the fletch angle changes. The group is the diameter of the smallest circle containing all the holes. With Easton X7 2315 aluminium shafts straight to within a thousandth of an inch and vanes set at 0.9 degrees, the group was 27 mm; bent to three thousandths it grew to 30 mm, and to six thousandths 48 mm. With straight vanes, no angle, the figures were 39 mm for straight shafts and 48 mm for shafts bent to three thousandths; with vanes at 2.1 degrees, 20, 28 and 40 mm. Between two carbon sets of the same model, too, the LightSpeed 3D, straight to one thousandth, grouped in 30 mm against 42 mm for the standard LightSpeed, straight to three.

The surprise was the nock. On straight X7 2315 shafts, with G nocks and a second adapter the group was 44 mm, with 3D Super nocks 33 mm and with MicroLite Super nocks 27 mm. On X7 2712 shafts, switching from 3D Super to MicroLite Super nocks took the group from 32 to 15 mm, and X7 2512 shafts, fitted with precision adapters made specially for G Pin nocks, grouped in 13 mm, only 3.1 mm more than the shaft diameter. The authors suggest that what matters is the length of plastic between the end of the shaft and the string groove (about 3.7 mm on the MicroLite, 11.7 mm on the 3D Super) and the concentricity of the adapters. Protour 380 arrows grouped in 25 mm, and a single Protour shot repeatedly in 15 mm.

For X7 2315 shafts with Diamond 235 vanes, the calculation gives about 8.4 rotations over 25 metres at 0.9 degrees and 19.6 at 2.1 degrees, with the first full turn at 6.9 and 4.4 metres.

The tests were run in Melbourne with a Hooter Shooter machine bolted to the ground and a Hoyt Pro Comp Elite compound with Spiral X cams, a 700 mm draw length and a 240 newton peak, with a hard draw stop holding draw length constant to within 1 mm. Aiming used a small peep and a black dot on the sight lens, with a bubble level to keep the bow upright, and the bow was re-aimed before every shot; the authors estimate the aiming error at within 5 mm at 25 metres. The bow had been tuned with a paper test, indoors or outdoors in a sheltered spot with no wind. Each arrow of each set was shot 10 times per configuration.

X7 aluminium shafts were chosen because they can be straightened and bent: 18 2315 shafts split into four sets, two to stay straight and two to be bent, each pair with straight vanes and vanes at 0.9 degrees. The shafts were bent at the centre with an Arizona Archery straightener, just to the yield point, and rechecked before each session because they tended to spring back slightly; within a set of four the bends were oriented at 0, 90, 180 and 270 degrees, to catch the worst case. For comparison the team tested X7 2213, 2512 and 2712 shafts, two sets of 400-spine LightSpeed with different tolerances, and Protour 380s. Arrow rotation was calculated with a model Park had published earlier.

The group is described by a single number, the diameter of the circle containing all the holes: the authors themselves say that recording the position of each arrow would have allowed a statistical analysis but would have needed a new target face for every shot. There are therefore no uncertainty intervals or repeated trials, and some differences of a few millimetres sit close to the estimated aiming error. The table is incomplete: shafts bent to six thousandths were tested only with angled vanes, and the other shaft sizes only when straight.

Everything was done with a compound bow and a machine, at 25 metres, on aluminium shafts bent on purpose; carbon straightness cannot be altered, so the comparison between the LightSpeed and the LightSpeed 3D mixes straightness with other manufacturing differences. The role of the nock emerged by chance, and the explanation about the length of plastic is the authors' hypothesis, not something they tested. That a skilled archer would match the machine's groups is an expectation based on separate tests by one of the authors, not a finding of this work.

For anyone shooting, the most concrete result is that three small things affect the group more than one might expect. Stated straightness matters: at three thousandths of an inch the group already opens up visibly, and the authors note that such an arrow, spun on the fingernails, can be told apart by feel from one straight to within a thousandth. Fletch angle matters, because an arrow that spins in flight partly cancels its own defects: with straight vanes the group was always the widest. And the nock matters, which in these tests changed the group almost as much as straightness did.

The authors draw a few suggestions from this: group-test arrows on a shooting machine before important competitions, set aside those that land outside, and rotate the nock on outlying arrows to see whether that is where the problem lies. These are reasonable pointers for anyone with similar compound equipment, but the study measures machine groups at 25 metres. How much of these millimetres would still show in an archer's groups, with the archer's own errors on top, depends on the archer's level: for someone with a wider spread, the difference may simply disappear into it.