Simulator for drawing historical bows: Butnariu 2018

The paper presents a virtual reality simulator designed to let the public, in museums or at fairs, feel what it is like to draw three historical bows: an English longbow, a flatbow and a horse bow. Instead of a bow, the user holds a wooden base carrying an electric motor that winds a cord: pulling the cord gives a resistance that rises with draw length, modelled on the force-draw curve measured on real replicas. The three curves are close to straight lines; written as F = a·x + b, with x in millimetres, their slopes are 0.2498 for the horse bow, 0.2837 for the flatbow and 0.3001 for the longbow.

Twenty people aged 19 to 62 shot the three replicas and then the simulator set to the three curves, and filled in a questionnaire scored from 1 to 7 on engagement, ease of handling, enjoyment, realism, usability and overall experience. According to the authors the ratings were largely positive and users could tell the three settings apart; all of them, however, could easily tell whether they were shooting the real bow or the simulator. The setting judged closest to the original was that of the "second bow", and the most common complaint concerned the weight of the device, awkward especially for women. Before the study, two experts from the History Museum of Brașov had tried the system: they found it realistic but noted that the feel of the string was missing.

The replicas, all fitted with Dacron strings, were measured on a Tinius Olsen H100KU testing machine: the bow was held in a vice and the string, attached to a moving hook, was drawn in small steps up to 50 cm while the force was recorded at each step. Stated draw weights are 13.6 kg for the longbow (177.8 cm long, weighing 1.3 kg), 12.5 kg for the flatbow (172 cm, 1.2 kg) and 18.1 kg for the horse bow (121.9 cm, 0.5 kg). The device uses a Maxon EC-Powermax 30 motor with its EPOS 70/10 controller and a 20 cm pulley; an encoder in the motor measures how much cord has been drawn and the software sets the current to give the matching force. The virtual scene, built in Unity and viewed through an Oculus Rift DK2 headset, computes the arrow's flight assuming that about 70% of the stored energy goes into the arrow.

Each participant made 20 attempts with the real bows and 20 with the simulator, in two sessions separated by a short break, with no time limit and no scoring: where the arrows landed did not count. The questionnaire has 43 questions; for the negatively worded ones the answer was subtracted from 8. Five of the twenty participants had used haptic devices before.

The questionnaire results are shown only as charts, with no means, standard deviations or statistical comparisons, so the authors' "positive" verdict cannot be weighed. There is no comparison group, the order of the sessions was the same for everyone (real bows first, then the simulator), and before shooting the participants heard a short story about each bow, told to raise their interest: all of these can inflate the ratings. We are not told how many of them already shot.

The simulator reproduces only the force during the draw: there is no string to feel, and at release there is just the cord winding back and a vibration. The replicas have about a third of the draw weight of the historical bows, as the experts themselves point out, and the idea that a Dacron string behaves like one made of natural fibres is an untested assumption. The text has some inconsistencies: the measurement is described for "both" replicas when there are three, the experts say they told "the 2 bows" apart, the unit of force in the linear fit is not given, and it is unclear which bow is the "second bow" the users preferred.

For anyone who shoots, the study changes little. It measures nothing about the shot itself, neither aiming nor release nor accuracy, and it was not built for training: the stated aim is an installation for museums and fairs that lets the public feel the differences between historical bows. The most concrete finding concerns those bows: up to 50 cm of draw, the force of the three replicas rises almost linearly, without the let-off that the cams of a compound give, and the authors name a comparison with the compound as their next step.

The result of most interest to an archer is an indirect one. All twenty participants said they could easily tell the real bow from the simulator, even with the same force curve. The study does not say what gave it away; the weight of the bow in the hand, the string on the fingers and the reaction at release, none of which the device reproduces, are the obvious candidates. On this evidence, reproducing the draw force alone is not enough to recreate the feel of a shot.