D1GP in Japan uses a system called DOSS (d1gp original scoring system). DOSS is a proprietary implementation of the DriftBox. Keeping in mind that the majority of this blog has been tech based, one would assume that I would love the DriftBox. After seeing its usage in D1GP, the response from drivers, and my own research, I’ve decided that the DriftBox is a half baked attempt that hurts more than it helps.
Subjectivity vs Metrics
Drifting has its roots in subjectivity. Inherently, that gut reaction/impression cannot be taken away. Many fans consider putting solid numbers behind judging to be killing the spirit of the sport. They don’t necessarily need to be mutually exclusive. For example, Formula Drift currently uses decel zones as judging criteria. But without the metrics behind it, there isn’t a final truth behind someone brake checking the car behind them.
The DOSS system is a driver’s upper roll cage mounted unit that mounts to the front of a vehicle. According to the spec sheet, it’s no more than a GPS and a 6 axis accelerometer/gyroscope. The GPS module is a 10hz unit that can theoretically do up to 10 samples per second. 10hz is pretty industry standard so the GPS module itself has stood the test of time. The accelerometer also seems pretty industry standard with its accuracy and update rate.
The screen doesn’t look too sexy, but that’s not really that necessary for judging.
Per this Japanese article for the 2019 season, the DOSS system scores by totaling , vehicle top speed, highest angle, maintenance of angle (not wavering), average speed (per section), rate to angle, and rate of change of angle in transitions. This is a pretty big difference in judging criteria in the US. Specifically Formula Drift. The difference is likely due to both the Japanese ideals, and the limitations of the DOSS system.
Breakdown by classic FD judging criteria
Internet searching for how the DOSS module judges line came up with nothing. Their literature says it scores “per section” but I don’t see specifics on how this works. It’s conjecture, but it seems like sectors/sections are done by a remote computer that live analyzes by GPS coordinates and arbitrary lines drawn on a map. By its own literature, being off line (off course) is human judged. So proximity to inner clips and outer zones are not judged by DOSS.
The spec sheet specifies an angle calculation but it’s most definitely wrong. Without external sensors and multiple accelerometers (I can’t find any literature pointing towards this), angle must be judged by the main unit on the front of the vehicle. The distance between the front axis of rotation (between the front wheels) and the unit itself will yield different angle measurements depending on the distance between those two points. These will differ greatly between car models.
Regardless of that, there is no intuitive way for the module to be able to tell if the vehicle is understeering or oversteering. With all my research of angle measurements, no one system is foolproof and each has their own pros and cons. The DOSS system in particular, uses an overly simplified system that can yield very incorrect information.
It’s a GPS. That’s what it does. Speed is the only measurement that is accurate. The 10hz refresh rate does leave something to be desired. But 10 measurements per second is more than enough for basic judging. This is also why the main complaint about the DOSS system is that it heavily favors speed. It’s the only good measurement possible.
Bringing it all together
The total judgement score of the DOSS system is heavily weighted towards speeds and acceleration. Assuming someone attempts to properly calibrate each unit for angle on each car and the flawed angle calculation at least being consistent between vehicles, the angle calculations are mainly used for “highest angle achieved”. The rest of the angle calculations are rate to angle and consistency of angle. These calculations all favor the fastest driver and not the highest angled driver or the most accurate driver.
Many drivers have complained that you “drive to what the box wants”. And that’s pretty accurate. The DOSS system was setup to use the data given to find a measurement that it could produce. The criteria given by the DriftBox is very limiting and judges to a very skewed ideal of what drifting is.