Suggestions for EXPERIMENTAL SERVICE LAW - 115 cm

[pcll99]

The experimental service law has been in place for quite a while now. I have a number of suggestions for improvements:

1. The two-sheet device should be calibrated between every games. After each game, a free standing pole should be placed at the court by the service judge himself without the help of others. After placing the pole there, he can return to his seat to calibrate the two-sheet device.

2. BWF should conduct a contest for the best smartphone app which can make the most accurate service fault call. The winner should receive US$10,000 while runner-up should get $5,000. The BWF shall retain the copyrights and all exploitation rights to apps of the winner and runner-up.

 

[phihag]

2. BWF should conduct a contest for the best smartphone app which can make the most accurate service fault call. The winner should receive US$10,000 while runner-up should get $5,000. The BWF shall retain the copyrights and all exploitation rights to apps of the winner and runner-up.

Application development is not really something you can easily make a competition out of: A competition attracts hobbyist developers, not professionals. If some hobby programmer already had a solution for this problem, they would have offered it to BWF or other badminton bodies by now. BWF also needs an ongoing relationship, for instance for porting to newer versions, dealing with different lighting conditions, etc. . The device likely needs to have some kind of stand, dedicated power, and other setup configuration, which cannot be done easily by arena workers.

I'm thinking about developing something similar myself (although for line calls), but other projects and insufficient hardware is currently the hurdle for me. The chance to win some money would not change much for me.

But even if the competition would yield sensible entries, what BWF really needs is not the best, but better than human. A competition might result in two apps that work flawlessly 80% and 70% of the time, which would be totally unacceptable for actual use. Iterative improvements on that may not ever come to something surpassing humans.

Smartphones are ubiquitous and very cheap, but not necessarily geared towards the workload this would entail. If you use pretty much 100% GPU and CPU for hours at a time, you may have way higher failure rates than one would usually associate with a smartphone.