Field sobriety tests are designed to be failed though. They are an evidence builder, by and large.
The fact that it is so difficult to design reliable FITs (field impairment tests), to me, says a lot about the questionability of sobriety testing in general.
If we can't measure impairment, how can we justify x% as an "impaired" alcohol level ?
The problem with associating impairment with levels of some chemical is that there's always going to be a new chemical on the scene - and people can have wildly differing reactions to a lot of chemicals. Bring on the self-driving cars!
I have always thought a "breathalyzer" should be some mechanical device that measures the same skills one needs to capably drive a car. If you can not adequately handle that then you can not drive no matter what the "impairment"; whether it be booze, pot, old age or simply some other cognitive or motor skill issue. And although I would like to see it happen faster, even google, benz, etc. admit that fully autonomous cars are a long way away.