Track tests alone won't reveal everything. That's why its still worth doing more standardised tests.
Well, if you create a prefectly balanced setup, it won't be perfect. In fact, it'll be undrivable due to shear lack of ANY stability. As to invalidity of standard tests, tell that to the Lancer Evolution VIII MR. At 0.98g (skidpad) and 70+mph (200ft slalom), it .... guess what, it's brilliant on mountain passes and anything that resembles the twisty nature of the monte carlo rally.
In comparison, the McLaren F1 posted 0.86g and 64.5mph (from Road&Track). It's faster on your average high speed F1 style track, but try doing the Japanese mountain passes with it. Truth is, it understeers heavily at any speed that doesn't allow a lot of downforce generation. And guess what, that's exactly what the numbers show.
So much for track tests only mentality. The truth is, both tests are necessary for those who want to thoroughly understand and tweak their cars. Racetracks tend to hide or overamplify some characteristics of your car, that's why standard tests are still necessary to attain a sort of basic balance. If it tests well, it also goes to prove if it is the basic balancr or the driving that is at fault. The problem with track tests only is that cars get overcompromised setups to compensate for some driver deficiency. This will not do the any good t the cars potential lap times.