Huh? The 'drag limited speed' you're talking about is exactly what I'm calculating. In practice, there are only two factors that limit the top speed: rolling resistance and aerodynamic/wind resistance. Rolling resistance is very negligible compared to the latter, so I've let that out of the equation. That leaves the wind resistance.
As a car moves faster, the air molecules will hit the front of the car harder, resisting it to go faster. In the meanwhile, because the car moves at a higher speed, it has covered more distance, meaning it has occured more air molecules as well. That's why the aerodynamic drag is dependent on the vehicle speed to the second power, because of the double 'resisting' effect. But that's forces exclusively. We're talking about power here, which is force times speed and therefore the needed power will be dependent on the speed to the third power. 'The simple threshold that varies from car to car' that you mention would be discribed with this formula.
As for the equation being empirical and therefore imprecise, I have to disagree with you because of two reasons.
Firstly, it is widely taught in engineering schools, so it must not be totally inaccurate. But the main reason is that the C_d coefficient is a dimensionless coefficient. The sole purpose of this type of parameter is to be able to scale the outcome when all other parameters are the same, save for one. The one parameter that's changed is of course the power to calculate a new top speed.
In this case, the car shape and surrounding is exactly the same (of course). That's the thing, the C_d value is a scaling factor, measured and explicitely tailored to this drag formula and therefore the formula is by definition exact (there is no standard deviation). The fact that it is measured (and not composed from theory) is that it is taking everything from real life into account and therefore is perfectly representative of the real world. Of course, this particular C_d value therefore only fits the exact car (2013 Viper) under the exact same real-life conditions (air density). And because it's the same car, the C_d value can be reverse-engineered from the real-world figures whilst at the same time knowing that it's true because those real-world power and top speed figures are true (well, unless Dodge/SRT is lying to us
).
Of course, the conditions are not always exact due to changing airflow due to random wind gusts, but the formula is a very usable one, as GT6 presumably doesn't simulate random wind.
