adrianjordan wrote: ↑
Wed Nov 28, 2018 2:23 pm
Edax wrote: ↑
Tue Nov 27, 2018 11:13 pm
M840TR wrote: ↑
Tue Nov 27, 2018 5:11 pm
Perhaps they revised their model of the current car and compared that to the on track data. If you remember, they spent a lot of free practice sessions running with sensor arrays etc on the car.
Probably but it is one thing to fit a model to an existing situation, and another to have the model predict a new design. The one does not necessarely guarantee the other.
Imagine that you try to fit a parabola with a straight line. At any point of the parabola you can get a fit by adjusting the slope and intercept. You can walk up and down a bit and you would be perfectly convinced that the world is flat and sloped. It is only once you make a big step that you discover the flaw in your model as you fall off the parabola.
Perhaps I am a bit too skeptical when it comes too modelling but I have too often seen engineering teams bring themselves into troubles by relying to much on models and functional testing.
I would be more at ease if I would have seen the Mclaren running with makeshift aero devices ductaped to it.