F1Tyrant wrote: ↑
Fri Nov 27, 2020 1:24 pm
Invade wrote: ↑
Fri Nov 27, 2020 1:14 pm
Yeh. I don't take the model seriously. Some strange cooking in there.
I think theoretically it's reasonably valid although most people including myself scoff at Button (11), Sainz (13), Perez (14), Hulk (17) and Rosberg (18) being rated higher than Senna (21) and Prost (20).
I think his list of the top 7 ATGs (Schumacher, Stewart, Alonso, Fangio, Ascari, Clark and Hamilton) is pretty spot on but I'd add Senna and Prost.
I think a well informed and honest 'eye test' complemented with basic stats would arrive at conclusions with less glaring inadequacies than this model has done. Also, the way the model treats the Senna Prost era, and Andrew's justifications and defence for it really don't hold up to any less myopic and more well considered theorising on the dynamics of eras and sports anomalies in general, across sports.
The model is so bound up in itself that it has serious big picture problems, it creates an image which suggests Senna and Prost weren't that good and that their era was weak and presumably it does so by the consistent use of its own integral parameters, and this view is then paraded as being conclusive and providing penetrating insight on the truth of that era. By any measure the model is applied it arrives to the same results, and constant tweaks providing the consistent outlook has given Andrew a quite unshakeable belief in the results.
There are so many variables and such a volatile range of influence in those variables for any model to hope to provide any real sort of 'objective' portrayal of driver ability and competitiveness. But common sense prevails when one considers the sort of massive advantages Prost and Senna often enjoyed, thus separating themselves generally from the field. So what flies, that the ENTIRE era and grid was weak and so this gap in performance must be completely absurd for Senna and Prost to score more highly, or that a couple of anomalies (Senna, Prost) were outstanding enough to separate amid a field which, more or less, remains quite stable in overall ability with slow arcs of average performance variance over time. The basic dynamics of sports and performance which can be analogically drawn across sports simply shows the outcome of this model to be a red flag. It doesn't mean it's definitely 'false', but far more serious thought needs to go into how such an atypical conclusion could be reached, in terms of far-reaching outside factors and moments of significant dynamical change which somehow unseated the model's baseline for the period.
My basic view is that the transitory carryover from driver to driver as to continuously calibrate ongoing driver performance measurement isn't so transitory after all. Rather, it builds in seriously obdurate judgment which causes the model to be maladaptive to unusually impactful 'outside' variables (as well as sudden drops or gains in performance), and that performance is judged with far too great a weight on a diachronic context rather than shorter windows of more synchronic, self contained performance which doesn't rely on the weight of huge strings and stretches of driver comparison across eras. That's my intuition on the biggest pitfall of the model. Because the '80s era 'in and of itself'
,relationally, would not yield such an outlook for Prost and Senna.
There's a strange connectivity in transitory performance between drivers which has handicapped the Prost/Senna era so profoundly that it demanded unprecedented dominance from them if they were to score in the same bracket as the likes of Alonso, Schumacher, Stewart, and even Hamilton, who scores lower than those three.