Emag wrote: ↑17 Sep 2026, 19:59
Well don't take my words in such an absolute stance. When I say top speed, I am not literally saying the single data point where the highest speed in a straight is achieved. My point is that McLaren haven't had traces that suggest "good" average time in the straights. Not even against some midfield teams, let alone Mercedes. I don't see how that graph makes sense. McLaren has not been the second best team in the straights throughout this season.
Looking at traces (literally) is the wrong approach, I think? We cannot see time on straight or slightly better acceleration by looking at a fastf1 velocity trace. It's beyond human perception, especially when you have to take into account differences in deployment when you look at a single straight. Anyone claiming to have such powers, I wouldn't believe them.
2nd, I looked at the methodology in your post.
https://i.postimg.cc/pLSB8M89/image.png
You are measuring different distances for every single car? You cannot do time on straight using different distances for every car. It's not normalized. In the best case, you could take the earliest braking point of any car and use that to fix the distance over which you report the "time on straight". That's how I would do it. I would also only use Merc, Mclaren, Alpine, RBR, VCARB, Audi, Ferrari. The others are so far behind on the chassis that it's going to mess things up too much.
I have no idea if that's the approach used by laptimeF1, and perhaps you will call it confirmation bias, but I find it interesting that an approach which measures time on straight would group 3 Mercedes powered cars at the top, followed by RBPT, followed by Ferrari, and that would also be exactly the order accepted in the paddock.
The figure that you shared actually looks closer to a top speed map, for whatever reason. Maybe the issue is in what I discussed with you using different straight distances for each car, because they have different braking points. Then there's the whole 5 track vs all of them thing. It could also make sense not to use the first 3-4 GP since it's too noisy with the deployment learning. You could just do the same tracks as the laptimeF1 and see what happens.
One type of analysis that you could do to make Mclaren's acceleration more obvious would be to just use the first half of the straight, or 75% and see what you get out of that (top 7 in WCC).
Working on the balance.