Ferrari F1-75

A place to discuss the characteristics of the cars in Formula One, both current as well as historical. Laptimes, driver worshipping and team chatter do not belong here.
User avatar
GrrG
86
Joined: 25 Feb 2022, 15:02
Location: Italy Rome

Re: Ferrari F1-75

Post


User avatar
F1DataAnalysis
23
Joined: 10 Apr 2022, 15:34

Re: Ferrari F1-75

Post

F1DataAnalysis wrote:
04 May 2022, 23:56
matteosc wrote:
04 May 2022, 23:24
AR3-GP wrote:
04 May 2022, 22:04


I challenged the legitimacy of this type of analysis on previous pages in the Merc thread but people defended it for whatever reason.
The difference Mercedes/Ferrari is reasonable. Red Bull results are definitely not.
Hi! I too have I doubt about the RB drag coefficient that I wrote in the analysis - I will check tomorrow, I could have calculated the ratio giving the relative drag of the RB compared to the other cars using a wrong value, relative to a previous calculation. This is a doubt I had even before reading your comment. I hope you’ll enjoy the sincerity 😁

In case I will follow up with a corrective post, as it is merely a typo - thank you for sharing it btw!!
I checked, and as I expected, there was a typo! I was looking at the wrong variable to obtain the relative drag of the RB compared to the other two cars - The RB drag was 9% lower than FER (and NOT 15%) and 11% lower than MER (and NOT 17%). It is still a significant difference, but it can now be explained by the inherent lower drag of the car concept and setup choices. A 9% lower drag with 1% less power predicts a top speed around 8km/h higher with DRS open, which is more or less in line with what we've seen so far. The RB really sheds off drag when the DRS is open.

https://twitter.com/F1DataAnalysis/stat ... Qm9ZYHFqBg

Thank you for pointing it out! I am constantly adding filters and conditions to improve the analysis accuracy, and I will post a more detailed description of my methodology on Friday :D
Follow me for F1 Data and Telemetry Analyses: Support the page and request custom analyses:

VacuousFlamboyant
VacuousFlamboyant
7
Joined: 22 Mar 2022, 02:45

Re: Ferrari F1-75

Post

GrrG wrote:
05 May 2022, 00:18
As impressive as Newey feat is, the difference between each concept shouldn't be that great. Around 180km/h (50m/s) porpoising begins to occur. Above 200km/h (56m/s) it severely impacts Ferrari and Mercedes, whereas Redbull deal with it nicely. It's a great representation of how Newey managed to mitigate porpoising with the suspension and floor. Porpoising uncontrolled flows generate drag. Hopefully, the new underside ski skirts and other upgrades bring these figures closer at high speeds.

Contrary to what is believed, Ferrari actually suffers a lot at high speed. That's interesting. Ferrari and Redbull drag ceofficient is evenly matched until porpoising hits hard. I wasn't expecting that. Then again, this is only straight line speed and high speed corners. I'm curious to see the unfiltered data.

AR3-GP
AR3-GP
341
Joined: 06 Jul 2021, 01:22

Re: Ferrari F1-75

Post

F1DataAnalysis wrote:
04 May 2022, 23:56
matteosc wrote:
04 May 2022, 23:24
AR3-GP wrote:
04 May 2022, 22:04


I challenged the legitimacy of this type of analysis on previous pages in the Merc thread but people defended it for whatever reason.
The difference Mercedes/Ferrari is reasonable. Red Bull results are definitely not.
Hi! I too have I doubt about the RB drag coefficient that I wrote in the analysis - I will check tomorrow, I could have calculated the ratio giving the relative drag of the RB compared to the other cars using a wrong value, relative to a previous calculation. This is a doubt I had even before reading your comment. I hope you’ll enjoy the sincerity 😁

In case I will follow up with a corrective post, as it is merely a typo - thank you for sharing it btw!!
There is very little discussion of error analysis in your work. I'm afraid that it leads to hasty conclusions, especially when your tweets are consumed by the layman. Anybody can slap a regression onto any group of points and make it "look good", but if the data set behind it is a mess (like the figures show), then what is the value of the regression when it exist, but it's not good at actually predicting any of the values in the real dataset.

For starters, what are the mean squared errors for your regressions? What is R square for each the regression? Is the error in the regression smaller than spread between the Merc and Ferrari datasets? Aka, are the differences statistically significant or does the regression error make everything a wash?

You have made a good effort so far :D . To be great, I think you need to add some error analysis.

Andi76
Andi76
415
Joined: 03 Feb 2021, 20:19

Re: Ferrari F1-75

Post

VacuousFlamboyant wrote:
05 May 2022, 00:52
GrrG wrote:
05 May 2022, 00:18
As impressive as Newey feat is, the difference between each concept shouldn't be that great. Around 180km/h (50m/s) porpoising begins to occur. Above 200km/h (56m/s) it severely impacts Ferrari and Mercedes, whereas Redbull deal with it nicely. It's a great representation of how Newey managed to mitigate porpoising with the suspension and floor. Porpoising uncontrolled flows generate drag. Hopefully, the new underside ski skirts and other upgrades bring these figures closer at high speeds.

Contrary to what is believed, Ferrari actually suffers a lot at high speed. That's interesting. Ferrari and Redbull drag ceofficient is evenly matched until porpoising hits hard. I wasn't expecting that. Then again, this is only straight line speed and high speed corners. I'm curious to see the unfiltered data.
I think "Concept" is the golden word here and the conclusions of this analysis are wrong. There always are different ways to achieve something. With simple words - one concept puts its focus on aerodynamic efficiency, another one puts its on maximum downforce and the others on the downforce being consistent. Especially with a new regulary. Just remember 2005 - Ferrari, by far the best team in relation to aerodynamics in the early 2000's, had put their focus on maximising downforce, while Renault and McLaren on aero-efficiency. This analysis only shows the teams different approaches, philosophies and trade-offs the teams put their focus on and not that Newey is a genius and much better than anyone else. With all due respect for Newey - he is great and his experience and abilities are worth a lot and a huge advantage and one of the reasons for Red Bulls great performance. But he is not that great that he can design a car thats 10% better in aero-efficiency than anybody else. Its about different approaches, philosophies and trade-offs the teams have made.

Also there are other conclusions one could make. Like, if its really true that Ferrari suffers a lot at high-speed, this is no good news for Red Bull and Newey. Because as soon as they can solve one problem, the F1-75 would be much faster than the RB18. Its already as fast as the Red Bull, sometimes even faster with this problem that makes them suffer a lot at high-speed. But at the end of the day it just shows two different approaches and philosophies in relation to how to achieve lap-time, trade-offs etc. Especially since its no secret that Ferrari was not surprised by porpoising and exactly knew porpoising will appear with ground-effect cars.

User avatar
F1DataAnalysis
23
Joined: 10 Apr 2022, 15:34

Re: Ferrari F1-75

Post

AR3-GP wrote:
05 May 2022, 02:19
F1DataAnalysis wrote:
04 May 2022, 23:56
matteosc wrote:
04 May 2022, 23:24


The difference Mercedes/Ferrari is reasonable. Red Bull results are definitely not.
Hi! I too have I doubt about the RB drag coefficient that I wrote in the analysis - I will check tomorrow, I could have calculated the ratio giving the relative drag of the RB compared to the other cars using a wrong value, relative to a previous calculation. This is a doubt I had even before reading your comment. I hope you’ll enjoy the sincerity 😁

In case I will follow up with a corrective post, as it is merely a typo - thank you for sharing it btw!!
There is very little discussion of error analysis in your work. I'm afraid that it leads to hasty conclusions, especially when your tweets are consumed by the layman. Anybody can slap a regression onto any group of points and make it "look good", but if the data set behind it is a mess (like the figures show), then what is the value of the regression when it exist, but it's not good at actually predicting any of the values in the real dataset.

For starters, what are the mean squared errors for your regressions? What is R square for each the regression? Is the error in the regression smaller than spread between the Merc and Ferrari datasets? Aka, are the differences statistically significant or does the regression error make everything a wash?

You have made a good effort so far :D . To be great, I think you need to add some error analysis.
Thank you! Additional info:
I use many filters, that must be all satisfied at the instant considered: throttle=100%, brake=0, ax>-0.5m/s^2 (to keep negative part of noise at top speed while removing clipping), rpm>10000 and so on. Additionally, I consider the points where is DRS on and/or off only, and I will add an upper limit on the lateral acceleration, as that increases the resistance.
R^2 is around 0.97, and I will add confidence intervals from the next analysis on
Follow me for F1 Data and Telemetry Analyses: Support the page and request custom analyses:

User avatar
Vanja #66
1430
Joined: 19 Mar 2012, 16:38

Re: Ferrari F1-75

Post

For what it's worth, I didn't see any significant difference from Bahrain until now and I think my assumption on drag difference is still valid

Vanja #66 wrote:
20 Mar 2022, 10:10
Decent analysis from reddit, thanks for sharing. I've noticed now a new rear wing on RB, looks a bit shallower than before, so that has a drag influence. Still, based on top speeds (323.2 and 316.6km/h) and assumption of 10HP more power of Ferrari, I got Ferrari with 8% more drag. That's very unrealistic in my view, and quite big. If an assumption is Honda being 10-15HP up on Ferrari, then Ferrari has 4-5% more drag. I think that's much closer to be honest. :)

Ferrari's good acceleration comes from torque, which is in line with their lower engine RPM compared to Honda and Mercedes, so I don't really see that as an indication of higher overall power. Then again, I'm not experienced in telemetry analysis, so I might have all this backwards in my head...
In any case, a lot of it comes from the difference in entire RW design, with RB going for a design that provides a large reduction in drag when DRS flap is open. Ferrari opted otherwise, they are said to bring a new RW for Miami. We might just see them completely level on top speed this weekend.
And they call it a stall. A STALL!

#DwarvesAreNaturalSprinters
#BlessYouLaddie

JPBD1990
JPBD1990
49
Joined: 22 Feb 2018, 12:19

Re: Ferrari F1-75

Post

As is always the case, the proof is in the pudding. It will see how things play out this weekend with respect to Ferrari v redbull and drag. EG will the new RW rumoured bring Ferrari level with redbull on top speed?

If so, will it have other implications like poorer traction or less downforce through high speed (of which there is a lot in Miami)? If that’s the case, will it prove the overall redbull concept to be superior?

If not, how close will Ferrari get? Will they retain their strengths in low speed?

Etc

Also interesting will be to see if progress has been made on porpoising. Latest rumours suggest the floor Carlos tested in the Imola tyre test will only be run in FP to validate the Barcelona upgrade. If that’s the case, how will porpoising look at a track with such long straights?

Exciting times.

matteosc
matteosc
29
Joined: 11 Sep 2012, 17:07

Re: Ferrari F1-75

Post

Vanja #66 wrote:
05 May 2022, 09:06
For what it's worth, I didn't see any significant difference from Bahrain until now and I think my assumption on drag difference is still valid

Vanja #66 wrote:
20 Mar 2022, 10:10
Decent analysis from reddit, thanks for sharing. I've noticed now a new rear wing on RB, looks a bit shallower than before, so that has a drag influence. Still, based on top speeds (323.2 and 316.6km/h) and assumption of 10HP more power of Ferrari, I got Ferrari with 8% more drag. That's very unrealistic in my view, and quite big. If an assumption is Honda being 10-15HP up on Ferrari, then Ferrari has 4-5% more drag. I think that's much closer to be honest. :)

Ferrari's good acceleration comes from torque, which is in line with their lower engine RPM compared to Honda and Mercedes, so I don't really see that as an indication of higher overall power. Then again, I'm not experienced in telemetry analysis, so I might have all this backwards in my head...
In any case, a lot of it comes from the difference in entire RW design, with RB going for a design that provides a large reduction in drag when DRS flap is open. Ferrari opted otherwise, they are said to bring a new RW for Miami. We might just see them completely level on top speed this weekend.
One thing that the "twitter analysis" does not consider is the different power deployment strategy of the different power units. It looks to me like Red Bull uses more power at the end of the straight (high speed) while Ferrari uses more power at the beginning (high acceleration). Not taking this into account would lead to underestimate Red Bull drag.

Sevach
Sevach
1058
Joined: 07 Jun 2012, 17:00

Re: Ferrari F1-75

Post


User avatar
GrrG
86
Joined: 25 Feb 2022, 15:02
Location: Italy Rome

Re: Ferrari F1-75

Post

Image

ryaan2904
ryaan2904
36
Joined: 01 Oct 2020, 09:45

Re: Ferrari F1-75

Post

Vanja #66 wrote:
04 May 2022, 21:49
GrrG wrote:
04 May 2022, 18:43
Yeah... that would make RB18 reach 350, while others couldn't reach 320kmh...
He corrected it, its 9% drag advantage, not 15% over ferrari
CFD Eyes of Sauron

.poz
.poz
50
Joined: 08 Mar 2012, 16:44

Re: Ferrari F1-75

Post

matteosc wrote:
05 May 2022, 13:55
One thing that the "twitter analysis" does not consider is the different power deployment strategy of the different power units. It looks to me like Red Bull uses more power at the end of the straight (high speed) while Ferrari uses more power at the beginning (high acceleration). Not taking this into account would lead to underestimate Red Bull drag.
I think it's the gear ratio: this year Ferrari uses shorter gears, similar to Mercedes

matteosc
matteosc
29
Joined: 11 Sep 2012, 17:07

Re: Ferrari F1-75

Post

.poz wrote:
05 May 2022, 15:14
matteosc wrote:
05 May 2022, 13:55
One thing that the "twitter analysis" does not consider is the different power deployment strategy of the different power units. It looks to me like Red Bull uses more power at the end of the straight (high speed) while Ferrari uses more power at the beginning (high acceleration). Not taking this into account would lead to underestimate Red Bull drag.
I think it's the gear ratio: this year Ferrari uses shorter gears, similar to Mercedes
Based on pre-season and in-season analysis, it looks like Ferrari uses longer gear than Mercedes (lower rpm for Ferrari). But this is not what I am referring to: I am talking about the different strategy of energy deployment, with Ferrari using more electric power at the beginning of the straight and then clipping at the end. Red Bull on the contrary seems to use their electric power more evenly, with little clipping at the end of the straight. Not taking this difference into account may lead to underestimate Red Bull's drag.

Andi76
Andi76
415
Joined: 03 Feb 2021, 20:19

Re: Ferrari F1-75

Post

matteosc wrote:
05 May 2022, 13:55
Vanja #66 wrote:
05 May 2022, 09:06
For what it's worth, I didn't see any significant difference from Bahrain until now and I think my assumption on drag difference is still valid

Vanja #66 wrote:
20 Mar 2022, 10:10
Decent analysis from reddit, thanks for sharing. I've noticed now a new rear wing on RB, looks a bit shallower than before, so that has a drag influence. Still, based on top speeds (323.2 and 316.6km/h) and assumption of 10HP more power of Ferrari, I got Ferrari with 8% more drag. That's very unrealistic in my view, and quite big. If an assumption is Honda being 10-15HP up on Ferrari, then Ferrari has 4-5% more drag. I think that's much closer to be honest. :)

Ferrari's good acceleration comes from torque, which is in line with their lower engine RPM compared to Honda and Mercedes, so I don't really see that as an indication of higher overall power. Then again, I'm not experienced in telemetry analysis, so I might have all this backwards in my head...
In any case, a lot of it comes from the difference in entire RW design, with RB going for a design that provides a large reduction in drag when DRS flap is open. Ferrari opted otherwise, they are said to bring a new RW for Miami. We might just see them completely level on top speed this weekend.
One thing that the "twitter analysis" does not consider is the different power deployment strategy of the different power units. It looks to me like Red Bull uses more power at the end of the straight (high speed) while Ferrari uses more power at the beginning (high acceleration). Not taking this into account would lead to underestimate Red Bull drag.
Today people tend to think they can do a proper analysis of complex things(like a F1 car) with few data only. Lets be true - this is not possible. Gear Ratios, engine settings, deployment stragtegies are just a few simple things here. So - you cannot say anything about efficiency with data like in this analysis.