Why do drivers still have to set up a car?

[autogyro]

Sorry gatecrasher but at the level we are discusing the drivers can adapt to practicaly anything.
There is no comparison at all with a go kart.
I think this high driving level has to be taken into account with any technical subject in F1.
The difference between the best driver and the worst out there is so close as to be irelevent to comparisons in most other formula.

[Jersey Tom]

Why? "Closed loop" data is a bitch to sort through and doesn't always tell you everything.. or anything.

Good driver feedback = win.

That all depends on how good the person or persons is at analysis. "Everything" is all there, it's just a matter of whether it gets lost in translation or not translated at all.

Good driver feedback minus data feedback will not beat good data feedback with driver input to the data.

A driver with terrible feedback and an excellent data analysis can outperform a driver with only his own feedback available, even one who has excellent recall.

It is possible to separate the driver, from his inputs to the car, his inputs because of the racetrack and the car from the driver inputs and the car from the track, given the "right" sensors in place. I've yet to see a F1 car without them.
The causes,the amount, the location and the length of time of any handling situation can be accurately tracked and analyzed, even before the driver pulls into a pit stop. He simply will be repeating what the analysis already shows...

Sorry but the amount of information that can be had, far out performs what the human brain is capable of ....driver's with great feedback isn't a requirement to win anymore and hasn't been for 15 years. IMHO

I used to think the same, until I started working with more and more on-track data. Very extensive, very well-recorded data. Just about every sensor you can buy.

And yes.. there are series where you absolutely need good driver feedback to win.

Other thing about DAQ... even if it tells you WHAT the car is doing, which will generally jive with the driver feedback anyway.. often doesn't clue you in as to why. That's the really sucky part. The 'what' is easier than 'why' is easier than 'how (to change it)'. There's a reason why teams and drivers will say they just can't get the balance or X or Y or Z right during a race. DAQ doesn't tell you everything.

Track telemetry better for engine tuning than handling IMO.

[marcush.]

+1 not quite sure if i would go to these extremes but the god drivers have a third sense for the area wich needs to be worked on..
A Truckload full of data can be a real pain in the ass to scrape and find out what really is restraining the car..

[speedsense]

Why? "Closed loop" data is a bitch to sort through and doesn't always tell you everything.. or anything.

Good driver feedback = win.

That all depends on how good the person or persons is at analysis. "Everything" is all there, it's just a matter of whether it gets lost in translation or not translated at all.

Good driver feedback minus data feedback will not beat good data feedback with driver input to the data.

A driver with terrible feedback and an excellent data analysis can outperform a driver with only his own feedback available, even one who has excellent recall.

It is possible to separate the driver, from his inputs to the car, his inputs because of the racetrack and the car from the driver inputs and the car from the track, given the "right" sensors in place. I've yet to see a F1 car without them.
The causes,the amount, the location and the length of time of any handling situation can be accurately tracked and analyzed, even before the driver pulls into a pit stop. He simply will be repeating what the analysis already shows...

Sorry but the amount of information that can be had, far out performs what the human brain is capable of ....driver's with great feedback isn't a requirement to win anymore and hasn't been for 15 years. IMHO

I used to think the same, until I started working with more and more on-track data. Very extensive, very well-recorded data.

There's a "major" difference between extensive, very well-recorded data and analysis of said data. 1% of the impact of data to a team/driver is "correct" collection of data. (setup,calibration,operation of the system) 99% of it is analysis, and of that 99% the majority of data people (85-90%) analyze data in the least effective way. That is absolute analysis (peaks and valleys, specific numbers) While this type of data analysis is common, the easiest learned and "Very" useful for an engineer to fulfill his math equations, the impact for a driver and his technique is almost nil. There are several techniques that are less common, yet have enough impact that they over ride the recall process of a driver by almost 100 fold.
Yes, I agree the amount of data present can overwhelm the inexperienced analyst, but it's all a matter of "known" techniques AND the "added" ability of the software package it's self to present it in an understandable fashion that the driver can relate to.
Personally speaking, I have yet to meet, out of the some 60+ drivers I've worked with, that can out recall even by 10 %, a competent analysis on their previous outing. Though their recall after wards and their performance that follows the analysis in the very next session, is 50 times improved over being without it.

And yes.. there are series where you absolutely need good driver feedback to win.

Specifically talking series were data systems (allowed) are in use during an event. As I presume your speaking Nascar as example, need I remind you that when they used to be allowed test, the testing was almost always with a DA system and even today when a team has data from a racetrack, that data is being run in a Sim program to help judgment of the forthcoming changes that the driver/crew chief are prescribing..... it ain't all the driver or even the crew chief there either....

Other thing about DAQ... even if it tells you WHAT the car is doing, which will generally jive with the driver feedback anyway.. often doesn't clue you in as to why. That's the really sucky part. The 'what' is easier than 'why' is easier than 'how (to change it)'. There's a reason why teams and drivers will say they just can't get the balance or X or Y or Z right during a race. DAQ doesn't tell you everything.

Strongly disagree, when the analysis is done correctly, it tells Why, What,Where, When and How very clearly. The ease of it has everything to do with who is doing the analysis and what they know how to do.
I do agree however data analysis doesn't tell you everything, it's doesn't tell you how to fix it. What it does do is lead to the right path being chosen, instead of the driver leading you down the wrong one often.

Track telemetry better for engine tuning than handling IMO.

Sure preventing an engine disaster is always good especially when it catches an impending failure. However when a pit stop is highly time limited, say in a 24 hour race, changes to aero and handling, and knowing what to change before the car gets to the pit, is the difference between winning and going several laps down while you interview the driver.

As Always In My Humble Opinion

[Jersey Tom]

You'd have to prove it to me

The more I work on this crap, the more skeptical I am.

[speedsense]

You'd have to prove it to me

The more I work on this crap, the more skeptical I am.

I understand, the data collected can be immense. And with someone doing analysis,that can give only numbers and absolutes, makes it even worse because drivers don't drive on numbers. While these numbers help engine builders and help tire engineers and help the aero guys, that don't help the recall process one bit.

With a steering sensor, throttle sensor, Lat-long G,Speed, and eng rpm, all running at least 20 samples per second (on at least a 12bit processor in the data system) and when they are combined together, there isn't one movement the driver can make without seeing it. The driving technique is completely exposed, as is the handling of the car exposed with understeer the hardest to define (the definition of understeer is different per each driver,)
Add in suspension sensors (at a minimum) and there isn't much the driver can make the car do, that you can't see (combined with the other "driver" sensors). As well as what the car (or track) makes the driver do. The responses to each are very clear.
Additional sensors only make the picture more, more clear, and the path crystal clear instead of just bright...

Now consider on more thing, on a road course, a driver may make on avg 300 decisions a lap. In ten laps, that becomes 3000 decisions. Without data, the driver (with great recall) may remember twenty of those decisions, mostly the ones where he saved his rear end. That's less than .01% recall of the decisions made. Data makes it possible to recall 90% of those decisions, good ones as well as the problem one....
Additional sensors only make the picture more, more clear, and the path crystal clear instead of just bright...

[PlatinumZealot]

The driver's feedback is very necessary I feel.

Remember Button said that the MP4-25 was "undriveable" on low fuel in Silverstone, while Hamilton was content with the car qualifying Button by 1 second. These are two drivers with similar setups, but Button is not good at controlling over-steer. The engineer probably looked on Button's data and say "why are you making the car do that? why are you so slow, the data is showing the steering doing this, and suspension doing that."
Now the engineer can just say to Button "Just make it like Hamilton's, do your steering like this, your brakes like that, change dampers this, wing to that etc make it like Hamilton's."
Button is the driver so he can now come in and say "No I don't like doing that." or "I cannot drive like that. I can only turn my wheel like so and I can only tolerate the setting at this level."
So even though the engineer knows what the problem is, he cannot solve it without feedback from the driver because the driver is his only link to get the performance from the car. Even if the engineer sets the car up perfect, it won't do any good if the driver can't drive it. Using the same example, the engineers could have given Button Hamilton's setup which was 1 second faster, without consulting Button but more than likely Button would have ended up going slower.

[Gatecrasher]

Sorry gatecrasher but at the level we are discusing the drivers can adapt to practicaly anything.
There is no comparison at all with a go kart.
I think this high driving level has to be taken into account with any technical subject in F1.
The difference between the best driver and the worst out there is so close as to be irelevent to comparisons in most other formula.

Sorry I was not trying to compare an F1 driver to my poor skills rather that each driver no matter how good has a style that is the best for them. Sure you can buy the best supercomputer in the world but without the best people behind it the data it analyses is not the best. I work with engineers and statisticians at my work, I can get 10 different ways to analyze the same data and get different assumptions from the same data. Until a computer can drive as good if not better than a human then you will always have to accept the driver and engineers gut feel.

Each year computers get better, that's part of my job, I just hope it does not happen in my lifetime.

[Gatecrasher]

FYI, I also enjoy tracks days on motorcycles. My bike has downloadable DA. I initially looked on it more as a toy however when I started to discuss it with racers I quickly found where I was loosing the most time on the track. My lap times took a considerable drop. It does work.

[speedsense]

The driver's feedback is very necessary I feel.

Remember Button said that the MP4-25 was "undriveable" on low fuel in Silverstone, while Hamilton was content with the car qualifying Button by 1 second. These are two drivers with similar setups, but Button is not good at controlling over-steer.

Is this an opinion? A struggling car and driver is probably the easiest thing to spot in data especially when you have another car to compare it to. It's night and day obvious.. It's much more difficult to find a .1 between two drivers that are a .1 apart in total lap time. Again, when data is analyzed properly nothing slips by, especially glaring problems....with or without the driver's feelings involved.

The engineer probably looked on Button's data and say "why are you making the car do that? why are you so slow, the data is showing the steering doing this, and suspension doing that."

As a F1 driver, when they have a problem with handling, most will drive around the problem, if they can, to make the car drivable, if the time in the session is not permitting to stop and change it.
Another obvious point that data will "see" every time.
AND answer all the "why's" your pointing out. Like I said, every movement of the driver and the car are recorded and digested in analysis. Why, how, where, when, and what- are all answered for every foot of the race track. The driver cannot "hide" anything and his driving technique, the lines he's using, problems he's driving around, etc. are all naked and exposed entirely.

Now the engineer can just say to Button "Just make it like Hamilton's, do your steering like this, your brakes like that, change dampers this, wing to that etc make it like Hamilton's."
Button is the driver so he can now come in and say "No I don't like doing that." or "I cannot drive like that. I can only turn my wheel like so and I can only tolerate the setting at this level."

By Silverstone, I would hope that the engineer already knows whether Button can drive Hamilton's setup or not and what needs to change between the two to make that possible. Somethings don't need to be said, they are mutually agreed upon by way of experience between the two of them. Now if Whitmarsh OR RD demanded he drive a certain setup, a far greater problem is created for the driver.....or for one more example, Button may chose this himself, knowing full well what to expect and what he needs to change in his technique to do so.

So even though the engineer knows what the problem is, he cannot solve it without feedback from the driver because the driver is his only link to get the performance from the car. Even if the engineer sets the car up perfect, it won't do any good if the driver can't drive it.

Then it isn't a perfect car is it?

Using the same example, the engineers could have given Button Hamilton's setup which was 1 second faster, without consulting Button but more than likely Button would have ended up going slower.

Your right, data cannot tell nor has there been a sensor invented that can tell what a driver feels or what he wants for lunch.
Data is way, way beyond your comments and a complete dissection of the driver, his session speed and times, and same goes for the cars performance.
Sorry but nothing you have said, escapes the microscope of data, except that is, the feelings of the driver, which aren't measurable, though everything else about his driving is....
With data at the current level it is, about the only feedback is... "I don't like it or yes, it's great don't touch the car..." A far cry from life as a driver in the early 90's....
I might also point out that both Button and Hamilton have pretty much grown up in racing with data in all the cars and karts they competed in, Unlike Schumi and Barrichello, who didn't have the luxury until they got to F1. It's impressive to see Hamilton when the car is garaged, with his nose in the data, truly impressive indeed. Hamilton may not have the recall experience of Schumi or Alonso, but he doesn't need it, it's right there in his laptop for him to digest. Takes away any advantage that recall skills may give another driver and actually improves his feedback way beyond someone just using the memory of their brain cells.
I guarantee you that not one driver in F1 today, is strictly relying on his feedback or recall abilities, not Schumi, not Alonso without first consulting the data that they themselves just created in the last sessions. It would be then and only then (after debriefing with the engineer) that a change is made in any direction on the car or changes to their driving either.
There is feedback and recall occuring all the time, but not without first consulting the data, to make sure it's the right feedback...IMHO

[marcush.]

don´t get me wrong...a driver who will need to see data to know where he could improve is possibly not to be found in F1.
You can find out where the difference in laptime stem from but that does not mean it is helpul to look at speed and steering angles trying to be as quick as your teammate...I´m sure these things only are possible at low level of driver skill .Some sit in the car and just know where the time is to be found ,others just have no idea and will soon find another pasttime..

Of course you need data aquisition with underfloor aero dominated cars as the main aim is to tailor the cars behaviour to the aeromap (assuming you have a reliable set of data for that) to get the most out of the package and of course the driver will not be able to integrate car attitude changes with downforce levels and cop movements..on the fly..

[speedsense]

don´t get me wrong...a driver who will need to see data to know where he could improve is possibly not to be found in F1.
You can find out where the difference in laptime stem from but that does not mean it is helpul to look at speed and steering angles trying to be as quick as your teammate...I´m sure these things only are possible at low level of driver skill .Some sit in the car and just know where the time is to be found ,others just have no idea and will soon find another pasttime..

Yet, a F1 car has 150-180 sensors operating at any given minute the cars runs...they are not all looking at aero maps. No matter the skill level, there isn't a driver alive and active in F1 that has the recall or feedback that can match a DA system. Not Schumi, Alonso or anyone else you care to name. As an example a driver who is 1 or 2 or 3 tenths of a second slower than his teammate on a 3 or 4 mile racetrack, is not and I repeat not going to find the time losses or the reasons why, by simply thinking it out, observing it himself, following his teammate, or even discussing the differences. In about 30 seconds he would find the difference on a DA system....the driver movements are broken down to 1/20th to 1/50th of a second (depending on sampling rate selected), well below the reaction time of those with the quickest reactions. And every shift,braking modulation,time gaps to throttle, time spent at full throttle,quickness of steering, quickness of correction, shift rpm,lifts that aren't felt/noticed by the driver etc. etc.. every single driver response and every single driver imposed input, charted and recorded. It ain't just about some speed graphs and some steering angle graphs.

Of course you need data aquisition with underfloor aero dominated cars as the main aim is to tailor the cars behaviour to the aeromap (assuming you have a reliable set of data for that) to get the most out of the package and of course the driver will not be able to integrate car attitude changes with downforce levels and cop movements..on the fly..

[/quote][/quote]
Find me a series that allows DA, aero dependent or not... and show me a car or driver that is winning on a regular basis that chooses not to use DA. You can't and you won't, because DA is that powerful. DA extends it usefulness not only to the driver, but the chassis, tires, engine, aero package (if there is one), but all aspects of motor sports. Bike or car or plane or boat... the dependence on DA, is there at all levels.....

AA IMHO

[marcush.]

I will not debate if DA is useful or not .It is necessary .some effects will plainly go unnoticed or a matter of speculation with out having access to a wel thought out and executed measuring and logging and analysis of data...that was not the point.

My point is :How will you be abble to extract the RELEVANT data from your truckload of gigabites...you might look and analyse for days months or forever but just by looking up data or even putting maths functions on top of it it still is only a snapshot of reality
and has big potential for missing the broader picture ,eg getting lost in the jungle.

The point is nly the driver can tell you what he needs to go quicker when you have reached a certain level.By this i don´t think of him demanding more wing at the front but something like:
If the car would be able to turn in a bit sharper I could open the exit a bit more and flor the throttle 15m earlier..this is something you will be unlikely to extract from a DA analysis as you will not be able to even imagine that there might be something to find in that corner.
Or the other way round :how will you know if your driver is on the correct path (no good teammate available) looking up your data stream? If he´s on the wrong path you can analyse for ages and the time just is not to be found...so having a spotter or more than one out on the track is also something to consider.

[tarzoon]

Agree with marcush.

Telemetry is the great equaliser here! Not all drivers are as gifted as, say, Fangio. And even those rely on data to see what is wrong. Except Kimi. He relies on big balls. And intuition. And his engineers are constantly lost. And a splash of that good stuff that they produce in the Highlands.

The rest of them (IMHO) have to work hard, see a lot of data, understand what was wrong and getting it sorted. This is a learning process. It's like any other job that requires specific skills and a touch of intuition.

Besides, maths can be as accurate as a prehistoric scalpel. From my university years I learned that the data will always give you want you want to see. It's only a matter of squeezing it in the right way!

[marcush.]

Agree with marcush.

Telemetry is the great equaliser here! Not all drivers are as gifted as, say, Fangio. And even those rely on data to see what is wrong. Except Kimi. He relies on big balls. And intuition. And his engineers are constantly lost. And a splash of that good stuff that they produce in the Highlands.

The rest of them (IMHO) have to work hard, see a lot of data, understand what was wrong and getting it sorted. This is a learning process. It's like any other job that requires specific skills and a touch of intuition.

Besides, maths can be as accurate as a prehistoric scalpel. From my university years I learned that the data will always give you want you want to see. It's only a matter of squeezing it in the right way!

oh yes.I have had lengthy debates with collegues about how to look at data ...and the conclusion coming out of it was entirely different with the two methods we were considering.Luckily only my approach showed a correlation with reality and explained the failure mode.
Nevertheless i am aware that i could as well be wrong and the final truth wa something different alltogether...
so data and data analysis alone is absolutely nothing without the ability to put things in context.and that is were the driver and the engineer asking the questions (!) is coming into the game and making a powerful tool out of it.
The Driver engineer will go a long way without it.. just look at racing till the 80s ..in F1 1976 the time spread 1st to 22nd was as little as 2 seconds in Qualy..
are you sure they all were equally dumb not able to come near the ideal setup arriving at time spreads close to what we see today? very unlikely .the spreads varied more from race to race so we can conclude that some teams had easier to set up cars or had better handle on adapting their machinery to all conditons ...but
to say DA is all and everything else is worthless is far from reality.