Drivin’

Back when I was autocrossing and racing all the time, it became apparent that I had lost a significant amount of driving skill after not driving for about 3 months. After a gap of any significant length, it would take several events before I was back up to my usual performance level. During the 4 year gap while I worked for Jim Hall’s Indycar team, I completely lost it all. My first several events after that were humbling. It took me about 1.5 years of getting my character built to get back up “on the step” in terms of competition driving ability. Those experiences convinced me that driving a race car is not at all like riding a bike: race driving skills fade quickly with disuse. Considering the fact that most professional race car drivers started in karts at age 5 and have been racing continuously since then, I think that no one starts out as a natural driver. We all have to work at it really hard, and we get worse rapidly when we don’t.

During the last couple of years, I have inadvertently learned a lot about re-learning how to drive after long stretches away from competitive driving. I had a 16 year gap, about 6 months of supremely embarrassing races, and a one year gap. The latest gap was caused by a new job that was 150 miles from home. Since I was away from my family during the week, I reserved the weekends for family-only stuff. Now that we are all moved in to a new house and the kids are in their new schools, I have some time to go play cars again.

I have spent a fair amount of time looking for good advice on how to speed up the process of re-learning competition driving, but came up with very little good stuff. In the process, I may have accidentally destroyed the fledgling web site raceruniversity.com by asking the question and getting no responses from the instructors there. A simple question like that should not have stressed the system to the point of collapse, but it got really quiet there after that. I’m surprised that there is so little good advice available on this topic, particularly since I can’t possibly be the only intermittent driver out there.

I finally realized that I should ask Paul Costas for advice on re-learning how to drive. He’s the Witchdoctor of Strategery for ThinkFast Lab, he’s a proven hot shoe, and he has a lot of experience as an HPDE instructor in Texas. And boy, did he ever come through with the goods. Here are some excerpts from our recent email exchange:

Paul: Advice? Tons.

This is all mental. Step into the car. Perform.

Neil: Dood! Love it. Very zen. This is getting printed out and living in my helmet bag, in the garage, and in the front of my setup note binder. If it would fit on the steering wheel, I would put it there too. So, while you are feeling philosophical about driver coaching, I have a question for you about over-driving vs. under-driving. We usually spend the vast majority of a lap providing smooth control inputs, and a small percentage of it making very rapid corrections to keep the car under control. What’s your opinion of the percentage of a lap that should be spent catching the car vs. making smooth control inputs? The Prost/Mears approach looks like 0% catching, but the Schumacher/Vettel approach definitely doesn’t.

Paul: The slower the speed, the faster the hands. The slower the speed the more the stopwatch can be influenced by aggressive moves and by pitching the car. On an initial run on cold tires overdriving the first bit will put heat into the edges of the tires quickly and get them gripping better, quicker. Once the tires are warm or on a second/third lap or a lap after a bit of cooldown, you may only need a slightly aggressive first corner or two and then work back to smoothness and ‘he who turns the steering wheel the least, wins’.

The inverse is true. The higher the speed the less tolerant the stopwatch will be of fast hands and the likelihood of the watch being kind to a pitch/catch maneuver is abysmal. Only in a last ditch keep-it-out-of-the-weeds move where the lap is likely screwed already. So I agree that the majority of a lap should be smooth inputs, and fast smooth inputs at slow speeds and slowish-smooth inputs at high speed.

For a % of lap guesstimate, it’ll come down to the lap or autocross course. The more that is low speed the more you’ve got to command the car. The more of the lap/course

that is high speed the more you’ll be herding the car and shepherding the car to go exactly where you want.

And insert tradeoffs and the ratios change again. If your car is really happy with low speed directional change you can of course be smooth. If your car is really happy at high speed but kinda fights you (worn fronts maybe?, bandaid setup?) on low speed then you’ve got to be ready to pitch it a bit in the low speed sections.

First lap of a race you need to be overdriving the car by a few percent regardless (assuming you’ve got some space and no cars up against you), at the very least for the first half to 3/4 a lap….must get heat in tires. Then you can back off a bit, be smooth and precise and make the tires last the distance. Mears was the smoothest person I’ve ever seen and I know he preferred setups that reflected that. Having seen Vettel in person now and watching F1 prac sessions 30′ from the track, they really REALLY rely on aero as they twist the nose WAY further/quicker than you think is possible and then the aero makes it stick and damn it if they dont make the corner and fly on through. Completely different animal than Scratcher or even Freddie as we dont have that downforce-parachute to save us.

The older I get the more I realize that the skippy school I took ages ago stressing how you get OFF the brakes is more important than how you get ON them is so true. I dont really care how I get the car slowed down to the right speed that allows the car attitude to be ready for turn in and power at the earliest moment and I’m still just caressing the brake pedal as I go for the gas.

In autocross driving Miss B4C (92 Camaro) I’ve had tons of folks come up to me and tell me that my brake lights are on and I’m on the gas at the same time. I’m constantly feeling for which tire can provide a bit more grip and using that. Also, transitioning from one to the other to ease the weight movement fore/aft as I turn the wheel to ensure the quickest sector time….I have two feet and I’ll use them at the same time. Ditto with road racing.

So any percentage number I give you will be weighted with the course type, the car setup and the condition of the tires….I’m sure the number would move quite a bit based on those factors. Driving style is obviously critical. We always hear of teammates running almost identical times with very different setups to suit their style.

Find your style. Find what you LIKE to have for feedback and what you are good at controlling and driving and set that up. Be aggressive out of the box and then bring it down a notch to precision when the tires heat up and you know that overdriving results in slop.

Neil: With these emails, you have just started writing an absolutely brilliant book. Or at least a massively huge and popular series of blog posts. I have spent a fair amount of time trying to find good advice on how to re-learn how to drive after a long gap. The advice you gave me today is a billion times more useful than anything else I have found. You have a gift. Use it.

I really hope that Paul takes up the challenge of spreading his driving advice, and his style of delivery, to a much wider audience than he has considered before. We can all benefit from it. I’m supremely thankful to be the first to benefit from the specific bits of Paul’s excellentness that you just read. And boy, did I ever benefit from it!

This nifty little image is a graph of the 12 run times that I got during a recent autocross test day driving Freddie in the California Speedway parking lot. Considering how bad my driving was a year ago, the clear progression towards an optimum that is apparent in these run times is proof of a dramatic improvement in my performance. Taking Paul’s advice to heart, I started slow and worked my way faster by a small amount each run. My natural tendency is to wildly over-drive, so it takes real effort to restrain myself a the wheel. Run #7 was slightly slower because the first 6 were in the morning session and the second 6 were in the afternoon, and I didn’t have the course fully memorized after that long. I drove just a bit more aggressively during each run, very much on purpose, including the last 2 which were apparently more aggressive than the optimum. I didn’t get close to spinning all day, and my cone count was much lower than usual. I think I only hit 2 in those 12 runs. I managed to exceed my modest expectations by a lot!

Here is the GPS-generated course map. The crosshair is at the start line and the path is color coded with speed. Blue is slow and red is fast. For reference, my top speed on that run was only 57 mph, so there wasn’t much of a straight anywhere. There were 3 places where I was up against the rev limiter in 2nd gear, but they were very brief. In cases like that, it’s usually faster not to shift up, then right back down, so I didn’t. With a Honda engine, I don’t have to worry about running it against the limiter.

Because this event wasn’t about going fast, I left the old dead Hoosier R35A road racing tires on Freddie that I ran at its last event. The grip was expectedly modest, but the car handled better than it ever has before. That’s partially because I changed the setup drastically from anything that I have run before. The setups that I had run in the past were all severely out to lunch, so I went with known good spring rates and took a huge leap of faith on damper adjustments. Based only on my recollection of how the car rode on rough pavement before, I adjusted the bump damping nearly full soft and the rebound damping nearly full stiff on both ends of the car. I also used the same tire pressures that I autocrossed with on my DB-1 almost 20 years ago.

That setup turned out to be very nice indeed! I wouldn’t call it balanced, but it understeered where I expected it to (in the slow stuff) and the rear stepped out enough times that the ratio of push to loose was near 50/50. It never really settled into a balanced drift, but that could be due to the nature of the course, my deliberate under-driving, or a still-non-optimum setup. For the first time ever, I was able to use the cockpit-adjustable front and rear anti-roll bars to try both different front/rear balance adjustments and different overall roll stiffness levels, and the handling responded just like I thought it should. Also for the first time ever, the balance was close enough to right that I was able to dial in the brake bias for autocross. It ended up 4.5 turns to the rear from where it started the day. That’s a really big adjustment. My DB-1 taught me that the optimum brake bias for autocross is much more rear-heavy than road racing, but on that car the difference was 2.5 turns. I still haven’t figured out why the brake bias is different for autocross and track events, but it most definitely is. If you can clue me in on the physics behind this, please do.

The nice behavior of the car and my semi-decent driving resulted in a somewhat reasonable looking ( to me ) picture:

That’s the friction circle plot for my best run. It’s an X-Y plot with lateral acceleration on the horizontal axis and longitudinal acceleration on the vertical axis. Forward acceleration is up, braking is down, and left and right lateral accelerations are left and right. The data logging rate is 10 Hz, so the data points are 0.1 second apart. The distance between data points indicates the rate of change of the combined acceleration vector of the car. The outer boundary of the data points is the maximum capability of the car in combined lateral and longitudinal acceleration.

A perfect driver would use all of the car’s grip capacity nearly all the time, so nearly all of the data points would be on the outer boundary. Professional racing drivers on road courses can get really close to that. The two things that I like about this plot are that roughly 3/4 of my data points are near the combined grip limit of the car, and that the shape of the grip boundary is close to round. That means that I was able to trail brake effectively and transition from cornering to accelerating smoothly, using all of the combined-vector grip that was available. In older posts such as this one, I have shown examples of my friction cirle plots that were distinctly heart shaped, indicating an inability to trail brake. Because about 1/4 of the data points were well inside the boundary, I was under-driving the car about 1/4 of the time. That leaves me with a lot of driving technique improvement to work on, and I will work on it as often as I can. I knew when I wasn’t maxing out the car, and why. It was a matter of not having the course and the line fully committed to memory. That is a skill that takes years to develop, but I’m confident that more seat time will slowly improve that particular weakness.

The grip boundary of the car is considerably less than I have seen with new tires. That’s not a surprise because the tires are old, and they are meant for road racing, so they never got warmed up to their design operating range during the brief autocross runs. The only track session that I have run on new tires showed max lateral at 1.88 g, compared to 1.40 g this time. That’s 34% less grip. New tires are awesome, but they are only awesome once. When I start buying new tires again, I will try to heat cycle each set correctly. I have never managed to do that before.

The IR tire temp sensors were on the car for this event, and they showed that I have more work to do in optimizing the cambers and pressures. The data clearly shows what I need to adjust, and which way, but not how much. So, more experimentation will allow me to home in on the optimums.

As you can tell, I was a happy guy all day long. Paul’s brilliant advice, and my ability to implement it somewhat skillfully, made my day.