Getting fast fast in motorsports using data

In this post, I talk about how to get fast fast using data. Anybody can get reasonably fast with a little time and effort. For most drivers, this a year or two. However, as with any craft, the last 10% is what separates the really good drivers from the so-so drivers. Using data will help you break into the remaining 10% much faster than relaying on seat time alone.

Key Performance Indicators (KPIs)

There are lots of books available on the topic of driving faster. If you’re an experienced driver you probably have your favorites and a lot of what you’ve learned should be instinct by now. If you’re just getting started, I highly recommend you first read some books or take a class, because this post assumes an experienced driver. My favorite book is Drive to Win by Caroll Smith.

When working on going faster I keep several KPIs at the forefront of my mind while driving and as part of measuring performance using data. They are:

  • How soon am I getting on power relative to apex? Sooner is better, and if after apex then that’s really bad.
  • After getting on power, am I staying in it or correcting? If correcting (with throttle or steering), then probably because I got on power too soon or my line was bad, or both.
  • What is my top speed just prior to braking for the next turn? Faster is better, and it’s reflective of how well I exited the previous turn.
  • How is my braking affecting corner entry speed? You want to maximize corner entry speed and carry it throughout the turn.
  • How are my steering inputs into and throughout the turn? Ideally one input and then hold, with little or no corrections.
  • Am I taking full advantage of the track surface? Even a couple inches off line can cost precious tenths — something about which a data acquisition system makes you hyper-aware.
  • Am I being smooth? Smooth is fast. Ideally, I look like I’m bored out of my mind and about to fall asleep at the wheel, as observed.

It’s far from being a comprehensive list, but it’s manageable and relevant to where I am in my driving development. Your list can be different, but have a list of KPIs for evaluating yourself on and off the track.

Notice that I didn’t include lap time. This, of course, is the ultimate goal but it’s what happens between start and finish that lowers lap time so focusing on it as a means of improving performance is not the answer.

It’s what happens between start and finish that lowers lap time so focusing on it as a means of improving performance is not the answer.

Data Collection

When I first started driving in 2000, data acquisition systems were prohibitively expensive for most drivers — including me. We had crude ‘tools’ at our disposal like a visual speed or RPM check at some arbitrary reference following turn exit. While much better than a seat-of-the-pants approach, it doesn’t compare to the even the lowest cost data acquisition systems readily available in market today.

A very good low-cost option that I’ve used and recommend is the CMS Lap Timer — it’s all that’s needed for all but the most advanced drivers, this is all that’s needed. Don’t let waiting to buy and install an expensive solution keep you from gathering data. (Yes, I’ve talked to people who use this as an excuse.)

A low-cost system like the CMS Lap Timer is all that’s needed to get started collecting data today.

I currently use a Racelogic Video VBOX Lite system to measure performance. Details on why I use VBOX and how it compares to CMS Lap Timer are discussed at the end of this post.

The following video was recorded using my VBOX at the Ridge Motorsports Park in Shelton, WA. I’m sure you’ll agree the amount of information available to the viewer far surpasses what would otherwise be available using a GoPro. You may notice the speed indicator sticking in areas. It’s annoying but it doesn’t affect the log data and I’ve since resolved the issue.

VBOX From the video overlay alone, you can see that I’m able to measure and record the following data relevant to analyzing driver performance:

  • Speed (slight issue with sticking in this video)
  • RPM
  • Brake pressure (%)
  • Throttle position (%)
  • Lateral, longitudinal, and total g-force
  • Lap times

Be sure to note track conditions, tire data, and any equipment changes when logging data. Here, I’m running Pirelli Trofeo R comp tires at 29.5 PSI (hot, all 4 corners) with 8 heat cycles. Track is dry and in good condition, weather is sunny and 78F.

I should be wearing a driving suit at these speeds, so please don’t follow my example if you’re pushing your car at or near the limits on the racetrack. Wear a driving suit. I’m addressing the situation for myself.

Data Analysis

I use three different approaches to analyzing the data:

  • Qualitative
  • Corner
  • Differential

The approach used depends on time available and what I’m hoping to learn.

Qualitative Analysis

Several key factors involved in going fast cannot be easily quantified:

  • How smooth am I?
  • Am I hitting my apexes?
  • Am I taking full advantage of the track surface?
  • Am I needing to correct a lot?
  • Am I relaxed and not rushing my inputs?
  • Am I being smooth and progressive with my throttle input?
  • Am I backing off throttle in places that doesn’t seem right, and why?

These can, however, be evaluated qualitatively using recorded video. In reviewing the previous video, there’s lots of room for improvement based on these factors alone. Do the same of your own driving, be self-critical, and apply what you learn in subsequent sessions. Over time, you’ll see your driving improve and lap times go down.

Using this method I’ve become much better at self-accessing my own driving in real-time. At first, my perception of how well driving in real-time was very different compared to as viewed later on video. This is why musicians record themselves while practicing their craft. We sound much better in our own minds when practicing as compared to reality because the human brain filters out mistakes. Playing back the recorded session makes mistakes readily apparent and correspondingly raises awareness, which in-turn contributes to becoming a better player.

One of the benefits of qualitative analysis is that it can be performed quickly and easily between driving sessions, with no special tools or software required. In doing so, you’ll become faster faster because you don’t need to wait until the next track day to apply what you learned. Instead, for a typical track day has 6 sessions, you’ll get 6 opportunities to review and apply what you learned while details are fresh in your head.

Corner Analysis

The key to lower lap times is in maximizing cornering performance. While the novice driver is inclined to believe its about how fast driven through a turn, the experienced driver knows it all about corner exit speed — especially if exiting onto a long straight. As with all rules there are exceptions. For example, you might sacrifice corner exit speed in a turn if it sets you up for a better turn exiting to a long straight.

Take a quantitative approach to maximize cornering performance by evaluating the following:

  • What is my corner entry speed?
  • Am I using all available grip in turns? (For mid-turn, steady-state I look at total g-force.)
  • How soon am I getting on throttle relative to apex?
  • What is my top speed at the end of a straight? (Reflective of corner exit speed.)

Apply the scientific method as a means of improving performance by establishing consistency and then controlling one variable at a time.

By consistent, I don’t mean 80 mph, then 77 mph, then 82 mph, if evaluating — for example — corner entry speed. That’s not even remotely consistent. The variance needs to small else you have no way of knowing the input change you’re experimenting with in responsible for the change.

Considering all the human, mechanical, and environmental factors involved, I’m still amazed how consistent we drivers can get in terms of laying down lap times over 2-3 miles of pavement. Tenths of seconds are normal. While consistency is important, I don’t drive around the track telling myself to be consistent. Instead, focus is on nailing turn in, hitting and pushing my brake zones, smooth corner entry, etc. There’s a lot of feel involved too. So, again, when you think about it that way, it’s amazing how consistent we can be.

The following videos shows two side-by-side laps driven in a reasonably consistent fashion. Please turn down the volume because it’s annoyingly loud.

Once consistent, start experimenting with different inputs one at a time, and observe how performance is affected. You might be surprised by what you discover. For example, a lot of drivers think braking late and hard is the key to fast laps. In a racing situation it can make sense if needed to protect position, but corner entry speed is more important for clocking a fast lap — assuming braking is reasonably aggressive and not coasting, or worse, accelerating into turn entry as a result of braking too early. It does no good to go barreling into a corner hot if it means spending the rest of the turn collecting the car.

Besides braking, some of the other inputs you can experiment with are:

  • Turn-in point
  • Line through the turn
  • Throttle application in the turn

Regarding throttle application, I typically like maintenance throttle from turn-in all the way until time to start powering out of the turn, and then the sooner the better — ideally, prior to apex bearing in mind I drive a car with an engine where it belongs (behind the rear axle). The only exception is if I’m trying to assist rotating the car through a tight turn.

Using this method you’ll watch your personal bests become your new norms over time as part of process that continues to repeat itself. Since it’s normal for a road course to have anywhere between 10 and 16 turns, go after the turns leading onto the big straights first since this is where you’ll see the biggest return on investment. As you gain more experience with a particular track, turn your focus next to the turns leading to medium-length straights, and finally the remaining turns which have diminishing returns.

Before moving on to the next section take a look at the following track map of The Ridge Motorsports Park. Write down on a piece of paper the turns that you’d prioritize working on after you’ve acquainted yourself with the track and are ready to start finding speed. If you have experience with this track then then you should have an opinion here. I’d love to hear from  think you have a better approach — write me so we can compare notes. I’m always looking for more speed.

The Ridge Motorsports Park - Track Map

The Ridge Motorsports Park track map. What turns would you prioritize working on after familiarizing yourself with this track for the first time?

Differential Analysis

VBOX includes a software application called Circuit Tools that runs on the PC and Mac. Among other things, it enables you to analyze your driving using graph data time synchronized to video. Once in Circuit Tools, you’ll have access to much more information than is available from looking at the video alone. It also enables you to view data over the length of the track with user specified channel data for visualizing patterns and identifying complex relationships, such as the impact of cornering performance on overall lap-time and braking on corner entry speed.

The most powerful feature of Circuit Tools is being able to compare two or more laps side-by-side. Laps compared can be from the same lapping session, across multiple sessions for the day, or even from different days. You can also see how you compare to another driver if you’re OK having he or she drive your car — just be prepared to check your ego at the door because you can’t blame equipment or tires if it turns out you’re slower.

The following screenshot shows a delta-t plot comparing two laps from a 30-minute lapping session at The Ridge Motorsports Park. Lap 7 (red) is the fast lap and it serves as the basis for comparison, as shown by the horizontal red line. Lap 10 (blue) is 1.26 seconds lower and we’re about to find out why using delta-t and various other channel inputs as we work our way around the track. The red dot on the track map encoded onto video shows track position. Adding brake channel data is a good way to mentally rough in the turn locations without having to scrub through the timeline when evaluating the delta-t plot.

circuit-tools-delta-t-partial

Using delta-t, it’s easy to spot where the additional time is being accumulated in Lap 10. For the purposes of this post, I Photoshopped in the 5 segments that are readily identifiable as being responsible for most of this time — they’re highlighted in red and labeled A, B, C, D, & E. I’ve also Photoshopped in turn references and calculated the time accumulated for each of the highlighted segments since this is not an interactive demo. The keen eyed will note these segments sum to 1.30 s instead of the 1.27 s. This is not unexpected since we’re not summing across the entire path, and it supports we have a good approximation.

circuit-tools-delta-t-full

It’s easy to spot that most of the time added (60%) is in segments D and E. This isn’t surprising since T12 and the T14-16 complex lead onto medium and long straights. If trend data shows I’m not able to consistently nail these turns then I will prioritize working on them over other turns because the ROI is much higher. That said, let’s closer look at A, B, & C. They’re a bit more interesting and goal here is to simply demonstrate an approach.

Segment A

Section A is unique in that there’s no braking involved. In my opinion, T2 through T5 is the most technical part of this track. What happens in T3, T4, and T5 starts with how well entered into T2. This is not something that can be easily accessed using data, but hints of it are evident in the following screenshot.

Here you can see how I turned in sooner in Lap 7 (red), which in turn enabled me to start releasing sooner. Note the steering angle inputs and corresponding increase in speed where I’m a full 5 MPH at instance shown. The slower Lap 7 is due to bad setup (off line, wrong attitude) coming out of T3, which starts in T2, and I needed to correct as a result. In this case, I was able to salvage T5 so I didn’t get penalized in the back straight leading to T6. Conceptually, one should think of the back “straight” as starting at T4 exit.

circuit-tools-delta-t-a1

Segment B

In general, I seek to brake as late as possible without compromising corner entry speed. If I need to choose between the two, sacrifice braking. This is also where things get real busy and — in my experience — is what separates the professional drivers from doctors and lawyers. (That’s a Carrol Smith reference by the way, and I include myself in that group even though I’m not a doctor or a lawyer.)

I like to rely on LongAcc for how hard braking. BrakePos is not a good indicator because heat and tires can affect how much pedal force is required for a specified braking force. However, BrakePos is good for monitoring pedal inputs, including how early or late to brakes.

I like to rely on LongAcc for how hard braking. BrakePos is not a good indicator because heat and tires can affect how much pedal force is required for a specified braking force.

The following two screenshots show activity in the breaking zone leading up to T6 and T6 entry, respectively. Let’s start with the braking zone.

Braking in the slower lap (blue) is earlier and more abrupt as can be seen from the BrakePos and LongAcc channel plots. The early brake causes start scrubbing speed sooner and delta-t starts to grow. The abruptness unsettles the car as evident in the LongAcc plot line and combined G (table to left). The impact manifests itself trail braking into T6, where entry speed is about 2.5 MPH slower.

circuit-tools-delta-t-b1

Braking is earlier and more abrupt in the slower lap. This starts the delta-t clock ticking and the trend continues past the compromised turn entry.

circuit-tools-delta-t-b2

Trail braking into T6 entry, Lap 7 is 2.5 MPH faster because the car is more settled during braking. From experience, 1.2g combined is in the zone for what my car can do with R-comp tires in steady state cornering so any improvements in T6 are going to be in entry and exit.

Segment C
[Coming soon]

circuit-tools-delta-t-c1

circuit-tools-delta-t-c2

circuit-tools-delta-t-c3

circuit-tools-delta-t-c4

Predictive lap timer

A predictive lap timer is a great way to see in real-time the impact a change or mistake has on your lap-time. Suppose you’re half-way around the track and your knocking on the door of a new personal best according to your timer (e.g., -0.07). Now you’re just one mistake away from blowing it. Focus!

It wasn’t until I started running a timer regularly that I truly appreciated just how much using every inch of track surface matters. Being 6″ off line didn’t seem so bad until I started seeing the results in real-time on my timer — it was a rude wake-up call. This also holds for the impact of otherwise seemingly small mistakes. Putting down fast laps requires focused execution and no mistakes. The predictive lap timer has been indispensable in training me to think this way — as opposed to try driving faster, which is a surefire way to go slower.

Equipment upgrades

I don’t change equipment very often because it interferes with my ability to access how I’m improving as a driver. It’s also expensive.

I don’t change equipment very often because it interferes with my ability to access how I’m improving as a driver.

I’ve seen and talked to a lot of people who spends thousands on go-fast parts only to discover their lap times are virtually unchanged. The disappointment is evident in their expressions. I know the look and feeling because I’ve been there. When racing, I tried out 4 different differentials at great expense, blood, and sweat. I was convinced each ratio would be the magic bullet but my lap times didn’t change one iota. Very frustrating and, again, disappointing.

I’m not going to be that annoying guy that goes on and on about how you shouldn’t upgrade. Some upgrades are necessary for even the beginning driver. Safety equipment, brakes, and a good set of pedals come to mind. If there’s something about your car that’s interfering with your driving, fix it! For example, lots of cars have soft transmission and engine mounts that cause the components they’re designed to support to shift under load. This can lead to the infamous “money shift” where you drop into 2nd instead of 4th, and pop goes the motor. Yep, I’ve been there too.

Let me just offer that if you’re not driving your car near or at its limit then making it faster isn’t going make you go faster, and it certainly won’t make you faster.

If you’re not driving your car near or at its limit then making it faster isn’t going make you go faster, and it certainly won’t make you faster.

If and when you do make upgrades, change only one component at a time so that you can correctly assess the cause-effect relationship of the change on lap time. This hold true for setting changes well, such a tire pressures, camber, spring rates, sways, and damping.

Always have a plan for the day

I like to have fun at the track, but you’ll never hear me say “I just want to have fun.” I always set aside sessions to have fun with drivers I trust on the track. Taking cool video footage, hanging it out in a particular turn just for the fun of it, doing some lead-follow, etc.

I also always dedicate at least 3 or 4 sessions to working on getting faster. In a nutshell, I recommend actually doiong what we all get taught in driving school, which is to have a plan for the day on what to improve. It can’t be get new personal best lap time — that’s not actionable. Instead, focus on 1 or 2 turns (or some other aspect of your driving) based on what your data it telling you. If you’ve analyzed it correctly then your lap times will drop as a consequence of your improvements in these areas.

Track day strategy

It’s pretty funny actually. All track days start and end pretty much the same way. In the morning, when it’s often cold and dewy, everybody’s in a hurry to get out on the track and go fast. There’s almost always an incident and traffic is so heavy, people are complaining about how others aren’t pointing-by. If you’re a regular, you know what I’m talking about.

I generally don’t waste my time going out for the first session. If I do, it’s just for a very mild warm-up or to do an equipment check — and even then, I’ll go out mid-session. The first couple laps are usually under yellow so the entire group only goes as fast as the slowest car, which can be agonizingly slow.

On the flip side, it’s very common to find the track virtually empty at the end of the day. Refer back to the videos above where I appear to have the track to myself. All those people who couldn’t wait to get on track in the morning only to drive in a traffic jam are nowhere to be seen.

It’s very common to find the track virtually empty at the end of the day, and this is the ideal time to work on improving driving skills.

There are lots of reasons why people may need to leave early — I’m not faulting or ridiculing them. Family commitments, equipment problems, tired, wreck, etc, are all very good reasons. However, to the extent possible, use this to your advantage as you plan out your day.

I always use the last couple of sessions of the day to focus on improving my driving and finding speed. I’m warmed up, have previous sessions to reflect upon, and — because traffic is light — I stand a chance at clicking off a new personal best.

Back-to-back Days

I also much prefer back-to-back driving events at the same track as compared to two events separated by days or weeks. It’s a challenge to fully analyze session data during the day without a support crew. In between session is spent checking tire pressures, torquing wheels, refueling (which can mean a trip to the gas station), dealing with equipment issues, and socializing. Time disappears. When driving back-to-back days I use the evening of the first day to delve deep into my data and apply what I learned the next day, while thing are still fresh in my head.

When driving back-to-back days I use the evening of the first day to delve deep into my data and apply what I learned the next day, while thing are still fresh in my head.

Unfortunately, there’s a trend by some track organizers to run back-t0-back days clockwise one the first day and counter-clockwise the next. This effectively makes for two completely different tracks, but lots of people like this format because it’s fun. I don’t attend such events.

 VBOX and CMS Lap Timer

I’ve used VBOX and CMS Lap Timer and they’re both great systems for their respective price points. In this section, I’ll briefly touch on the relative pros and cons. I also include video footage taken by each for the same time to help you see how they compare.

20131004-12168

Videos below are from my VBOX and CMS lap timer, as shown here.

VBOX Lite

As mentioned ,VBOX is a more expensive and capable system. You can forget about impressing your friends with cool in-car video footage but it gets the job done in terms of providing visual references while analyzing your driving, which — in fairness to VBOX — is the point of this system. For more information and installation instructions, see this post.

Some key points to consider:

Pros:

  • VBOX support multiple video inputs (2 for VBOX Lite, more for higher-end units).
  • System utilizes bullet cameras that can be positioned as desired.
  • SD video quality means not needing to worry about running out of storage.
  • External microphone input
  • Can be integrated and wired directly into the car so you never miss footage.

Cons:

  • Moderately expensive
  • SD video quality
  • Logging data is rendered directly onto video output.

The last bullet requires a bit of explanation. By rendering log data directly onto video output you cannot get the original video footage back. Practically speaking this isn’t an issue since the video quality is so poor you’re probably not going to want it for any thing other than driver analysis.

CMS Lap Timer

The basic CMS Lap Timer requires only an iPhone or Android device, a windshield mount ($10-$20), and the app (free version and pro version from $20). A high resolution GPS receiver is also recommended for better accuracy and costs around $100. Assuming you’re already an iPhone or Android user, this means you can analyzing your driving with less than $200 invested. A bargain! If you don’t have an iPhone or Android phone, an iPod Touch also works (no phone contract required). Still a very inexpensive solution. Details can be found here.

With the basic setup, you get the essential information needed to analyze your driving:

  • Speed
  • G-forces
  • Video with track position indicator for reference
  • Analysis software
  • Ability to share and compare session data

Depending on your vehicle, you may also be able to purchase an optional ODBII interface that can capture and log additional parameters such as RPM, throttle position, steering input angle, and more. I have just the basic setup.

Pros:

  • Very inexpensive (if you have an iPhone or Android phone)
  • HD video quality
  • Video is stored separately from logged data so original captured video is retained.
  • SD video quality means not needing to worry about running out of storage.

Cons:

  • No ability to customize how data is displayed
  • Depending on device and usage, amount of available storage could be a concern
  • Limited to single video input (the phone camera)