Photo Credit: Zach Hunter Owens (Vicious Corvus)
Any datalogger and analysis tool worth having will give you these two useful but sometimes misleading datasets:
- Theoretical Best
- +/- or ‘delta’ trace (a way to compare where you gain/lose time across two or more laps)
I remember my first forays into using data – my first logger being my iPhone running Harry’s Laptimer (still my go-to recommendation for anyone wanting to get started in data analysis). Theoretical Best told me I could go faster – a LOT faster! It was eye-opening. As I drove more consistently and improved my pace, the gap between my actual lap times and my Theoretical Best shrank (but still tended to be 1-2 SECONDS).
I’d see major gains in micro-sectors of the track. Typically, brake zones. It made sense that I needed to optimize and consistently execute the latest-possible braking and perfect car rotation in a corner… Or did it?
The next micro-section would show a net-loss of as much or even more time than I’d gained in the braking zone. What gives??
At that point, I gave ‘Theoretical Best’ much less weight than I did before. It was ‘lying’ to me, if I didn’t take into account the fact that it was a simple algorithm, taking the smallest time values of any sectors I gave it, added them all up, and spat out a very optimistic time to shoot for.
Have you hit the same point? Is Theoretical Best useful, or useless?
Here are some things to think about that can make Theoretical Best and Time +/- traces useful again if you’ve lost faith.
A Great Article on Your Data Driven
Samir Abid’s Your Data Driven newsletter is again a wonderful muse for me and I would like to expand a bit on his article and video — I’ll embed the video in just a bit.
In this video, Samir gives us a live review of a few sim laps with live data-traces of Speed and Lap Time +/- report. It’s very interesting and may help you put some pieces together on what makes a lap slower or faster — it has a lot to do with understanding what you did as a driver and how that impacted your pace.
Using data, Samir is able to show us how the Sector Time +/- that pops up is only a small part of the story. You might end a sector with a report that you shaved 0.1 second off your best lap time but… it can’t tell you that you were actually 0.5 seconds ahead before you made a bobble in just one corner. Furthermore, Samir goes into some of the things he did as a driver that created those gains and losses — if you pause the video and look at the two graphs he shows, you can start to see great examples of how the data tells you exactly where you did better or screwed up.
Some of the keys I was reminded of as I watched the video:
- Sometimes you need to ‘give up’ speed in a small bit of the track in order to be faster in a more important bit of the track (and be faster overall). As Samir gets into laps 2–3, you’ll see a few examples of this.
- Alternatively, you can destroy a good lap by overcooking a single corner — sometimes as a result of going faster in a prior corner, especially when you get a great run onto a longish straightaway
- Look for the ‘false gains’ and ‘false losses’ in the Lap Time +/- graph.
What are these ‘false gains’ and ‘false losses’ in the Lap Time +/- report?
‘False gains’ are where the graph dips sharply below the midline (meaning you are suddenly a lot quicker than the comparison lap) only to see a sharp incline (meaning you’re “giving it all back”). This is usually due to totally missing an apex or over-shooting a braking zone. You’re “faster” because over a few dozen feet of the track map, you’re several MPH faster than ever before (good, right?) but… you are actually going into a corner too hot, have to gather the car back up, and are slower out of the corner and all the way down the next straight at a bare minimum.
‘False losses’ are where you’ll see a rise in above the midline as you enter a corner a little more slowly than before or brake a little earlier and softer than before — you’re going more slowly because you’re off-gas sooner, but… you bear the fruits of your labor in the exit of the corner and the whole next straightaway as you can get to gas earlier. You see this in the Lap Time +/- as the trace gradually dips below the midline and keeps going down until late in the straightway or into the next corner.
What About ‘Theoretical Best’?
I’ll focus on how AiM’s works in RaceStudio 2 and 3 and I bring this report up because it, like the Lap Time +/-, can “lie” to us if we don’t fully understand how the report works and how to interpret it.
Dividing a lap into micro-sectors (meaning, straightaways, corners, and even parts of corners), the software logs your lap time in each of those sections. Then it mashes together the fastest sectors from any lap in a session (or across sessions if you wish) and gives you a ‘Theoretical Best’.
After reading this article and watching Samir’s video, you can already tell how this report can be misleading. It will take those braking zones you overcook and plug them next to a lap where you were able to take the corner properly — as if you were able to ‘hero’ a braking zone and absolutely nail the apex.
Is is possible to do for one lap? Maybe. It depends. On what? That’s the $1M question now, isn’t it?
If you can actually brake super-late, rotate the car, get on the gas, hit the apex, and you’re not absolutely smoking your tires — it’s possible.
Is it possible for more than one lap? That’s a $10M question.
Provided you’re not thrashing the hell out of your tires to make all of that happen, then you can theoretically make this pace for lap after lap.
This is why the raw Theoretical Best report can be an absolute fantasy: it might be taking bits of laps that can never exist as part of a continuous, real-life lap! To their credit, AiM does point this out in their user manuals — but who reads manuals? I usually don’t, but AiM equipment is an exception because… oh geez, talk about tough to initially set up and understand!
So… what to do?
Some Things to Keep in Mind
As we analyze our data, we should always keep in mind:
What was happening in the car and on the track to make the data look like this?
If we can understand that, then we can begin to identify if we’ve found a mental model that is worth keeping and repeating on-track. A good mental model will be something we can call upon over and over again to produce consistent results, be “nice” to the car (I’m discounting the “flier” qualifying lap on-purpose), and fit certain conditions (weather, tires, etc).
When attempting to identify the ‘fast line’, do what Samir did: compare your data alongside your video. Live, if possible. This should help you identify if you were having a ‘helmet fire’ or if that fast bit of lap was at least semi-intentional.
Finally, if you’re looking for a more useful way to use the Theoretical Best report (at least, the AiM version), then I recommend you look at the ‘Best Rolling Time’: this greatly reduces the likelihood of a fantasy lap because it takes the best lap from any 2 laps as so:
If you’re keen to fiddle with things a bit more, you can even re-make your sector splits so that there are fewer of them. One (very fast) racer friend suggested setting up sectors that began in corner braking zones and included the entire corner and following straightaway. This method helps “smooth” that data when looking at the split time report and allows you to more easily see “good” sectors at-a-glance.
Until next time — have fun, take chances, and learn as you go.