Tire Testing

If I was going to pick one tire to use for everything, I’d use the Hankook RS4. It’s not the fastest, but it’s fast enough, decent in the rain, has great feedback, lasts forever, and is available in three 15” Miata sizes. The 195 is really a 205 width, and a good choice for Miatas on stock power and skinny wheels. The 225 is an ideal choice for most upgraded Miatas on 8-9″ wheels. The 245 is good for tight courses and for more powerful cars on 9-10″ wheels. This is my favorite tire and I see no reason to do HPDEs or endurance race on anything else.

And yet… I keep buying different tires! Some of this is wanderlust, to just see what else is out there. Some of this is bargain hunting: I got Yokohama S.Drives on closeout and was out the door for $200 mounted and balanced; I bought 225 Maxxis RC1s on closeout for $108 each, with free shipping; I’ve bought Douglas all-season tires for $36 for sliding around and training. And then I’ll stumble across cheap take-off slicks that are priced so well that I don’t have a choice.

As a result, I have 36 tires in my garage right now, and 28 of them are mounted. After upgrading my NA6 to NB Sport brakes, I gave up on 14″ wheels and so all seven sets are 15″ wheels. One set of wheels is 7″ wide and has street tires on it, all the others are 8-9″ wide, with track rubber. Shockingly, only two sets of wheels have RS4s, the rest are rotated regularly for tire testing.

Tire testing is a hobby, but also something of a responsibility. I wrote the rules for the Pineview Challenge Cup, and I rank every tire individually. Some of that ranking comes from online tire reviews, but a lot comes from my personal testing. The primary way I evaluate tires is to look at the lateral Gs in Turn 2 at Pineview Run, a long right hander.

Turn 2 for tire testing

I export the data to a CSV file, and then average the lateral Gs through 200 feet of that corner. This removes the peaks and valleys from the 10hz GPS data, and also idiosyncrasies of driving style and line, and gives me a solid number I can use to compare to other tires.

Tires that I have personally tested and have the lateral G data for are the following: Achilles ATR Sport 2, BFG Rival 1.5 S, Bridgestone RE71R, Champiro SX2, Continental ECS, Douglas all-season, Dunlop DZ102, Falken 615K+ and RT660, Hankook RS4 and Z214, Hoosier A7, R7, and SM7, Maxxis RC1 and VR1, Nexen N’Fera Sur4G, Nitto NT01, Pirelli PZero PZ4, Toyo RR, and Yokohama S.Drive. I’m sure I’m forgetting some, but you get the idea. I often put one of my Aim Solos in other people’s cars, and the list of tires that I have lateral G data for, but didn’t personally drive, is about twice that many.

I don’t know of anyone else who has such an extensive database of tire grip, but then I wouldn’t expect to! It takes a lot of time, money, and effort to gather and collate this data, so this is the kind of thing one keeps to themselves. Or at least it’s not something that’s shared with other people without some means of compensation.

Tire Test: Accelera 651 Sport

I recently tested the Accelera 651 Sport, and I’m not shy about sharing this data, because it’s not a competition tire. Most of the people using this tire are drifting, and even though the treadwear carries a 200 UTQG rating, this is a mid-300 TW tire, grip wise.

I’ve driven on the 651 Sport previously, but in a Honda minivan, in a 24 Hours of Lemons race. That car is brutal on tires, and I didn’t feel it was a fair test of this tire. You can read that report over here, but the gist of it is that the 651s were fast in a straight line, but lacked cornering grip compared to the N Fera Sur4G.

The reason I hadn’t previously tested the 651s on a Miata is because Accelera only made them in a 195 width. Given that these tires run narrow to begin with, I wasn’t interested. But starting this summer Accelera started offering the two most popular Miata sizes: 205/50-15 and 225/45-15. I ordered a set of 225s as soon as I’d heard about them.

I had the tires mounted at Shade Tree Auto, my favorite car mechanics in Ithaca. Jack, the owner, races in Champcar, and his team regularly kicks my team’s ass, and so I know that he knows what he’s doing. Jack did my alignment at the same time, and when he was finished called me to report something very strange going on. The car pulled to the right on acceleration. He knew it wasn’t the alignment, and suspect diff bushings, a worn out Torsen, or something else in the drivetrain.

I picked up the car, drove it home, and noticed the same thing. On the gas, and on the brakes, the car would veer to the right. At 80 mph and on the gas, the car felt unsafe.

I really didn’t want to take my rear end apart, but I had to see what was going on. When I pulled the tires off, I set them up against the wall instead of stacking them on top of each other, as usual. I don’t know why I even looked, but I noticed they were different heights. Two of the tires were about 1/4” taller than two others, weird. I installed two of the tires with the tread facing backwards, so that I could get the same size diameters on the front and rear, and viola, the on-throttle steering returned to neutral!

I got two tires flipped on the rim so that my tread pattern is all going the same way now, and now as long as I run the larger ones on the rear (or front), the car tracks perfectly. I contacted the tire importer Tire Streets to tell them about this, and they rectified the situation with a new set of tires. They didn’t have the 225s in stock, so I got 205s instead, but in the 100 TW compound instead of 200. A+ for customer service, even if it did take several pictures and measurements before they believed me. The replacement tires are in a heated basement waiting for next year and another tire test.

I had no idea that 1/4” in diameter would have this kind of effect. I posted about this on the HPDRE group on Facebook and found that other people have had similar problems with different brands of tires, even well known name brands! FWIW, three of the tires all had the same date codes, one was different, but two of them measured undersized.

If I can pass along this one piece of knowledge – measure the diameter of your tires – and keep someone else from disassembling their car looking for answers, then my work is done here.

When mounted on 15×8 wheels, a stack of four 225 Acceleras measured exactly the same height as a stack of 225 Maxxis RC1s. The 651s I drove on the minivan were narrower by comparison, and so perhaps new Accelera sizing is more in line with other tires now?

My initial plan was to use these tires in a Lemons race, but that plan when to shit, quite literally. My first chance to test the tires was at the final Pineview Challenge Cup race. I hadn’t driven in 14 weeks, and still not feeling 100%, but I figured that getting back on track might kickstart my system.

My first few laps with the tires were tentative, and I initially thought the 651s might be similar to all-seasons. But I think it was just cobwebs, since I got better and better as the day went on. The time trial race is just three laps per session, one warm up, and then two hot laps. In my first run I barely broke a 1:20. My second run I did a low 19, and started feeling more myself. In the final run I did a 1:18.6, and that felt like a pretty good lap. I initially set the tires to 28 psi cold, but after checking the pyrometer, added 4 psi all around to bring the center of the tire up, and that helped lap times as well.

The tires felt really good, with great audible feedback and predictable breakaway. They are a little vague on turn in, but are responsive to mid-corner steering and throttle inputs. I was a little disappointed in the lap times, being 1 second off the time I set using Continental ExtremeContact Sport (ECS). But like I said, I’ve been out practice for a bit, and I’m comparing to a different day, with different conditions.

So let’s look at the data. The following image is lateral Gs comparing four laps on Conti ECS (red) vs four laps on Accelera 651 Sport (blue).

Lateral Gs – 651 (blue) vs ECS (red)
  • A – This is Turn 2, where I do most of my lateral G data gathering. If I went by peak lateral Gs, then the 651s would win, and this is why I average the Gs over 200′ of distance
  • B – This is a right/left going down the hill, and it’s not a peak G corner, but it’s a good measure of driver confidence. On average, the ECS are giving me a bit more confidence.
  • C – This is T11, also called the Knuckle, a long left hander. You can see red and blue overlap quite a bit here, and lateral Gs are just about the same.
  • D – This is the Blind Hairpin, a cambered 180 with a downhill braking zone that requires a lot of turn-in confidence. In this section, the Conti ECS has more grip.
  • F – This is the S-trap, a super tight right-left switchback that loads the tires a lot. Again, a very slight advantage to the ECS.

Now this wasn’t a back-to-back comparison; track conditions were different both days. The wheels were different as well. The 205/50-15 Contis were mounted on 15×6.5 wheels, while the 225/40-15 Acceleras were mounted on 15x8s. If the Conti was available in a 225, and it was on a 8″ wheel, I think the Conti would be a clear winner, but as it is, there’s not much to choose between them.

While I haven’t tested the Accelera 651 Sport for longevity, they are so similar to the Conti ECS I bet they are the same in that way as well. Like the ECS and many other 300 TW tires, the 651s have three center grooves, and I would guess they make a good rain tire. All in all, I think the Accelera 651 Sport is a solid alternative to the Continental ECS, Firestone Firehawk Indy 500, Michelin Pilot SuperSport, PS4S, and other 300-340 TW tires. The 651s are a step above the older 300 TW summer tires like Yokohama S.Drives, which Miata people generally liked.

Accelera tires are imported by Tire Streets, and have a unique 30-day money-back guarantee. They are priced well, and are a bargain in the larger sizes. I plan to use up this set as a dual-duty tire, and then mount the Accelera 651 Sport Xtra tires I got in exchange. These carry a 100 UTQG rating, but based on the 200 TW 651 Sport being more like a 300 TW, I’m betting the Sport Xtra 100 TW is more like an actual 200 TW. But I’ll only know that after I flog them through Turn 2.

Setting Miata Lap Records at NYST

On Sept 14th 2021, three Miatas set lap records. If you don’t know New York Safety Track (NYST), it’s two miles of undulating asphalt near Oneonta, New York. Despite the name, I wouldn’t call it the safest track; there are trees, tire walls, and not quite enough runoff before you encounter them. This weekend (9/21) they had their annual roman candle war, where people run around holding fireworks in their hands and shooting them at each other. The teams are shirts vs skins; I’m not making this up. Ahem, “safety track.”

But as tracks go, it’s a good measure of a car’s performance and a driver’s skill. A track like Watkins Glen is more about car performance than driver skill, and a track like Pineview Run is more about the driver than the car. But NYST strikes a balance, treating cars and drivers fairly equally, and I’d say it’s similar to a track like Mid Ohio in that respect. Average minimum corner speeds at NYST are in the low 60s, and there’s a long and uphill front straight, which doubles as an airstrip for small planes.

18 turns, one long straight, lots of elevation.

This HPDE event was organized by Doghouse Track Days, which is a group of Porsche instructors from the Niagara region who wanted to do something different than the typical PCA event. They started their own HPDE organization and have succeeded in creating a fun atmosphere, with lots of track time, and a great price point. I will be going to more track days with these guys in the future, and I hope they bring their party to more race tracks.

The conditions on Sept 14th were cool and wet in the morning, cloudy initially, but clearing throughout the day. An 8-9 mph tail wind down the front straight would turn into a head wind down the back straight and help dry the track out. With a maximum of 70 degrees all day, these were ideal conditions. Record setting conditions, if you will.

Lap record classes

The overall lap record at NYST is a 1:27.81 held by Mark Petronis in a C5 Z06 with aero. You can read about that here, and watch his humorous vlog here. Mark says it wasn’t a perfect lap, and the data says there’s a 1:26 in it.

Well, that’s the overall lap record, and Miatas will never be in that discussion, but there are other records for the taking. In official racing, you have different classes (Spec Miata, NASA TT5, Gridlife Club TR, etc.), but there are no “official classes” for HPDE. Generally, people say “fastest Miata” or “fastest street tire”or some other unofficial classification.

<rant>As a category, “Fastest street tire” is bullshit. The 200 TW autocross wars have ruined the entire concept of what a street tire is, and it’s ruining endurance racing as well. The fact is that A052s are faster than Toyo RR, Maxxis RC1 and many other slick or treaded tires in the 40-100 TW range. Anyone on Yoks should be put in the slicks category, which is what Gridlife does. If we’re going to standardize on a street tire, let’s standardize on Hankook RS4s. They don’t fall off drastically in a 20-minute session, last five times longer than A052s, and embody what a 200 TW street tire is all about </rant>.

So if there are no official classes for HPDE lap records, chassis and engine generations are at least a good place to start. Miata classes that make sense to me are the following:

  • ND – I’d put ND1 and ND2 in the same category. But since there’s no easy upgrade path from ND1 to ND2, maybe they belong in different categories.
  • NC – All NCs in the same class. MZR 2.2-2.5 swaps would go into a different category.
  • BP – I’d put all the 1.8s into one group. The 1994-95 made a couple less hp than the 1996-97 , but many of these are on standalone ECUs now that they are 25 years old. You can also bolt up the later heads and intake manifolds to the NA8 bottom ends, so there’s not that much difference between a modified NA8 and a NB2.
  • B6 – Compared to the 1.8s, the 1.6 has 15% less displacement, but more importantly, there’s no easy button for improving the intake manifold (square top, Skunk2) or port geometry (NB head). Selfishly, I have a 1.6 and the only way I’ll ever set a lap record is being in a class without 1.8s.
  • Forced induction – Each one of the above “classes” should have its own FI class. Or maybe by by transmission, since the strength of that’s often the limiting factor. I wouldn’t distinguish between superchargers and turbochargers, it’s all a mistake forced induction to me.
  • Engine swaps – This is a pretty broad category, and a good way to split these up might be by number of cylinders: normally aspirated swaps of 4 cylinder, 6 cylinder, and 8 cylinder.
  • Kit car – I wouldn’t put Exocet, Catfish, or other Miata-based kit cars into the same classes as any of the above. I don’t really think of these as Miatas any longer.
  • Open – Turbo V8 Exocet anyone? Anyone?

That’s a lot of Miata record classes, and after writing it all down, it’s kind of stupid. There really should be a better way to standardize and rank the performance of Miatas, or really any car, based on a universal performance index. I’m working on that, but let’s save that discussion for another time.

On this day, the lap records were set for fastest normally aspirated BP, fastest forced induction BP, and fastest Miata overall. Who done what?

Alyssa Merrill, N/A-BP – 1:37.908

Alyssa has a 1999 Miata with a few bolt ons, the engine probably makes 120 hp at the wheels. The aero package is good, with a DIY splitter, 3D-printed splitter ramps, and a 9 Lives Racing wing without a Gurney flap. None of that explains how she can go under 1:38 on Hoosier R7s. See for yourself.

As impressive as that lap is, she also did a 1:39.586 on RS4s. Going under 1:40 on an honest-to-god street tire (not a Super 200), in a street-legal NB1 still using the stock ECU… fucking unbelievable.

Chris Safranski, FI-BP – 1:36.540

Chris and I might have met once before; he was the head mechanic on a team that was pitted next to us at the Lucky Dog 24 hour race at Buttonwillow in January 2018 . At that race, our Yaris ended up beating their Civic, but it was a pretty close race, and it’s a shame we didn’t start our friendship back then.

Chris has gone through just about every iteration of forced induction, from a M45 supercharger to his current turbo setup. It makes about 250 horsepower and has proven to be reliable over many seasons of tracking and instructing. His aero setup is a little different with a carbon wing and a homemade splitter that pivots upwards when it hits things, but otherwise it’s standard trackable Miata fare.

Chris is an instructor, but did his laps in the crowded Advanced group. He would have had more clean laps and a faster lap time in the Instructor group. Nevertheless, a record FI BP lap on this fine day.

Michael Giurintano, V8 Miata – 1:35.0

Michael’s car was originally an NA with an automatic transmission. He scrapped the driveline, gutted the tub, caged it, and swapped in a LS3. You can imagine it’s not easy to corral 400-plus horses in a small chassis, and so he’s got a homemade carbon fibre splitter and 9 Lives Racing wing to help high-speed stability.

Like most of us, Mike is not a professional driver, and he’s still learning. I’d expect him to drop another second next year, but for 2021, this is an impressive lap time, and the overall fastest Miata ever at NYST!

Fastest Miata at NYST in 2021.


I got Lyme disease back in June, and I haven’t driven since, so I’m transitioning more into a role as track support and data nerd. I have a couple Aim Solos and I had them in both Alyssa’s and Mike’s cars. Chris has his own Solo, and so I was able to get data from all three cars. Chris had a different Start/Finish line set, but I was able to change this in Race Studio with Modify > Beacon Shift.

When I look at theoretical best laps in Race Studio, I throw out the first lap. I don’t know why, but it sometimes gives unrealistically low sector times. After that, I construct a map with five or more sectors, but usually less than nine. The point is to group the compromise corners together and divide the track into portions that are clearly doable in real life.

My 7-sector map. I don’t know why Race Studio numbers some of them and not others.


Alyssa’s record lap only had one sector that was her fastest; her theoretical best lap was 7/10ths faster, for a 1:37.200. I know Alyssa would beat herself up for that, but I’ve looked at a lot of other drivers, and a delta under 1 second is lapping consistently.


Chris’s Aim file included all of his laps together, and so I’m not going to show you all his laps and histograms, it’s just too large for the page. Stitched together, his best sectors showed that he could have done a 1:35.228.

That’s a difference of 1.3 seconds, but Chris didn’t get a lot of clean laps, and was often stuck in traffic. It’s actually somewhat humorous (and frustrating for him) how many slow sectors he had stuck behind people in the Advanced group.

It’s also worth noting that Chris was on Toyo RRs, and the others were on Hoosiers. The lateral-Gs show the Toyos are at a slight disadvantage, but it’s honestly less than I would have thought. Suddenly RRs are on the shopping list!


When I put the Aim in Mike’s car I asked him which side he wanted it on, and in hindsight, I shouldn’t have asked. It was out of reach once he was strapped in, and he didn’t turn it on for his record run. However, I got data from an earlier session that included a 1:35.830, and that’s still damn fast.

In the histogram you can see that he had good rhythm in the middle of the 1:35.830 lap, but botched the last two sectors. Putting it together he could have done a 1:34.574, one and a quarter seconds better.



Here’s all three drivers compared on speed/distance and time/distance. Alyssa is blue, Chris is red, Mike is green.

All three drivers.
  • A – Alyssa and Chris are both excellent on the brakes, going immediately from full throttle to threshold braking. Notice the sharp peaks and steep the slopes. This is an area Mike (green) can work on. Everyone reaches min speed at about the same spot on track at 1660′.
  • B – Chris takes a different line, and has a higher minimum speed. However, this doesn’t translate into a better lap time because he’s later to full throttle. The end result is everyone is pretty equal in this corner.
  • C – Chris’s line (red) has a hockey stick shape, indicating blending inputs, or trailbraking. If you watch his video, it looks like he’s scrubbing speed with wheel angle and yaw, rather than brakes. But whether through the hands or the feet, this is excellent work.
  • D – We should put Alyssa’s name on the esses, she owns them. I looked at her data on RS4s and she’s faster here on street tires than everyone else on slicks.
  • E – Mike backs up T12 very well, gets his braking and turning done early, and is accelerating earlier. This is how to drive a powerful car.
  • F – Despite what looks like overbraking into T13, Alyssa continues to gain time in the autocross section (not the official name, that’s what I call it). The trend in this flat section is all downward for the blue line.
  • G – Alyssa has a 6-7 mph min speed advantage over Chris and Mike in the final corner. They gobble her up on the front straight, just the same.
Driving line.

I don’t want to sound too much like Alyssa’s cheerleader, because it’s easier to drive a low-powered car to higher limits, and it’s not surprising that a turbo Miata and V8 Miata have more variance in the time deltas. But this was great driving by all three of them, and I challenge you, or anyone, to beat them. In the end, that’s what records are for: breaking. Who’s next?

Get used to this view.

Wing Efficiency Doesn’t Matter

A wing that is set up properly doesn’t add a lot of drag. My real-world testing showed that my 9 Lives Racing 60″ wing set at 4.4 degrees AOA and roof height added .03 to the coefficient of drag (Cd) of my Miata, no matter if it was an open top, OEM hardtop, or fastback.

On a Miata with an OEM hardtop, adding a wing accounts for an increase of 6.25% drag, which is slightly more than the two mirrors combined (see Where Drag Comes from on a Miata). That’s nothing compared to the benefits of a wing, and yet wing drag and wing efficiency are a constant source of conversation and consternation. There are much better things to concentrate on, like the lift/drag ratio of the entire vehicle. Let’s get into it.

The effect of changing wing angle

If you want to go faster around a track, then downforce outweighs drag reduction. Every. Fucking. Time. And yet people make purchasing decisions based on which wing has less drag than another. Or they adjust wing angle for more efficiency rather than more downforce. Going after efficiency only makes you go slower, and I’ll prove it to you with a few simulations.

For the simulations I’ll use a Miata with round numbers: 2400 lbs, 120 hp, 1.2g, Cd .45 (plus wing), Cl 0 (plus wing). This represents the average Miata with bolt-ons running on 100 TW R-comps or Super 200 tires. The Cl and Cd values will change with each configuration, as I adjust wing angle. I’ll simulate a 9 Lives Racing wing, using their published CFD data for wing angles of 0, 5, and 10 degrees. Many wings operate in a similar window, and so this is mostly a generic wing choice. I’ll run the make-believe car around Mid Ohio, which has both fast and slow sections, and is about average for a race track. (I’ll also simulate the car at 75 hp, but more on that later.)

Here’s how it shakes out:

Wing angle 0, 5, and 10 degrees at 120hp, and then reduced to 75 hp.

In the table above, the 10-degree setting wins with a 1:40.75 lap (100.75 seconds). It’s only a little faster than the 5-degree setting, but both of them beat the zero-degree wing by a fair margin.

Does the most efficient setting ever work? In a straight line, yes. And even in this simulation, the zero degree setting has the highest top speed. On an oval track zero degrees would probably work well, but on a typical race track, maximum efficiency never wins.

Take a look at the last three columns where I reduced power to 75 hp. This is like a Miata running on three cylinders. In this detuned state, the 5-degree angle wins over 10 degrees, but just barely. Meaning, if your car has 75 hp, then you can babble on about optimizing your wing angle for less drag. Everyone else STFU.

Wing maximum efficiency is baloney

Given that the most efficient wing angle was the slowest, it begs the question: does wing efficiency matter? As a static number of “maximum efficiency”, no. At a certain coefficient of lift, a tiny bit.

You can research this yourself on Airfoil Tools, or read the article I wrote Car Wings Examined. The gist of that blog post was that there are different wings for different uses. Of the 1638 airfoils in the Airfoil Tools database, there is no single wing that is the most efficient at all speeds. I used the site’s search feature to find the most efficient wing, and each of the wings below is the “most efficient wing” at a different Reynolds numbers (which you can think of as different speeds, or different size wings, or both).

Which is the most efficient airfoil? All of them.

Consumers who believe that a wing that has a 17:1 lift/drag ratio is going to be better than a wing with a 12:1 ratio (or whatever) is a victim of marketing, bad assumptions, and lack of knowledge. The airfoil that has the highest maximum efficiency won’t make the most downforce when set at a high angle. Likewise, the airfoil with the most total downforce is probably not very efficient at low angles.

To sum it up, as it relates to car racing, an airfoil’s maximum efficiency is total bullshit.

I’ll tell you what’s to blame for this: advertising. Wing manufacturers like to compare dick sizes and somehow wing efficiency became their ruler. Touting their higher “17:1 lift/drag ratio” gives them a chubby. Every wing manufacturer seems to do this, and so I understand having to keep up with the Jones’s, but it’s still utter and complete nonsense as it relates to the only thing that matters – lap times.

It would be better if wing manufacturers stopped competing with useless information and told us the maximum coefficient of lift, what angle of attack that occurs at, and the lift/drag ratio at that point. That’s the data we need to make purchasing decisions.

Aerodynamic efficiency of the vehicle

I previously mentioned that a wing added .03 drag to the vehicle, for a total of .48 Cd. That might or might not seem significant to you, so let me put this another way: the wing is responsible for 1/16th of the drag of the entire vehicle. If you read my post on Where Does Drag Come From, you’ll see that reducing drag on any other part of the car is going to return larger gains. Optimizing for wing drag is a waste of time.

In the OptimumLap output (the spreadsheet image above), notice the Aero Efficiency field, the fourth row from the bottom. Aero Efficiency is a measurement of the total lift/drag ratio of the vehicle, and this increases with wing angle. Now wait a goddamn minute; increasing wing angle makes the wing less efficient, right? Correct. You have to think of aero as a system, and a wing is just part of that system. A wing doesn’t add a lot of drag to the entire vehicle no matter what you do, but can add a lot of downforce at a much higher rate.

A Miata with the wing set at the highest-downforce least-efficient wing angle (without stalling) gives the entire vehicle the most efficient lift/drag ratio. If you care about aerodynamic efficiency, care only about this.

But don’t misunderstand this I’m saying and set your wing to 10 degrees! You have to take into account the downdraft angle of air induced by the roofline shape, which changes depending on the height of the wing, and changes across the length of the wing as well. You can read about that in my post on Visualizing Airflow, but at roof height on a Miata, the downdraft angle is an average of about 5 degrees (5 degrees in the center, 7 degrees at the wing stands, and zero at the ends). When I say the wing is most efficient at 10 degrees, that’s free stream air. On a Miata, the best you can do with the average car wing is 5 degrees. Unless you can get the wing higher, where there’s less change in the angle of the air coming over the roof, or you’re using a wing that can operate at a higher angle of attack.

The fact is that different airfoils work better at different angles of attack. Most of the wings I’ve reviewed in Car Wings Examined have a range of about zero to 10 degrees before stalling, but some work up to 15 degrees. I typically cite 9LR wings for examples because a) awesome, b) tested it, c) have not found anything better. I’m making some wings myself, but I don’t think there’s going to be a marked improvement in the shape. I’m just going after more chord and I like DIY projects.

When drag matters

I oversimplified to exaggerate a point, and there are some outside cases where wing drag matters. Certainly anywhere top speed is more important than cornering speed, such as land-speed records. An oval track is another area where wing drag reduction could be beneficial over downforce. But I haven’t seen Miatas setting land-speed records or racing oval tracks, and for normal racetracks where Miatas race, there are very few situations where wing drag matters. The only time it does is with low powered vehicles (under 75 hp) and the following corner cases.

Endurance racing

I’ve run a lot of endurance racing simulations in OptimumLap, and downforce wins over drag 99% of the time. This is true even when fuel consumption forces you to take an extra pit stop over the course of a day. However, there are some cases where you can optimize your fuel consumption and lap times using a lower-drag lower-downforce configuration, and ultimately complete more laps.

That one in a hundred times is always the result of one fewer fuel stop. If you’re racing in AER, with 90 minute stints and 3 minute pit stops, then this will never work. If you’re racing in 24 Hours of Lemons (where stint time is unregulated), and can do a two-driver change the first day, it can work. For Lucky Dog and Champcar where there are 2-hour stints and 5-minute pit windows, you have to be right on the cusp of a 1:55 stint time, and then wing drag matters. Although a single full-course yellow incident during a stint would have the same effect.

Watkins Glen

At Watkins Glen, in a Miata with less than 100 hp, drag matters. I’ll spare you the data, but I ran the simulations and the 5-degree and 10-degree wing angle came out to exactly the same lap time for a Miata with 100 hp. Both of these beat the most efficient wing setting by half a second. When I increased the power above 100 hp, the high-downforce setting won every time. Ergo, if you have less than 100 hp, then drag matters, at this track and similar tracks. At most any other track, wing drag doesn’t matter, go for maximum downforce.

Speaking of Watkins Glen, I’ve never been to an endurance race there when the track wasn’t under full-course yellow for less than 45 minutes. Someone always hits a wall (not pointing fingers, one time it was me), and this brings out the pace car and many laps under caution going 40 mph. This usually happens three or four times in a 8- to 9-hour day. In these situations you save a lot of gas, and optimizing your wing for more efficiency would just slow you down when the track goes green.

Bad wings and bad setup

I tested a cheap 53″ double wing at Watkins Glen, and it had almost as much downforce as my 60″ 9 Lives wing, but a shit ton more drag. A shit ton in technical terms is 300% more drag than the single wing, and this changed the total vehicle Cd from .48 to .55. Even worse, the rear-biased drag lifted the front of the car, and so the front lost .2 Cl downforce! In this case, wing drag absolutely matters.

It’s worth noting that OptimumLap simulations predicted the car would go faster with the double wing than without it, and so even a crappy, draggy, $70 wing is better than no wing at all.

A similar situation would occur if you set a single element wing with too much angle, and put the wing into a stall condition. This is often the result of setting the wing angle without accounting for the downwash angle of the roof.

Image from F1 Dictionary

Aero balance

Going after maximum wing downforce may produce a slower car. If reducing wing angle makes the car go faster, then either the wing was set too high (stalling) or the car understeered so much that it ruined the balance of traction between front and rear tires.

An understeering car can be fixed in a number of ways, the easiest is adding chassis rake (raising the rear of the car, or lowering the front). According to Supermiata alignment settings, Miatas are sensitive to chassis rake and this can be an easy way to adjust balance. Another way is to reduce the front roll couple by increasing rear spring rate or sway bar thickness, or decreasing those in the front.

Aerodynamic balance is really a topic unto itself, and I’ll get into that in the future. But if you don’t care about lap times, and you’re just having fun in HPDE, play with wing angle and aero balance. Less angle means less rear grip, and many people enjoy a bit of oversteer, even if the car has slower lap times and is less stable at high speed.

Finally, if you’ve set your wing up correctly, you can ignore wing drag and efficiency as variables of any consequence. The only thing that matters is the aerodynamic efficiency of the entire vehicle. The benefits of downforce far outweigh the penalties of drag, and as a whole, the vehicle will have the highest aerodynamic efficiency with the wing set at the highest, least efficient angle. The next time I hear someone talk about reducing wing drag, or how efficient their wing is, I’m going to punch them in the dick point them to this article.

Car Wings Examined

Airfoil Tools is an amazing website, and is the primary way I research wings. When I look at wings, the most important factors to me are the maximum downforce (Cl), and the efficiency (Cl/Cd). For downforce, I ignore anything under Cl 1.5, and for efficiency, I’m only interested in Cl/Cd of 100 or higher. These are nice round numbers, and easy to remember.

I look at these two values at 500k Reynolds, because that represents a normal-sized wing at a realistic car speed. For a 9″ wing, 500k Re is 73 mph. If I have a single bone to pick with Katz’s Race Car Aerodynamics, it’s that he commonly cites Reynolds numbers of 2 million, which would be a 9″ wing traveling at 280 mph!

For low speeds (or small wings) I might look at Re 200k, and for huge wings or really fast speeds, I’ll look at 1 million. But there’s really no reason to look outside of the 200k-1M Reynolds numbers and 500k is a happy medium.

I also set the turbulence value to Ncrit=5 because Airfoil Tools doesn’t allow me to set it any lower (in some cases there’s data for lower Ncrit numbers, but it’s rare.) The default setting in Airfoil Tools is Ncrit=9, and that replicates what the wing would experience in a wind tunnel, and doesn’t represent what’s happening in a race, behind other cars, with cross winds and other factors that create turbulence.

The next important factor I look at is how the wing deals with stalling at high angles of attack. I look at this for two reasons. 1) Most cars will go faster with the wing set to maximum downforce rather than maximum efficiency, and 2) car roofs are often cambered and this means air hits the wing at different angles across its length. The fact is, at some point along its length, a wing may be stalling. When some wings stall, they do so gradually, and that’s good. Other wings stall dramatically, and that’s bad – lift takes a nosedive and drag spikes way, way up.

Airfoil Comparisons

Now that you know my criteria for looking at wings, in the rest of this post I examine different wing profiles and give my thoughts on them. I’ve put these in order by how much downforce they make. At the end of this post I’ve included a table with summary values and some parting thoughts.

With all of that front matter and grey matter out of the way, let’s check out some wings!

Clark Y

The Clark Y airfoil (wikipedia) has a flat top, which can make it easier to manufacture, and for an airplane, it’s good for training because it has gentle stall characteristics. But as a car wing, the flat top means more drag and less downforce than a wing with a cambered topside. As such, the Clark Y doesn’t quite meet my downforce threshold of 1.5 Cl at 500k, and the efficiency (Cl/Cd) is below 100 as well. Personally, I wouldn’t use this wing profile.

Clark Y flat top.

Wingmen Aerodynamics makes a wing with this profile, or something very similar. I’ve seen this wing at a couple different races, and I was immediately impressed with the build quality. From the construction pictures on Facebook, it doesn’t appear that they need to make the top flat – it’s not like they are using a flat piece of foam (or whatever) to simplify construction. So for the amount of work that goes into making such a wing, I would have chosen a different profile.

Wingmen Aerodynamics, flat-top fiberglass wing.

If you have one of these wings, I’m not trying to make you feel bad; I’m sure your car goes a lot faster with it than without it! Also, the flat top has two advantages: it’s easy to set the wing angle, and it doesn’t fall off a cliff when it stalls. This means it would perform well behind a highly cambered roofline, such as Miata, where the wind angle changes over the length of the wing.

NACA 6412

MacBeath often cites the NACA profiles as examples in his books, and for good reason, they are easy to understand. The first number in 6412 means percent of camber relative to chord (6%), the second number is where the camber occurs (4 means 40% of the chord), and the third number is thickness (12%). NACA 6412 meets my criteria for a good car wing with a Cl over 1.5 and a Cl/Cd over 100.

NACA 6412 looks good.

I like NACA profiles for another reason: they allow me to change the variables and see what happens to lift, drag, and efficiency at different Reynolds numbers and angles of attack. For example, I’ve read that maximum lift on single-element wings occurs at 12% thickness, and after experimenting with different NACA profiles that are identical in other respects, I know this to be true.

You can also use the NACA 4-digit generator to create your own wing profiles. For example, this is a NACA 9512. I’ve maxed the camber allowed in the tool (9.5%), set the camber further rearward (50%) to increase lift, and used the max lift thickness of 12%. I’m certain this would be a good car wing.

NACA 9512 looks better.

Cambered Plates

Probably the easiest way to make a wing is to cut a metal pipe lengthwise into strips, and then lay two of the curved pieces on top of each other. Put a semi-circular nose on it and weld the three pieces together. This is a cheap and easy way to make blades for small wind turbines, I don’t see why you couldn’t do the same for a car wing. It’s so DIY I want to make one for 24 Hours of Lemons.

The simplest of wings, two cambered plates connected together.

You can make these wings in different thicknesses, and at 12%, it has a high Cl of 1.7. However, the efficiency is less than half of my threshold value of 100. Still, for a low-speed wing where drag is inconsequential (autocross), this would totally work. And for 24 Hours of Lemons, it’s better than a snowboard, skateboard, angled plywood, etc.

FX 72-MS-150A

I have three different made in China wings, one came as a double wing, the other two are single wings. They are fun for experimenting with, cheap, and disposable. Whoever designed them chose a similar profile for all of them, which is akin to the FX 72-MS-150A.

Made in China wings.

Some of these wings are sold as “universal” and so they are flat on the bottom because they have two mounting rails underneath. I modify these by adding wood to the bottom and rounding it.

MIC wing modified.

By the numbers, this is a good airfoil for a car, it makes a lot of lift (1.8) and is very efficient (121). The only drawback is when this wing stalls, it falls out of the sky. This isn’t a wing that you want to set for maximum downforce. Make sure you take into account the downwash angle on the roof and then set the wing to around 5-6 degrees max. Behind a Miata hardtop, this is about 1 degree negative.

GOE 464

This is a very thin wing, almost potato chip in profile. The only reason I find this airfoil interesting is because the APR GTC-300 carbon fiber wing uses a similar profile. APR’s wing has more camber, and I didn’t find an exact match on Airfoil tools, but the GOE 464 is close.

Potato chip profile.

The GOE 464 has a max Cl of 1.85, which is a lot of lift, and the efficiency at 500k almost reaches 100. It’s an interesting wing, but would be difficult for me to build, and I feel there are better choices.

GTC-300 is not unlike GOE 464

The previous airfoils were interesting in one way or another, but I personally wouldn’t put the effort into building a wing using those shapes. All of the following airfoils are superior, and would be better choices for a car wing. There’s always a trade off between lift, drag, and stall, and so each wing below has a niche where it outperforms the others.

Church Hollinger CH10

Any wing that makes around Cl 2.0 is in the category of ultra high lift. The most efficient of these is the Church Hollinger CH10. At 500k, this wing has a Cl/Cd of 132, which blows away the others.

Get thee to church. Church Hollinger, that is.

The 9 Lives Racing wing is a close cousin in shape, although the Big Wang has more camber. Adding camber adds downforce, but this usually peaks at around 10% of chord, which is where the CH10 sits as is. I’m sure that Elan figured it all out, and my guess is that those changes add downforce, at the expense of some drag. In any case, this is a great wing for a car, and would be a solid choice for a low powered car or endurance racer.

Big Wang with Gurney flap slot cut off for experimentation with chord.

GOE 652

The first thing you notice about this wing is the blunt and rounded nose. The purpose of that is to create a thicker boundary layer, which delays separation at steep angles of attack. It carries that thickness over much of the wing, and the 17% chord thickness is phat.

She thicc.

In some ways the GOE 652 airfoil is the opposite of the FX72, because the 652 has a very gradual stall. Meaning this wing is tolerant of being set at too steep of an angle. This would be a good wing on a car with a highly cambered roofline, or where you have to mount the wing closer to the trunk. In these cases, there are large changes in apparent wind angle across the wing, and this wing won’t care that much. For the same reasons, this isn’t a great candidate for a 3D wing, it just wouldn’t be necessary.

The high lift of Cl 2.0 and efficiency over 100 put this wing into elite company. I’d wager this would make a good upper element for a dual-element wing, not just the because of the shape, but because the added thickness would make it stiffer in a smaller chord.

Eppler 420

The Eppler 420 isn’t as efficient as the CH10, but has slightly more downforce and a gradual stall. It’s a good all-purpose shape, and because it’s thicker, would be a strong contender for either element in a dual-element wing. As an all-purpose wing, it’s hard to choose between the CH10 and E420. The former is more efficient, the latter makes more downforce.

Eppler 420 is a solid all-around choice.

It’s also a pretty good wing for low Reynolds (low speed or small wing). The Porsche Cayman R has a tiny rear wing, and it’s probably not a coincidence that the profile looks a lot like the Eppler 420.

Porshe Cayman R wing looks like an Eppler 420 at 5 degrees.

Wortmann FX 74-CL5-140

This airfoil wasn’t on my radar, but in writing this blog I looked at almost every single airfoil on the Airfoil Tools site. Glad I did, this one is a keeper! With a Cl that’s nearly the same as the Selig wings, and an efficiency closer to the CH10, this wing sits in rarefied air.

It’s nice, I like.

You don’t get something for nothing, and the tradeoff is a steep drop when it stalls. If you want to go after maximum lift with this airfoil, mount it high where the angle of wind doesn’t change much. I have more to say about this wing after reviewing the next wings. I’m tempted to build one, so stay tuned on that.

Selig 1223 and 1223 RTL

The Selig 1223 and the Selig 1223 RTL are Downforce Royalty. The RTL version is slightly thicker, which results in higher lift and drag. The RTL can be set to 15+ degrees and approaches a Cl of 2.5. That’s huge.

Selig 1223 (red) and 1223 RTL (green).

Both airfoils make a lot of downforce, but also a lot of drag, and their Cl/Cd efficiency is less than 100 at all angles. Ergo, I would use this airfoil for low speed or for a car with a lot of power. Those are also usecases for a dual element wing, which might be a better choice if your racing rules allow that.

AeroDesign wing from Australia appears to be Selig-ish.

Let’s compare the two Selig wings to the FX74. These graphs are from Re 1M because I plan to use these for a larger chord wing. In the comparisons you can see how the S1223 wings are clear winners in downforce, but the FX74 is far more efficient at Cl 2.25 and below. I’ve drawn a blue dashed line at 2.25 Cl, and you can see that the FX74 has a sweet spot where’s it’s making a lot of downforce without much drag.

Re 1M, Ncrit=5

Summary Data

Here are all the wings sorted by Cl. For each wing, I’ve also listed the max efficiency, and the angle where that occurs. Note that the fastest way around the track is often right around max downforce, not max efficiency. But efficiency is arguably important for endurance racing and momentum cars.

WingMax ClMax Cl/CdNotes
Clark Y1.4590.2 at α=4°Clark why?
NACA 64121.6111.3 @ 6.25°Good reference
Cambered plate1.742.5 @ 5.75°Lemony
MIC FX 721.8121.1 at α=6.75°McWing
GOE 4641.8597 at α=7.5°Potato chip
CH101.95132.4 at α=3.25°Effin efficient
GOE 6522.05102.8 at α=2.25°She thicc
Eppler 4202.1106.1 @ 6.25°Dude, 420
Wortmann FX 742.25115.5 at α=6.75°Gimme
s12232.397.4 at α=5.75°DF Queen
s1223 RTL2.3585.4 at α=5.5°DF King
Values at Re 500k, Ncrit=5

This is how five of the airfoils stack up at 500k Re. I didn’t include the RTL because that’s in the previous charts. I’ve pointed out a few areas that differentiate one wing from another.

How the players stack up at 500k Re.

My wings

Here’s a collection of wings and cutoffs in my shop. Two are missing, the GLTC 250 sq-in wing that I sent to Justin Lee, and the upper element of that wing, which is lounging in a corner somewhere.


From the bottom and then left to right: Selig 1223 RTL 41” x 16” prototype; MIC 110cm x 11.6cm chord (I have three of these); APR GTC-200 59.5” x 8” (end cap shown, the wing is on a car); MIC 135cm x 14.2cm; 9 Lives Racing Big Wang cutoff; 9LR-based wing 48” x 11.4”.

When is a 1.6 Miata faster than a KMiata?

The biggest mystery of last year is when Stefan Napp of Napp Motorsports brought his K24-swapped Miata to Pineview Run and we both went faster in my car than his. I had mentioned this in my driving other people’s cars writeup, but didn’t dig into it.

Stefan’s car is the stuff of dreams. Very wet dreams: K24Z3 motor, DIY Bilsteins, BFG Rival 1.5 S tires, Enkei RPF1s. Is there a better specification for a normally aspirated Miata? No. Well OK, a K24A2 is better than a K24Z3, but that’s nitpicking. (I originally wrote that his car was on Xidas, but stand corrected.)

Sweet dreams are made of this.

On the other hand, My 1.6 Miata is much more pedestrian. At 2/3 the displacement and no VTEC, the B6ZE is no powerhouse, but as engines go, it’s a sweetheart. I’ve written about this ad nauseam (1, 2, 3): it has all the bolt ons, 264 cams, and a standalone ECU. Most importantly, this was all tuned by the high wizard Rick Gifford.

The shocks are Tein Street Advance, which are a constant source of embarrassment. I mean, the spring rates (392/336 lbs) don’t even appear to be made for a Miata, I know, I know! And finally, the car is on Hankook RS4 tires, which are a step below Rival S and other Super 200s.

My car from later in the year when I added the splitter lip and taller spoiler.

Given the different specifications of the two cars, you’d think Stefan’s car would walk all over mine. Comparatively, my car is a dog, but every dog has its day, and on this day, we both went faster in the 1.6 than the 2.4. How is this possible? Let’s look at the data.


First is Stefan driving my NA6 (red) and his K24 (blue). I’ve marked some points of interest, but the most telling is the time/distance graph that shows he’s 1.3 seconds faster in the slower car.

A – Up to this point Stefan is driving both cars pretty equally, but he shuts off going down the hill in my car. It’s a sketchy off-camber spot, and this isn’t his car, so that’s normal.

B – Stefan makes a mistake and brakes too deep at the top of the hill and loses a lot of time on the entry to T11 (Knuckle). Notice the acceleration slope of the blue line. The K24 can really get out of the hole, except that this time the hole was too deep.

C – Stefan takes a long time to shift my NA6 from 2nd to 3rd gear, that’s the dip you see in the red line at the top. Earlier in the corner he took different lines, but they averaged out.

D – Braking too deep loses a little time. What’s interesting here is that I expected to see a much stronger acceleration slope on the blue line. This section is uphill, the NA6 is wheezing, and the K24 should be reeling it in, and isn’t.

E – Stefan loses a chunk of time by slowing too much before the final right hander. Again I’m wondering why the acceleration slope isn’t steeper out of T15, and while the K24 gets to a higher top speed, it’s not that much higher.

Some of the differences you see on the speed trace are the result of him taking different lines. This is a new track for Stefan, and it’s good to see him experimenting.

NA6 red, K24 yellow (blue doesn’t show up well)

I will say that this man can drive! Both Stefan and Dylan were instantly up to speed on an unfamiliar track, and while they would benefit from learning the layout better, they were on the limit of traction from the get-go. Could they go faster? Yes. Could they drive harder? I think not. Check it out in their video.

Theoretical best laps

The previous data was from the best lap in each session, but we saw that Stefan made a couple mistakes and was still experimenting with line. So maybe if we stitched together his best sectors, the K24 would be faster? Indeed, it looks like he could have gone .4 seconds faster, and do a 1:19.423.

Stefan’s theoretical best in the K24.

Next let’s see his theoretical best in my NA6, and he could have done a 1:17.792. Now this is flatly astonishing, because you’d expect more variability in the car with more power, and more consistent times from the anemic engine. But instead of .4 seconds, he could have dropped .8 seconds in my car. Now the delta between the cars is 1.63 seconds in favor of the 1.6.

I don’t think we’ve unravelled this mystery at all, if anything, it’s even more convoluted. Let’s see how a driver who knows the track did in both cars.


I’ll use the same colors, red is my NA6, blue is Stefan’s K24. I know the track better than Stefan (in fact I wrote the book on it, literally), and go about 1.5 seconds faster than Stefan did in his car. The freaky part is I go another .7 second faster in my car.

A – The biggest difference comes from going through the Crick, Turns 3-5. I get the car turned in earlier and this helps my min speed through the corner and down the hill.

B – I roll off going down the hill. This is not my car, on a sketchy part of the track, so I’m slower on the blue line. Recall that Stefan did the exact same thing in my car. We’re kind to each other than way.

C – Another big time delta is in the uphill esses where I can maintain a higher min speed.

D – Here’s where I finally see the K24 dominance, top speed just before braking into the Blind Hairpin. This happens again before braking into T13.

E – Notice the top speed advantage of the K24 on the front straight, this gains some time back at the very end of the lap.

Theoretical best laps

This is the same deal as before, stitching together the best sectors to see what my potential best lap time is in each car. I use a 7-sector map because this is about as granular as I can make it without getting unrealistic times that could never be duplicated in real life.

K24 best 1:17.280

My best theoretical lap in Stefan’s car would be .615 seconds faster, for a 1:17.280. That’s better, but still doesn’t match the time I did in my car. I’m more comfortable driving my own janky hardware at the limit, and I could have gone .38 seconds faster, for a 1:16.802. See below.

NA6 best 1:16.802

That’s more like the data I expected, with the slower car having less of a delta in theoretical best than the more powerful car. But once again, I did not expect to go faster in the slower car.

In the GPS trace you can see I drive a slightly different line in his car. Some of that is me being inconsistent, the other part is I drive a powerful car differently, and I’m sure you would, too.


Driving Stefan’s car back-to-back with my own, the K24 felt amazing: flat torque, instant push, and then a top-end rush. What’s weird is that I’m not seeing that at all in the data! If I look at the longitudinal Gs, the K24 maxes out at .44g, and the B6 at .41g. I’d have expected more out of the K and less out of the B.

Longitudinal Gs, NA6 red, K24 blue.

The VTEC hit comes really late, especially at this track, with this gearing. Right when you feel the VTEC kick in, you have to get on the brakes; it’s both annoying and a disappointment because you really want to ride that wave. And this may be partly why the min speeds are lower in T4, T12, and T13, because the VTEC caught both of us off balance and forced us deeper into the corner on the brakes. However, this doesn’t explain the difference in min speed in the esses.

Weight could be a factor here, especially in the fast transitions. While the cast-iron block of the B6ZE weighs more than the K24, the brackets and other associated hardware for the motor swap may have added some weight. Stefan’s car also has a soft top under the hardtop, but that should be about the same weight as my rollbar, which his car didn’t have. And my car has a tow hitch. In any case, the weights should be pretty close, and I’d guess there isn’t more than 50 lbs between them.

Both cars started out with 6” ring gears and have been upgraded with 7” Torsens. I believe both cars had the same 4:1 final drive ratio at the time. And both cars had manual racks and no ABS. Even Steven, all the way around.

What about lateral grip then? Rivals are supposed to be stickier than RS4s, but the data shows them having slightly less lateral grip. In left turns, the Rivals were about the same or better, but in right turns (which there are more of), the RS4s had more grip. RS4s FTW! (I’m punching the air as I write this.)

Which tire has more lateral grip?

This left/right imbalance could be because of sub-optimal alignment on Stefan’s car. I didn’t use a pyrometer on that day (I wasn’t officially car testing, I was just a helping hand), so I don’t know if the alignment or tire pressures were correct. And then perhaps his suspension was set up poorly, and mine was better. But seriously, I have to think that Bilsteins on their worst setting are better than Teins on their best.

Personally, I have a lot of time driving RS4s, and I can drive them at the limit easily. I had a harder time with the Rivals. Factor that into my lap times.

And for sure some of this mystery is the track itself. Pineview Run is like an autocross, but with massive elevation and camber changes. Top speed is around 70 mph in a Miata, and average corner speed is in the low 40s.

At any other track, the K24 would have had a chance to stretch its legs and run away. However, Pineview Run continues to be a riddle, wrapped in a mystery, inside an enigma. High-powered cars are hamstrung here, front-wheel drive cars are hobbled at the knees, and aero doesn’t have a leg to stand on. To that point, both cars had R-package front lips and rear spoilers. My spoiler was an inch taller, but otherwise the cars had identical aero at that time.

If I had to guess what’s going on, it has to do with drivability. There are so many low-speed corners connected to each other that throttle modulation becomes perhaps the most important factor. This is not simply for accelerating out of the corner, but for balancing the chassis for optimal grip. I saw this when driving other cars that had abrupt power deliveries, and they were always slower than expected, sometimes much slower. Which is not to say that the K24 had an abrupt delivery at all, but it didn’t have that partial throttle crispness and exactness that my NA6 has. And it’s hard to get out of shape when you have no torque!

Dylan’s NB2-swapped NA was there on the same day, and it was even better in this regard. But his car was shod with Yokohama S-Drives, and so no direct comparison can be made there. I’ve said this before and I’ll say it again: this was the best normally aspirated Miata engine I’ve ever driven. Stefan can assemble the same spec engine for you at a reasonable price, and if I wasn’t so far down the path of 1.6 absurdity, I’d have him build me one.

Well that was in the before time, and in the present day both of the Napp Motorsports cars are ruined have turbos. My car has evolved slightly with a unicorn 4.625:1 final drive and better aero. Both cars are diverging from each other, but perhaps one day we can do another test at Pineview using the same tires, and shed some more light on this subject. Or the 1.6 will win again and extinguish the lights completely.

Visualizing Airflow for 2D and 3D Car Wings

As air moves over the top of a car, it follows the contour of the roof, changing in shape and direction. Most car wings are a 2D design, meaning they have the same shape across the entire wing, and thus aren’t optimized for a curving roofline. Ergo, at various points along the wing, the effective angle of the wing changes.

The easiest solution is to get the wing as high as you can, into the cleanest air. Many racing series limit wings to roof height, and even if you’re allowed to get it higher than that, there are practical limits to how high you can brace that much weight, and the ill handling that results from it.

Ideally, a wing should follow the shape of air as it goes over the car’s roof, and so 3D wings make a lot of sense. However, 3D wings have a complex shape, and thus aren’t easy to build. And they’d have a different ideal shape for every car, and maybe even on the same car at different heights.

Finally, for any wing, you need to know where to set it for the best lift/drag ratio, or the most downforce, and for sure you want to avoid too much angle, which will make the wing stall. Many wing manufacturers publish CFD data, but how does that relate to what’s happening on your car? I wanted to find out.

Airflow visualizer

To investigate the shape of air as it comes over the roof, I’ve created what I call an airflow visualizer. It’s basically a metal rod suspended where a wing would be, with small pivoting airfoils that move into the position of least drag. With this visualizer I can measure the angle of air as it across the back of the car. This allows me to discover the ideal wing shape at any position behind any roof.

9 Lives Mini Wangs rotate into position of least drag.

To build the airfoil visualizer I used all-thread so that I could attach it to wing stands of any width, and placed 9 Lives Racing “mini wangs” at intervals using binder clips so I could slide them around anywhere. I placed the last airfoil at body width, which would be a 64″ wing on this car. Given that car roofs are symmetrical, I only put airfoils on half the wing, and mounted a camera on the other side to capture what happens at speed.

My airflow visualizer has two end plates, one is wood, the other a clear acrylic. The all-thread rod goes through the end plates and I can mount the visualizer anywhere I want by drilling a couple more 1/2″ holes. On the clear side I put 2″ grid lines for measurement, and so that I could see the angle of apparent wind. The grid lines are at a 7 degree downward angle, which was a mistake, but turned out to be a happy accident.

Now I thought this was a great idea at the time, but because of the mass of the mini wings, I had to go very fast before they would pivot, and their inertia meant they also swung when I hit bumps in the road. In the end, this brilliant idea didn’t give me a clear picture of the airflow.

I then switched to using yarn in place of the wings, but that didn’t work great either, too much amplitude. I needed to dampen the signal as it were, so I taped 2″ wide strips of cardboard around the all-thread, which was a lot better. This was such a cheap and easy solution that I’m regretting cutting up all those mini wings now.

Two runs: roof height, then 8″.

In the next section, I’ll explain what I’m seeing in the video. (If you see something different, drop me a note in the comments.)


In the following video still, I’m doing about 20 mph and all the cardboard telltales are going straight back without any up or down movement. The telltales at the end of the wing are level with the road, and the ones in the middle of the wing are level with the black lines on the acrylic plate. So this means there’s about a 7 degree downwash angle on the sides of the roof, and what looks like 5 degrees in the center.

Wing mounted at roof height.

As I picked up speed, the telltales stayed at the same angle, so it doesn’t look like downwash angle changes much with speed. However, the telltales that are next to the acrylic plate started oscillating up and down like crazy, the result of turbulence. The middle of the “wing” shows some increased turbulence at speed, but nothing so extreme.

Angle of air doesn’t change, but turbulence goes way up half way between the middle and ends.

I’m guilty of confirmation bias; I like predicting results and then finding out they are true. I predicted nothing like this, and I liked that even better! Well, let’s see what happens when I lower the wing to half the height.

Lower height

The lower mounting location replicates a wing at 8″ high. The telltales at the end are just higher than the grid lines on the clear end plate, so it looks like a few degrees negative. I expected to see them at zero degrees, but perhaps there’s some interaction with the sides of the car or the road at this height? Whatever the case, something is creating a slight downwash angle.

Visualizer mounted in the low position at about 40 mph.

In the middle of the car, the downwash angle is greater, maybe 15 degrees, and as speed increases, so does the angle. If you watch the video, it looks like that the air near the trunk is pulling the center foil downward at speed. As you move from the center to the sides, the angle changes a bit. And again there’s a shocking amount of turbulence on the three telltales near the wing stand. You can see this in the video: at about 20 mph, all of the cardboard goes straight back, but by 40 mph, the telltales near the mounts are flapping up and down like crazy.


The cardboard strips have a bend in them at the end, and I wondered what would happen if I cut this flexible section off. I did that and re-tested both heights and much of the turbulence went away. That dampened the signal too much! It occurs to me that if I started that way, I never would have discovered the turbulence.

I then moved the airflow visualizer 4″ higher and 12″ further back. Theoretically the air should be a little cleaner here, with less change in angle across the wing.

4″ up and 12″ back.

The downwash angle flattened out a little bit. Look at the two telltales on either side of the acrylic plate, that’s where the angle is the greatest. In the middle of the wing, the downwash angle is little bit less.

I then kept the hight the same and moved the visualizer 6″ forward. So this splits the difference between the first run and the last. It was pretty similar, and so I’m not sure moving the wing rearward further than this has much benefit. In fact you lose downforce and create front end lift by moving a wing too far rearward.

4″ up and 6″ back

I then retested at roof height and forward, and here you can see the 7 degree downwash angle on the telltale that’s inside the wing stand (level with grid lines), and the one on the outside is level with the ground. The telltale that’s in the exact middle of the roof appears to have slightly less angle.

Re-test at roof height.

Early conclusions

My testing is far from over, but I can draw some early conclusions.

For a 2D wing, get the wing as high as your rules allow, but within reason. At some point the extra weight and moment of inertia will be detrimental to fast changes of direction; a tall, wide, heavy wing at the polar end of the car is the antithesis of mass centralization. In addition, a higher wing causes front end lift because it has more leverage to rotate the entire car around the rear axle. All of this is to say that from from the wing’s perspective, it wants to be as high as possible, but from a handling perspective, lower is better. Somewhere around roof height is a good rule of thumb. This is not groundbreaking knowledge, I think most people do this already.

If you mount your wing at roof height, do you need a 3D wing? No. I’ve done the calculations for a 5-degrees offset, and there are very minor gains: 5% max downforce and 10% average gain in efficiency (entirely drag related), which is completely irrelevant at practical racing speeds. At 7 degrees, it wouldn’t be that much different.

It’s worth noting that with the wing 4″ higher than the roof, the downwash angle flattened out slightly. I would imagine there’s less turbulence at that height as well, but I couldn’t tell because I’d cut the telltales shorter before I tested that. In any case, if your rules allow you to place a wing above roof height, there’s some small benefit.

If you are limited in height to say 6″ below the roofline (SCCA Super Touring), then a 3D wing might be a better option. On a Miata, at this height, the change in angle between the ends of the wing and the middle is 10-15 degrees and this would likely cause a 2D wing to stall somewhere along the length. This of course depends on the roofline shape, and a car with less curvature in the canopy (E30, etc.) would be fine with a 2D wing.

The wing stands are 41″ apart, and so if I was designing a 3D wing for a Miata I would make the center section about 41″ wide, and put the outsides at a steeper angle. My guess is that the wing stands themselves are affecting this, by guiding the air along them, and that the center section that has more downwash angle is in reality a bit narrower than 41″.

If you have a 3D wing, I wouldn’t mount it at roof height unless the offset between the ends and middle is 7 degrees or less. All of the 3D wings I’ve seen have 10-15 degrees offset, and it’s likely that part of the wing will stall if mounted at roof height. You can find out the ideal placement of a 3D wing by making your own airflow visualizer, but lacking that, I would find pictures of wind tunnel streamlines for your car and measure them.

In fact I could have gotten pretty close using the following image from a wind tunnel test. However, it wouldn’t have told me the width of the center section that’s at a steeper angle, nor the massive turbulence hitting the inside third of the wing.

Maybe I didn’t need to make the airflow visualizer?

The cause? My guess is that air is wrapping around the sides of canopy, add it must detach in this region. This makes a lot of sense, because air doesn’t like to change direction at more than 12 degree or so, and needs help in the form of strakes, vanes, and other tricks to do so. This is true whether we’re talking about air going over a roof, under a diffuser, or around the sides of the canopy. The Miata’s hardtop wraps around too abruptly on the sides, and air can’t stay attached. The result is downstream turbulence hitting the wing.

And that’s probably the biggest discovery in all of this, which is that there could be enormous gains in feeding the wing cleaner air. The OEM Miata hardtop looks cute, and it functionally covers the hole where the soft top used to be, but probably 25% of the wing area experiences turbulence that decreases downforce and increases drag.

Changes to roofline shape, adding guide vanes, a spoiler, and perhaps even vortex generators on the sides of the canopy could be tricks that decrease flow separation in this area, and make a wing perform better. A 3D wing is not the answer to this problem, it would experience the same turbulence in this location, or perhaps more, being mounted lower.

Location of turbulence.

The following photo is from a wind tunnel test of six generations of Corvettes, this one an early hardtop version that’s similar to a Miata hardtop in shape. You can see the turbulence and separation along the side of the canopy. If you watch the video, you can see the flow stays attached down the middle of the car; it’s the air trying to wrap around the sides of the canopy that’s the problem.

Photo from Motor Trend.

Future tests

Now that I’ve built the tool, there are a number of tests I still want to do.

  • New telltales – The sensitivity of the winglets plays a big part in visualizing what’s happening, and I have some ideas that are better than strips of cardboard. Ideally I’d measure the angle, amplitude and frequency using a potentiometer.
  • Spoiler – You don’t see many people using a spoiler and a wing, because racing rules seldom allow that, but a spoiler might be useful to change the shape of air before it hits the turbulent portion of the wing. Yes the spoiler is behind the canopy, but it could build up a local high pressure region in front that might help in some mysterious way. Anyway, probably can’t make it worse.
  • Custom top – I have three custom tops I’ve built (fastback, shooting brake, coupe), and I’ll update this post when I test them. But knowing what I know now, I have an idea for a new roof that has the specific task of feeding the wing clean and level air.
  • Different cars – I can clamp the airfoil visualizer to any set of wing stands and change the width using the all-thread rod. With that I can measure the airflow on any car, at any height. Maybe I’ll create a database of “wind shapes” for different vehicles?
  • Wing stands – My airflow visualizer is basically two enormous wing stands, and perhaps this is affecting the telltales that are immediately next to them. I’ll mount the airflow visualizer to standard wing mounts and see if there’s any change.
  • Open top – In my aero tests at Watkins Glen, an open top reduced the wing’s downforce by 2.5x. At the time I didn’t know if this was turbulence or a change in wing angle, but at this point I’m leaning towards turbulence. On second thought, I probably won’t bother testing the airflow visualizer with an open top, because anyone using a wing and open top needs more help than I can give them.
After cutting the cardboard shorter and coloring the telltales for better visualization.

Gridlife Touring Cup Wings and Aero

Justin Lee of Miatahubs recently asked me what I would do for a Gridlife Touring Cup (GLTC) wing that measures no more than 250 square inches. This sent me down a rabbit hole of possibilities, and to consider all their aero rules and options.

GLTC Wing Rules

Racing rules change all the time, and aero rules even more, but let’s take a look at the current rules.

  • A rear aero device, oem or aftermarket is allowed. If the surface area is above 250in2 up to 701in2 (chord × length) the adjustment must be taken.
  • Only one aerodynamic element such as a wing or spoiler is allowed. The only exception to this is OEM hatchback spoilers which are utilized to mount wing brackets, are not considered an aero device. For example, a dual element wing is illegal, a single element wing with a trunk mounted spoiler is illegal, a single element wing mounted to the oem hatchback spoiler is legal.
  • Entire assembly (including endplates, and wing mounts) may not extend more than 5” past the most rearward part of the rear bumper when looking from above. There are no height restrictions.
  • Each wing endplate is unrestricted, however, should not exist in space that causes potential contact to cars during close quarters racing. Wings causing potential or actual contact problems will be asked to be removed.
  • Gurney flaps are allowed. Active aero is not allowed.

In the first bullet it says, “the adjustment must be taken.” The adjustment is a 3% hit to the lbs/hp ratio. GLTC is based on 12.5 lbs/hp, and if your car is 2500 lbs and 200 hp, then a wing costs you 6 hp or 75 lbs. That’s not a very large penalty for aero, and if I was racing GLTC, my first instinct would be to use a 700 square inch wing. That’s 9″ x 77″ or 12″ x 58″, and fairly large.

You can also run a wing that’s under 250 square inches for free. Well that’s interesting. I don’t know why they chose 250 square inches, perhaps some sedans with wings (Integra-R, WRX STi, etc) measure in under that size? In any case, that’s a pretty small wing and one wonders if it will do anything. Let’s find out.

250 sq-in Wing Options

I have a 53.1″ wing that measures 4.7″ chord, which calculates to 249 square inches. It just barely squeaks under 250, perfect. Originally the wing was flat on the bottom, but I modified the bottom so it’s curved.

Cheap wing with modified bottom profile.

It now looks like a bit like a Wortman FX72 airfoil or a GOE-525. I’ll explore that in Airfoil tools and set realistic Reynolds numbers (realistic car speeds) and see how it performs.

fx72150b-il at 120% thickness, -5 degrees, upside down, and reversed.

Notice first in the Cl v Alpha (angle of attack) chart, the wing makes the most downforce at around 10 degrees. There isn’t much difference in Cl (downforce) between the 200k and 500k Re lines, and so the wing is generating downforce, despite the low speed.

Reynolds numbers 200k and 500k with Ncrit 5.

Below that chart in the Cd v Alpha, you can see that the drag really spikes up after 10 degrees. Lift and drag are combined in the Cl/Cd v Alpha, and you can see the wing is most efficient at around 5 degrees, but I’d go for more downforce and aim for 8 degrees. Downwash from roofline shape would come into play here, and so I’d probably set the wing to about 3 degrees and then play with the AOA from there.

Notice also in the Cl/Cd v Alpha chart the difference between the 200k and 500k Reynolds numbers. At low speed (or small chord), the wing is about 60% as efficient. If I included 1 million Re, the wing would be about 2x more efficient than 200k. In other words, a bigger wing would be more efficient.

All in all, this is a good wing profile for low Re, and there aren’t many wings that have a max Cl above 2.0 at 500k. A custom wing built specifically for this purpose would have more chord and camber, but wouldn’t give that much more downforce at these speeds. Anyway, for a $50 wing modified with a pine baseboard glued to the bottom, this is not so bad.

To optimize the wing I made oversized end plates, added a 1/4″ Gurney flap (5% of the chord) to both the wing and the end plates. I sent the wing to Justin for testing. He sent me back a pic, and it looks good, but I wonder how it’ll work? He’s racing GLTC this weekend, and I’ll look at the data after the race and report back.

250 sq-in wing doesn’t look half bad on Justin’s car.

Calculating Downforce

I previously wrote a post on a simplified way to calculate downforce, and I’ll use that same calculation here: (mph*1.47)^2 * .00119 * sq-ft.

I won’t make you go through that, so I’ve done the calculations and listed downforce for three speeds: 50 mph, 62 mph and 82 mph. These speeds are specific to Mid-Ohio, where Gridlife will be racing.

  • 50 mph is about the min speed through the slow turns: T2, T5, and T12.
  • 62 mph is the average minimum cornering speed of all corners combined.
  • 82 mph is the min speed in T1, the fastest corner.

Let’s see how the no-points 250 sq-in wing stacks up against Justin’s 67″ 9LR wing, in downforce.

MPH 250 sq-in wing623 sq-in wing
50 mph11 lbs28 lbs
62 mph17 lbs42 lbs
82 mph30 lbs73 lbs
Downforce at cornering speeds

The standard 67″ 9LR wing is making almost the same downforce at 50 mph as the little wing does at max cornering speed. In addition, the larger wing would be more efficient at these speeds because of the larger chord, which brings the Reynolds number up to a more useful range. Before you write me and say that I have this backwards, please go look at data for wings at low Re. At low speed, a larger chord is more efficient. I know this totally contradicts your aeronautics courses, but they were talking about flying, and we are never getting off the ground.

Downforce increases grip relative to the weight of the car, so a lighter car gets more out of a 250 sq-in wing than a heavier car. Justin’s car is pretty light, and I wouldn’t even bother with a small wing like this on something 3000 pounds or heavier.

But a free wing is free, and it adds a not insignificant 30 lbs of downforce in T1, which is about 1.2% more grip on a car that weighs 2500 lbs. That could be useful, but through the slower corners, the downforce is negligible, while causing drag on every straight. Still, T1 is important, and extra rear grip would be useful when braking from high speed into China Beach.


Another free option is a 250 square inch spoiler. I wrote a whole article on Miata Spoilers, but to recap it, spoilers are effective at three things.

  • Changing the shape of air as it passes over the car. This can reduce drag and cancel lift.
  • Creating a high-pressure zone on the rear deck lid. Pressure is akin to weight, and that means downforce.
  • Moving the center of pressure rearward, which usually makes a car more stable.

I’m a fan of spoilers, I like the way they look, I have data that prove they work, and they are cheap and easy to DIY. If you made one 8″ tall by 41″ long, with a semi-circular shape on each end, that would be 250.26 square inches. Shape the bottom center portion so it follows the cambered trunk line and it would be less than 250″. It could be made adjustable for angle, but for simplicity, I’d set it at 70 degrees.

Simple spoiler might work better

A spoiler of this size should have a coefficient of lift around -.45, which is pretty significant. (This is according data from MacBeath’s book, Competition Car Downforce, and also corroborated in my spoiler testing.) In the 250 sq-in size, I’m guessing a free spoiler might be faster than a free wing.

You might be thinking, wings are more efficient than spoilers, how could a spoiler be faster? Because wings are designed to fly fast, and that’s when they are efficient. At low speeds, or with a small chord (which is the same thing), wings have more drag, less lift, and in this size, might be slower. I tested a 9 Lives wing vs a 7” spoiler at Pineview Run, which is a very low speed track. The spoiler was .5 seconds faster in redundant tests. I’ll have to test both the 250 wing and spoiler at a faster track, just to close this loop.


Gridlife allows a splitter up to 3″ in front of the bumper. This is a bit on the small side, especially considering the 701 sq-in wing allowance. Most racing rules allow a longer splitter, but whatever. I tested my car with a 4″ splitter and it dropped Cd by .01 (a splitter is less drag) and increased downforce by .38 Cl (win, win). So I’ll estimate that a 3″ splitter will be worth .007 drag reduction and a .30 delta to downforce. I’ll use this later in simulations.

The GLTC splitter rules are pretty well written, and say that if you have a bumper or undertray or anything that acts as a splitter, they’re going to call it a splitter. This is a recent rules addendum: “Any upper horizontal surface that is non-OEM and exposed to airflow will be classified as a splitter regardless of protrusion from or location in the front fascia. Example: Flat section floor in front of radiator.”


I’ll run some simulations based on GLTC rules. For the cars with aero, I’ll penalize 3% for a wing, and another 3% for a splitter. That penalty can occur through reducing power or adding weight. If you read my previous article on the subject, then you know adding weight is theoretically faster, but I’ll do one simulation with the same weight and detune the car for less power.

Here are the 6 different builds:

  1. No aero: This is the baseline build at 12.5 lbs/hp. NC Miata, 6 speed, drive. I made the Cl about zero to simply calculations on my end, but in reality there would be some positive lift.(Cl 0.001, Cd .45, 197 hp, 2463 lbs.)
  2. GLTC wing: This is the 250 sq-in wing only. It’s free aero, it should go faster. I’m guessing on the drag and lift values. (Cl .2, Cd .46, 197 hp, 2463 lbs. )
  3. GLTC wing, splitter: Add a 3″ splitter to the above. Car pushes a bit already, so this should balance the car better. Notice the increase in weight to offset the 3% aero penalty. (Cl .5, Cd .453, 197 hp, 2537 lbs.)
  4. 67″ 9LR wing: Remove the wing and splitter, replace with Justin’s 9LR wing. I can’t simulate aero balance, but the car might push in this configuration. (Cl .6, Cd .49, 197 hp, 2537 lbs.)
  5. 67″ 9LR wing, splitter: Now add a splitter and get the car to be more neutral. Bigger weight penalty running two aero items. (Cl .9, Cd .483, 197 hp, 2611 lbs.)
  6. 67″ 9LR wing, splitter, detuned: This is the same configuration as above, but instead of adding weight, I’ve removed power. So the car weighs the same as configuration 1, but has 11 hp less. It’s the same power/weight ratio as #5. (Cl .9, Cd .483, 186 hp, 2463 lbs.)
Vehicle1. No aero2. GLTC wing3. GLTC wing, splitter4. 9LR wing5. 9LR wing, splitter6. 9LR wing, splitter, detune
Lap time1:37.351:37.871:36.321:36.361:35.931:36.03
Min Speed47.9648.1948.5248.6348.9449
Max Speed121.89121.31120.63119118.35117.15
Max Lat G1.251.291.341.371.421.43
Max Accel0.40.410.410.420.420.42
Max Decel-1.41-1.49-1.59-1.63-1.7-1.71
Aero Efficiency00.431.11.221.861.86

The fastest configuration is #5, the big wing and splitter, with added weight. It’s .1 seconds faster than #6, which has the same power/weight ratio, but is lighter. And that’s expected, because you need power to overcome drag, and that power results in a 1.2 mph top speed advantage. Notice, however, that the lighter car has slightly more grip, 1.46g vs 1.45g, and so OptimumLap figures that part correctly.

In addition, Gridlife allows the 2611 lb car to use a wider tire than the 2463 lb car, and so the heavier car would have wider rubber as well. Finally, if you win a GLTC race they add 150 lbs to your car. That’s a larger penalty to a lighter car than a heavier one. If you can exploit the rules, lighter is not always faster!

Configuration #3 is interesting, with the splitter and GLTC-250 wing, it beats the no-aero car by over 1 second. In the real world, it might have good aero balance and be a good choice overall. Hard to say without real-world testing tho. Oh hey, Justin did that.


Justin tested both configurations of wings at Gridlife, and you can watch the Gridlife live broadcast and find him. He had to add 75 lbs of ballast to use the bigger wing to make the race weight, and so this is a great real-world A/B test. Justin sent me the data, and I’ve marked the points of interest on the speed trace. The 250″ wing is red, the 9LR wing is blue, and I’m using three laps from each to normalize the data. Let’s jump into it.

Mid-Ohio speed trace.

Exhibit A – This is Turn 1, a fast left hander. The bigger wing has more downforce and grip, which translates into a higher corner speed, and that pays dividends on the next straight. Despite having more drag, the bigger wing is faster all the way down to the Keyhole.

Exhibit B – Turn 2 is a slow and long right hander. The small wing is making about 11 lbs of downforce here (probably less considering the low Reynolds number), which is negligible. The 9 Lives wing is making 2.5 times more downforce, and while this is only about 1.1% more grip, it’s significant because of the long straight coming up.

Exhibit C – You’d think the smaller wing would have a higher top speed, and on average, it does. The difference is just over 1 mph, 124.83 vs 123.56 mph.

Exhibit D – Through the esses, the area under the curve is greater with the bigger wing, meaning more grip. If you look at the time graph on the bottom, the esses are a big chunk of the time difference.

Exhibit E – The entry of Turn 9 requires confidence, and Justin is able to trail brake better through here. Notice the hockey stick shape is the same for blue and red, but the blue is done at significantly higher speed. That extra confidence in corner entry means a higher min speed, and he’s faster through Thunder Valley.

Exhibit F – In low-speed corners, the blue and red lines are close together. The faster the corner, the more they diverge. Simply put, the higher the speed, the more the big wing is at an advantage. You can see this clearly in T1, and again here in T11.

Exhibit G – The final corner is a low speed carousel where the little wing can’t do shit, and the big wing can. If you ever think to yourself 1% more grip is nothing, think again.

Exhibit H – The table in the upper left corner shows peak values for speeds. I should really do one of these for lateral Gs, because that’s where the real difference lies, but the results are so obvious I’m not going to add another graph.

Final Conclusions

Real-world testing confirms that a large wing with a 3% penalty to lbs/hp is faster than a smaller wing in the “free” 250 sq-in size. Simulations predicted that the big wing would be about a half second faster, while in the real-world, it was a difference of about 1.25 seconds. Justin reported that the bigger wing gave him MUCH more confidence, and that’s probably where the extra time came from.

Certainly there’s some noise in the data, but Justin is a really consistent driver. I looked at 11 laps from each wing and calculated theoretical best laps from a four-sector map, and the data was virtually the same. So while the numbers might be off a bit here and there, the trend is clear.

Is there a legitimate usecase for a smaller wing? Possibly. There’s been some chatter online postulating that a 250 sq-in wing on a very light car, or perhaps using a fastback, would somehow be better. Those things would add more grip, but think about it for a minute… a lighter car, or a Miata with a fastback would get even more out of a larger wing! If anything, the gap would increase and show that the larger wing was even better than the small wing.

It’s also worth noting that we’re comparing a 250 sq in wing with a 67″ 9LR wing which is only 87% of the area of the 701 square inches allowed in the rules. If we were comparing the rules limits of 250 vs 701, you’d see an even bigger gap.

However, there’s something compelling about getting something for nothing. If you don’t have aero now, and you can’t/won’t add ballast or detune your car, then a free spoiler or wing is going to be faster than nothing. With a clean sheet of paper I’d choose a different wing profile optimized for high lift at low Reynolds numbers (Selig 1223 RTL), make it 40″ x 6.25″ to maximize chord at 250 sq-in, add oversized end plates with Gurney flaps and slats and what not, get the wing high up and centered over the rear wheels as much as possible, and then brace myself for disappointment.

As a side note, it’s kind of neat to see that simulations predicted a 1:36.36 lap time, and Justin did a 1:36.872. OptimumLap can’t factor in things like elevation changes and camber, so I usually have to create a “fudge factor” by baselining off Spec Miata lap times and then I change the surface grip in the simulator to get more accurate predictions. With the Mid Ohio map, this wasn’t necessary.

That’s fast company, nice job Justin!


I ran several more simulations on a lightweight NB Miata to see what happens to power, weight, and aero using variables I didn’t use before. New variables include stock aero, weight changes, spoiler, and fastback.

HT stock170210412.3750.450.399.33
HT stock (2251#)182225112.3750.450.399.06
HT stock (2601#)210260112.3750.450.398.59
HT spoiler170210412.3750.5-0.1598.09
HT spoiler, splitter165210412.750.493-0.4597.29
HT 250-wing170210412.3750.46098.34
HT 250-wing, splitter165210412.750.45-0.397.49
HT 701-wing165210412.750.49-0.696.76
HT 701-wing, splitter160210013.1250.483-0.996.00
FB, spoiler170210412.3750.43-0.0797.89
FB, spoiler, splitter165210412.750.423-0.3797.06
FB, 250-wing170210412.3750.39-0.1597.35
FB, 250-wing, splitter165210412.750.38-0.47596.40
Various configurations using hardtop (HT) and fastback (FB).

Twin Studies: Driving Styles Examined

During our “Lemons from Lemonade” track day, Ian and I got to drive a few different cars and compare driving styles. I have a lot of laps at Pineview Run, and it’s not exactly a fair contest when it comes to straight lap times. However, Ian has been to PV before (three years previously, he got four laps) and has logged a lot of sim time on it in Assetto Corsa. Ian is also instantly quick out of the box on every track we’ve been to, and typically buries me on lap times, so I don’t think he’s particularly handicapped.

NB1 Miata

Clayton’s 1999 is a dual duty car that’s more street than track. We usually fit whatever take-off tires we’re trying to get rid of, and on this day it was shod with cycled-out 225 RS4s on 15×8.5″ wheels. On fresher tires, the times would have been at least a second faster, but for comparative times, used up rubber would be fine.

First let’s look at Ian’s three fastest laps. You can see his lines are mostly on top of each other, and that he’s got the track dialed. But he’s still experimenting in T1, the Esses, and the Knuckle.

Ian’s best three runs.

When I assemble his best sectors, he has a theoretical best lap of 1:19.927, which is about half a second faster than his best, 1:20.421. That’s damn consistent for such little track time.

Ian’s best theoretical lap.

I only did three laps in Clayton’s car, so I’m not going to show you my theoretical best, it’s my fast lap.

In general, Ian drives with more yaw, and he felt like the NB1 oversteered too much. Ian usually races a Yaris, which he has to aggressively trail brake to rotate, and a Miata will do that a lot easier, and that might have skewed his perceptions. I didn’t feel like the car oversteered at all, and my inputs and driving are more tidy. Here’s a lap of Ian first, then me.

One lap from each of us.

Let’s take a look at how those laps compare on the speed/distance and time/distance graphs. For the first 1200′ of track, Turns 1-5, we are dead even. At 1300′, where the cursor is, Ian is faster and gains a bit of time. I’m not surprised, everyone is faster than me on this downhill section.

Speed trace of our best laps.

But I trail the brakes better into T7, and have a slightly higher min speed. If you look at the view from above (switching Ian to yellow here), you can see my approach is straighter, and I sacrifice some track width on the entry, but it works because I don’t have to move track left and can brake in a straighter line.

Turn 7 differences.

The most significant difference is how we go through the Uphill Esses. From 1700′ to 2300′ on the distance graph, I gain an easy half second. This is mostly using a tighter line. I wrote about this in my track guide, which Ian helped edit before it was published, so I’m surprised he’s not taking my advice here. See below.

Tighter line is faster.

Next is the Knuckle, and this corner is a perfect example of how differently we drive. If you look at the graph, this isn’t a big deal, because from 2400′ to 3400′, there’s only .1 seconds difference. So while the outcome is close, the strategy is different.

  • Ian carries a higher entry speed and initially gains time on me. He brakes deep and stays in 2nd gear, using a line that optimizes acceleration out of the corner.
  • I short shift into 3rd gear before the corner, and maintain a higher min speed.

From there, I trail brake better into the Blind Hairpin and keep a higher min speed, and this gains me another two or three tenths. And I gain about that much again going through the S-trap and T15, which we do very differently.

Different lines through T13-15.

Ian’s Corner

This is the part of the blog post where Ian takes up the pen.

So which is the most important corner on the track? If you believe the usual wisdom, it’s the one that leads onto the longest straight. Alternatively, it’s the fastest corner. But if you look at the data, I’m actually losing a lot of time in the middle of the slowest corners. Any time spent going slow is bad, and in a slow corner, there’s a lot of time hanging around going slow.

If you watch the video, the most obvious difference is in the Knuckle, where I was experimenting with a lot of oversteer. I don’t normally drive that loose, I promise! In general, I like slinging the car all over and Mario likes driving tidy.

Beyond that one corner, the thing I noticed most was how differently we hold the steering wheel. You can see the muscles in his forearms quite clearly, his elbows are high, and he grips around the steering wheel with his thumb and fingers. In contrast, my forearms are more relaxed, my elbows are low, and I place my hands on top of the wheel, often with the thumb on top.


These differences in how we hold the wheel aren’t the causes of our driving styles but rather the result of them.

  • I create oversteer on corner entry and maintain it mid-corner. The car rotates off the rear wheels, which makes the steering light.
  • In contrast, Mario creates and maintains a little understeer as a way of feeling grip and maximizing minimum speed. The reason why his muscles are bulging is because the wheel is literally heavier as a result of understeer.

Despite being identical twins, we have completely different philosophies on how to navigate a corner. Which way is better? I think it depends on the car, track, and surface. I can say that over the course of the day I drove more and more like Mario. Understeer FTW.

NA6 Miata

Ian and I have driven my NA6 back to back a couple times. The first time was in 2018, when Pineview had just opened. At the time my car was at a lower spec (about 1/3 less power, an open diff, and Yok S.Drive tires), and we were going 9 seconds slower.

This was our first laps ever on this track, and so we were still figuring out which way the track goes. I see both of us making the same mistakes I wrote about in the Pineview track guide.

As it pertains to driving style, notice that Ian is howling the tires from the first corner, finding the limit, and making a lot of small steering inputs to correct. If you look at our runs through the Knuckle, not much has changed, for either of us. As Ian mentioned, I use understeer for traction sensing, and as a result I grab the wheel with Popeye forearms. Back then I was shuffle steering so much my hands looked like a groping teenager. These days I’m in a committed relationship with 9 and 3.

Ian and Mario, 2018.

Fast forward three years and let’s take a look at some data from our Lemonade day. Again we are using up old rubber (Pineview is brutal on tires), so we’re on 14-year old NT01 tires that have lost much of their grip and wear imperceptibly.

These laps are from the end of the day, where Ian is experimenting and trying to adopt my driving style. First, let’s look at Ian’s best theoretical lap. He doesn’t leave much on the table, and his best possible lap is a 1:16.828, just two tenths faster than he went.

Ian’s theoretical best

My theoretical best is a quarter second faster than my best, and I could muster a 1:16.424 if I could get my shit together.

Mario’s theoretical best

Next is the speed trace of our fastest laps. He owns the first sector, I take it back from the Crick to the Esses, he beats me at my own game in the Knuckle, I eke out a bit in the Hairpin and the result is the identical twins have identical times exiting T14. But the final corner….

In Turn 15, I enter wider, get the car rotated early, and I’m on the gas sooner. These are things Ian is usually coaching me to do. But this time I do it better than he does, and when we cross Start/Finish, the result is .3 seconds.


Ian and I have an ongoing discussion about why Pineview is especially cruel to FWD cars. I know this to be a fact, but he can’t wrap his head around it for the following reasons.

  • On any other track, FWD is not at a disadvantage. As proof, he’ll cite the Miroshi videos at low-speed Tsukuba, and how the Inte-R beats on virtually everything.
  • In the Pineview sim on Assetto Corsa, FWD doesn’t have a disadvantage. He’s tried all the cars, and I believe him.
  • His Yaris out corners Miatas all the time, I’ve witnessed this firsthand.

And yet, at Pineview, in the real world, FWD cars are slower than their RWD or AWD counterparts. I have data from a Mini R50 on RE71R, a JCW on RS4, a Civic on Sur4G, and a Yaris on Conti ECS, and from Miatas on the same tires. In every case you can see it in the data, the FWD cars don’t have the same grip, but they get out of corners quickly. Ian will counter this data with “every car that over brakes on entry has good acceleration.” He’s also opined that perhaps Pineview hasn’t seen a good FWD driver yet. And I counter that with the list of seasoned FWD racers that have been to PV. And on and on the argument goes.

On this day, Ian and I would finally drive a good FWD car on good tires, back to back and settle this debate once and for all. For this A/B test we’d drive Chris Gailey’s Hyundai Veloster N on Falken RT660s. (As a data point, I drove Chris’s car last year on the stock PZ4s, which are about three seconds slower.)

I won’t do the whole blow-by-blow thing, because the biggest difference in our driving style is that Ian is consistently later on the brakes, and I’m consistently earlier on the gas. Take a look at the blue lines below – all of the peaks are to the right of the red lines. On average, Ian is braking two car lengths later than I am.

I think Ian was also trying very hard to get the Veloster to rotate, and it wouldn’t. He felt that a smidgen of traction or stability control was still on, and that the factory doesn’t allow you to remove all of it. Personally, I felt the car handled great, but I’m an understeer guy, and I’ve probably never properly rotated a FWD car.

But apparently I can drive OK that way, because if you look at the valleys, the red lines are all to the left of the blue lines. I’m backing up the corner better, and on average, I’m accelerating about two car lengths before he is. However, in the final corner, it’s a difference of three car lengths, and I get a little more time. This is pretty much the same thing we saw in my NA6.

In the end, I put down a really good lap and almost broke the all-time FWD lap record, I was just .2 seconds off! If Ian and I combined our best sectors, we’d have beaten the record by .2 seconds with a 1:15.136.

Does this mean that FWD cars aren’t handicapped at Pineview. No. Josh’s FRS had about the same lbs/hp ratio in 2019, and he was doing mid 1:14s on Z3s. On RT660s, he’d be doing 13s or less. Heck, Alyssa can do 1:14.5 in her 18 lbs/hp Miata on RS4s. Now I’m not as fast as those two aliens, but I should be able to do 14s on this tire and lbs/hp combination, and I can’t. I blame FWD.

So while we still don’t know why FWD is slower, there’s one thing we can conclude from this test, which is that I can drive a FWD car! Of the myriad reasons a FWD car is slower at Pineview, it’s not the driver. Or at least not this driver. Now we can concentrate on eliminating other factors, and maybe one day solve this mystery.

Ian’s final corner

Let’s start with the facts: Mario is faster at Pineview than me in every car we drove. He knows the course better, and his technique in slow corners is better. This isn’t very surprising. Over the last 2 years, he’s gone to Pineview dozens of times. In the last 2 years, I’ve had 6 hours of track time. It would be shocking if he wasn’t faster.

Rewind nearly 10 years ago, when we first started racing, and you would see that Mario was always faster than me by a few seconds per lap. Why? I think he understood traction better because of his motorcycling experience and a general knowledge of racing. I was of the opinion that cars were for commuting.

But then I started studying with books and simulators. I also got into data analysis and coaching. It took hundreds of hours, but eventually I became fast by educating my mind and body. After all that work, it was gratifying that I was a few seconds per lap faster.

And now Mario is faster than me at Pineview. Why? He started studying. He got into data and coaching. He changed his philosophy from “I just show up and drive, man” to “how am I going to train today?” There is no shortcut to the process. You have to put in the hours. You have to practice. You have to study. If you’re not being coached, you have to become introspective and actually critique yourself. It’s not always fun. It is rewarding though. Eventually.

I made sure to say “at Pineview” above because I think I still have the edge at many other tracks, especially unfamiliar ones. The reason is sim racing. I’m much more tuned into reference points than he is. I’m also more comfortable recovering from bad situations because I have the luxury of being able to do stupid shit in the virtual world I’d never do in the real world. Again, these differences are the result of a commitment to hundreds of hours of study and training. At the end of the day, or the race, or the blog post, let that be the final lesson.

Making Lemonade from 24 Hours of Lemons

I’m standing in a puddle of my own shit and piss. I have it on my arms and legs, and I’m reflexively spitting because I got some in my mouth. I think to myself, maybe this is a sign?

Let’s rewind the clock a couple weeks. I’ve been preparing to race the 24 Hours of Lemons race at NJMP Thunderbolt. I’ve done Lemons races on the east coast and in California, but none in this Miata.

The first hurdle was passing tech. Most people think about 24 Hours of Lemons as cut-rate racing, and in some ways it is, but they cut no corners on safety. The Lemons cage rules are more stringent than any other series, and my car, which has passed tech in AER, Champcar, NASA, and SCCA, won’t pass Lemons tech. Whoever built my cage put the back stays in at a 29-degree angle, and they need to be 45 degrees, give or take.

I called my friend Tom Pyrek, whose minivan I race in 24 Hours of Lemons, and he agreed to help me out. Except that he didn’t have a lot of time, and things got pushed back a few days right when I was crunched for time. That was a headache I didn’t need, but that’s how these things go. Anyway, it gave me more time to work on the theme.

Most people these days don’t bother with a theme for Lemons, but I still think it’s an important part of the series. I wanted to do the famed Ferrari Breadvan, an iconic race car from the 1960s. Other people have done this theme (and probably better than we would), but I wanted to do it mostly to try out the aero.

The shape of things to come…

Our plan was to serve pizzas out of the back in the evenings, and so I wrote “Pizza is always the answer” on the side of the car. We were even going to deliver a pizza to the judges in the penalty box at some point during the race. If we didn’t get a penalty, we were going to do it anyway, just to recreate this moment.

Fast forward and it’s the Friday a week before the race, the new back stays are welded in, and we’ve just started the car for the first time in a while. There’s bit too much white smoke for my liking, and so we check the compression numbers: 165, 130, 130, and… 60. We put a bit of oil down the plug holes and the numbers come up, and so we knew it was rings.

Humph. Not terrible, but that one cylinder is concerning. The last time it ran was at PittRace, and we were the fastest non-swapped Miata. Alyssa was doing 2:03s on well-worn RS4s (faster than the Spec Miata record). We all felt the car was running well, so it never occurred to me to check if the engine was still healthy. There’s a mistake I won’t make again.

Ran when parked. Alyssa and Mario on grid at PittRace a few months before.

At this point I have three options. A) Run it as is and probably blow it up at some point during the race. B) Pull the engine and install new rings. C) Pull the injector and plug on one cylinder and run it as a triple.

Option C isn’t as insane as it sounds. In fact there are a whole group of motorcycle racers in the Pacific Northwest that neuter one cylinder of a 600cc four and effectively make a 450 triple out of it. They call these “Cripple Triples,” and it allows them to compete with 650cc twins on equal footing.

Now this option is so Lemony I want to do it, but when we try it, it’s super slow. I run it down the street and the engine feels like it has 50 hp. Somehow that missing cylinder, even with only 60 psi, is very important. So we decide to do option B, and rebuild it. Or at least throw in a new rings, hone the cylinders, and put in new seals. No problem.

The race was in a week, and so I ordered all the gaskets and a piston ring set from Auto Zone on Friday. I thought that should be plenty of time to pull the engine and assemble it, and in fact most of the parts came the next day. Except the piston rings. Monday came and still no rings. Fed-X said they were supposed to be there yesterday, and the new tracking info said Wednesday. Feeling queasy about that, I ordered another set from NAPA, this time next day air via UPS. I paid the price in rings for the shipping alone, but I wanted to cover my bases.

With the engine out of the car, it occurred to me that I had another block sitting on a shelf up at Berg Racing. That engine overheated and I’m sure the rings relaxed, and so it needs new rings as well. May as well do them at the same time. So I drive up and gather it, and we start the surgery of taking that one apart on the bench as well. I’ve got nothing else to do.

I just need rings

Tuesday comes and still no rings. Next day air, my ass. But I’m not sweating it, we can get the engine in quickly. I tidy up other things on the car and finish the theme.

Breadvan theme was ready

Wednesday no parts, either by truck or by air. I went to Auto Zone and Napa and made them call the manufacturer and find out what the problem was. In both cases the shipper hadn’t picked up the packages yet.

Are you fucking kidding me? The manufacturing plant in Tennessee says the parts are right there waiting for pickup, but neither UPS or Fed-X can be bothered to actually get them? I’m feeling like this is a sign or something. We are also having trouble getting rear brake pads. I ordered them a week ago, and they still haven’t arrived.

Around this time my support vehicle (Honda Element) started making belt noises. We use the Element for bringing extra tires and spares, and I really don’t need it to break down on the way there, or the way back. We mess with the belts and the noises go away, but then it throws an engine code. Bad omens or what?

And then my RV, which is race headquarters and also my tow vehicle, throws engine codes for a misfire. And then the brand new tire that I just replaced is slowly leaking. FUUUUUUCK! Everything was falling apart at the same time!

For the past few weeks this has been happening in little subtle ways. I’ve been looking at all these signs, portents, and omens and patently ignoring them. But I pressed on for good reason: My brother was flying in from California; My buddy Chris was flying in from Detroit for his first ever wheel-to-wheel race; And I don’t believe in omens. So I was going to make this happen, my dog spinning upside down on the ceiling and speaking latin backwards, or not.

I then went to dump out the RV tanks at a local state park. I was doing this on the down-low, without paying the sewage fees, so trying to be a bit sly and get it done quickly. Well, some genius (ahem) left the guillotine valves open last winter, and so when I unscrewed the cap to attach the sewer line hose, my own shit and piss sprayed out all over me. Yep, I literally shit all over myself.

It was at this time, standing in a widening brown puddle, my arms and legs covered, no longer the least bit stealthy about dumping my waste tanks without paying for it, reflexively spitting and wondering how sick this shit would make me, that I began to believe in omens. I stitched together all those signs and portents and finally figured it the fuck out. I contacted all my teammates and told them we were done. We aren’t racing. As my brother put it “when the shit hits the man, it’s time to reassess the situation.”

And Yet, Lemonade

Instead of racing, I invited all my teammates to Pineview Run. They were all coming east anyway, it was the least I could do. We spent the day hooning and gather data, and fun was had by all.

Ian drove my wife’s Honda Civic. We didn’t use it much because the VSA really overwhelms the brakes. (See my blog post Autocross N00b for the 10-step procedure to turn that off.) It wasn’t fast or fun to drive anyway, and Ian got in a lot of other cars.

Clayton drove his NB, and both Ian and I got to drive it and compare notes. I’ll follow up on this data in another post, it’s rather interesting to see how differently we drive the same car. We also got to see how Clayton drives, and while he’s off the pace, his instincts are really good. He’s a natural driver, and will keep getting better.

Jim brought his 240 hp turbo 1.6 Miata. I didn’t get a chance to try it, but I’ve driven it on the street and it’s a blast. Turbos are not my choice for track cars, but street cars, yes please. Unfortunately the charge pipe kept coming loose, both on the track and on the drive to and from, and by the end of the weekend Jim had replaced every hose and clamp.

The front end of Jim’s car is called a Wizdom. It wasn’t exactly an easy fit, but we made it work. I also made a custom undertray and left it long in front to make a splitter, and added a hood vent. It’s a bitchin car all around.

Fitting the Wizdom bumper, before splitter and other enhancements.

Chris brought his Veloster N, which is an impressive car and I regularly think about buying one. He fitted Falken RT660 tires, and while his car isn’t set up with lot of camber, he wisely got the tires heat cycled from Tire Rack before delivery. Meaning, the tires didn’t delaminate, which is what they do when run without heat cycling and with less than ideal camber.

I ran a 1:15.565 in the N, exactly two tenths of a second slower than the all-time FWD record lap time (CRX on A7s). On A052s or Hoosiers, I’d own the FWD record. In a bone-stock car. I need to make this happen.

Ian also got to drive the Veloster and I was thankfully faster than him, but only by half a second. He’s a FWD expert, and I’m a Pineview expert, but I sure as shit was not going to lose to my brother on his second ever visit to the track! Phew.

I drove my 1.6 Miata mostly on 14-year old NT01s (I shit you not), and they still grip, and wear imperceptibly. I also finally got a chance to try some take-off 245 R7s. Oddly I didn’t go any faster than I did on 205 R7s (1:14.5). The steering effort was absurd, and I overall didn’t love them. I’ll have to play with pressures and stuff and see if I can get them to work.

Honestly, it was a fantastic day, maybe even more fun than racing because we were all on track together, with five cars instead of one. Yeah, I had to eat the Lemons entry fee, the NJMP practice day, all the parts and labor to prep the car, and $800 worth of track fees at Pineview. So I lost over $3000 on this “race” weekend, but I guarantee it would have been worse if we had gone to Thunderbolt. Something bad was going to happen. Every force of nature was against me, every step of the way, I just wasn’t listening. I am now, tho.

My wife says, you win or you learn, and education costs money. It reminds me of the old joke: Do you know why divorce costs so much money? Because it’s worth it. Missing this race was oddly worth it; we made lemonade out of Lemons.

The Punchline

Saturday afternoon, the day we were supposed to be racing in New Jersey, I got a phone call saying that the rings arrived at NAPA (five days late). The next day, Auto Zone called to say the other set of rings arrived (seven days late). In the race of rings, UPS next-day air beat Fed-X freight by a day. But they are both fucking losers in my book.

Maybe I should have spent more time on mechanicals than theme?

Everturrible Race and Tire Report

This is an overdue race report from April. At first I was waiting for the Lemons wrap-up video to come out (it was late), and then I just got to doing important things. Anyway, it was a fun time, and there’s some neat data to look at if you stay with me to the end.

Every year I race Tom Pyrek’s minivan in the 24 Hours of Lemons. We’ve done NJMP Thunderbolt, Thompson, and PittRace together. The van has been the source of some really good themes, and is better on track than it should be. It’s a first gen Honda Odyssey (also badged as an Isuzu Oasis), the one with 4 doors instead of a slider, and with an anemic 4-cylinder Accord engine. Tom’s minivan racecar has thankfully been manual swapped, but it still struggles with an open diff, terrible aero, and a lot of weight.

Tom has been down the road chasing performance, and it’s just not his gig any longer. He’d rather have fun racing an underperforming slug than stress out over performance. I can understand that, but I’m also about incremental improvements. I can’t do anything about the power or diff, but I’ve been subtly adding performance year by year.

The last time we raced the minivan I begged Tom to allow me to add an airdam and splitter. I did that modification overnight between race days, and so we got to A/B test the modification. It was noticeably better, and so Tom even left it on for street driving after that. Yes, it’s street legal, and he drives it regularly.

Front end aero. Worth it.

This year Tom also allowed me to remove the spare tire from high up on the passenger side window, and remove the third row seat. I was like “Bruh, do I even know you?” The wide-screen TV set would still remain bolted to the rollbar, because I don’t have that kind of influence yet, and Tom is still Tom. But I’ve made some solid progress on both aero and weight.


The minivan is a source of countless themes, and we’ve had some good ones. The first time I raced with The Awkward Corner, we were the Teenage Mutant Ninja Turtles.

Clockwise from top: Tom, Brian, me, somebody’s borrowed kid, Melody, Dieter.

I think we kept that theme for Thompson. The first time we raced PittRace, we changed it to Mister Rogers Neighborhood Trolley. Mr Rogers is from Pittsburg, so this was appropriate. The Swangers in the pic below are just for show, we didn’t race on them; the spokes can come loose and they weigh like 80 lbs each.

L to R: Melody (driver), Eric, me, Dieter, Tom

This year we were stuck juggling theme ideas until the Evergiven tanker got stuck in the Suez canal. Lightbulb! It finally got dislodged before the race, but we stuck with the theme.

The Awkward Corner race team. Everturrible.

There are more theme pics at the very end, with keen details to appreciate.

The Race

The race went from good to bad, and then to good back to bad again, and then good. Pretty normal.

The first problem was tires. PittRace is notoriously hard on tires, and we’d fitted budget Accelera 651 tires. These are unusual 200 TW tires because they have three center grooves instead of the usual two, and come with a FREE MONEYBACK GUARANTEE. They are cheap (arguably free), and come highly recommended by my friend Brian Smith at Easthood Racing, so we figured we’d try them out. We chunked them in our second session.

The Accelera 651s chunked quickly

In foresight, I’d brought two wheels from my wife’s Honda Civic, with the intention of A/B testing them for one session, and then secretly putting them back on her car before she knew. Her wheels were shod with Nexen NFera Sur4G 200TW tires, which we’ve raced on the van before (in a 16″ size, not these 17″ ones) and they’ve worked well.

So we changed the tires quickly to my wife’s tires, and on the plus side, this gave me some great back-to-back data that I’ll share later. On the down side, the tires delaminated, and now I’m sleeping on the couch.

The Sur4Gs delaminated just as fast!

If you’re thinking the minivan racecar doesn’t have enough camber, it has -3.5 degrees. Tom says that tire deflection has a lot to do with wear, and that it was the 7″ wheels that were to blame. So we rotated tires around as best we could, and used some older tires that were well heat cycled (or cycled out) on the front, and did our best to finish the race.

So with our tire woes somewhat sorted, the next thing was our brakes. I could hear something grinding and vibrating, so I brought the car in. We got the wheels off and noticed it had a cracked front rotor.

We just can’t catch a break.

We replaced that, put the car down, and I didn’t get to the track entrance before I brought it back in with the same sound. It was, in fact, the rear pads that were shot. These were brand new parts-store ceramics, which usually last a couple races when the parking brake isn’t dragging. Which it was. We replaced these with a used set of organic pads, but those barely lasted one session. Luckily we already had someone out and about buying spare parts, and so we got more rear pads to finish the race.


At one point I got punted from behind by another driver in a E30. In a Lemons race, anyone involved in contact gets a black flag, regardless of who’s fault it is. And so this resulted in me collecting my first black flag since 2013. If you watch the video, I think you’ll see this wasn’t my fault; it may not matter to Lemons, but it matters to me.

I later confronted the driver, who was a total douche about it and called it my fault. But after the race the other guys on the team said their apologies, and I countered with “racing incident”, and we’re all good now.

It was a dicey moment at the time. I had to react quickly to go straight off, then turn and gas it right before hitting the tire wall, which vectored me away so that I only grazed the right rear. But this gave me more speed which I had to scrub quickly to avoid getting T-boned in T18. I stopped it before re-entering the track, barely.

These are, unfortunately, the moments I live for.


This was my first time with a three-driver team, and I prefer four. We had a lot of downtime in the pits fixing shit, and a fourth hand is helpful for that, and if we didn’t break so often, we’d have had the same amount of track time anyway.

I haven’t mentioned our new teammate Stephen Kent. Tom found him on the Lemons Rally, and so I already knew he was good people. What I didn’t know is that in his first race ever, he’d impress the hell out of me.

Stephen was fast (2.8 seconds off me), did a lot of work on the car, bought food and drinks, and generally was the ideal teammate. Never mind the fact that this autocrosser (I’m making the X sign and backing away) had no previous racing experience and had never been to PittRace, he had excellent situational awareness, stayed out of trouble, got consistently faster, and played up the theme to a 10. Yes, he was the one that blocked the track exit at the end of the race. PittRace officials didn’t find this funny, Lemons people did.

We also had a pit crew member Ari, and he was a great help all weekend. He’s been doing some autocrossing, and with some help from Stephen, maybe we’ll have a fourth teammate after all.

Staying “in theme” and Lemons Wrap-up

To be theme-tastic, we purposefully stuck our van in as many places as we could: We blocked the tech garage in the morning; We held up traffic at track in, and track out; When we were pulled in for a penalty, we wedged the car between the building and a pole. We were Everturrible.

24 Hours of Lemons did a wrap up video on the race, and we are featured singing a Whitney Houston song, and generally playing up our theme by being assholes.

Tires and Data

As promised here’s some interesting tire data comparing the Accelera 651 to the Nexen NFera Sur4G. I chose a couple representative laps from each without traffic. (If you want to measure your laps vs a portly minivan, my best lap was a 2:23.135 with a top speed of 97.4 mph.)

On the speed trace, Nexen is red and Accelera is blue. You can see the Nexen makes a better lap time by about 1.5 seconds, but what’s interesting is where the Accelera if faster: on the straights. Once into high gear, the 651s have less drag, less rolling resistance, and less weight, and this translates into a consistently higher top speed.

The Accelera’s are the lightest 225 street tire I’ve weighed and are a couple pounds lighter than the Nexens. When you put both 225 tires on their sides, you can see the 651 is a narrower tire. This weight and narrowness is probably where the top speed comes from.

Tire width, L Nexen NFera Sur4G, R Accelera 651

Next take a look at lateral Gs, and this is where you can see the Nexen’s have an advantage. And they should. Sur4Gs are known to be a 200 TW autocross tire, and the 651 is a budget 200 TW tire. You can see that the red lines are both higher and lower than the blue lines, indicating more grip (low means left turn, high means right turn).

As for how they drive, I liked both, but I was able to trail brake better on the Nexens. In the friction circle, you can see the red lines are not only wider than the blue lines, but more bulbous shaped, meaning I can blend inputs better.

I drove the Acceleras slightly better than my teammates. They were about 2.5-3 seconds slower on the 651s than the Sur4Gs, while I was only 1.5 seconds slower. I practice on all-season tires, and so I’m used to a car that moves around a lot. It could be that training, or it could be that their driving styles don’t mesh as well with this tire.

Anyway, the Accelera is a slower tire, but it still might be a good endurance racing tire, the jury is still out. I need to get these tires on a Miata with proper camber, and not a FWD van that destroys anything you put on it. I have a feeling that the three grooves on the 651 will make it a good wet tire.

I sent this data and some comments over to Kaylee at Tire Streets, to say that I’m intrigued by the tires, but until you come out with a wider 15″ tire, I’m not buying. First, because all they offered was a 195/50-15, and second, 651s seem to run narrow. Well, not long after sending that email I found out the Accelera 651 is now available in both a 205/50-15 and a 225/45-15. I bought them immediately.

My plan was to A/B test them at the NJMP Lemons race vs RS4s, but due to recent circumstances (which you can read in maybe the next late race report), I have not tested them yet. I can tell you that the 651s measure exactly the same width as 225 Maxxis RC1s, and so they might not run narrow in this 15″ size.

More Photos

24 Hours of Lemons is a weekend of fun on and off track. It’s not just about racing, it’s about having a good time. The series is going more and more towards performance cars, but I wish more people brought shitty cars with themes.

We parked here in the dark and blocked tech in the morning
We pretended to get stuck a lot (that’s me in the shark costume)
Just before taking a bite out of Randy Pobst
It’s all about the props.
The source of so many memes
It’s still about the props
We made custom containers on each side of our “ship”
All of the containers… find your team!
A rare shot of us not misbehaving
Living up to the theme