Lap time simulation project

[Revvd]

Ever since I discovered this lap time simulation software OptimumLap, I've been creating cars and adding new tracks, to see how they align with real lap times. This is my idea of fun.

Now it's gotten to a point where I have a pool of 48 cars and 60 tracks (10x NA, 33x EU, 10x UK, 5x Asia, 2x Australia) and I feel like it would be a shame not to share this. Perhaps you will find it interesting as well.

Where possible, I used real dyno graphs and manufacturer's aero stats (Sport Auto magazine is a great source of these). Sadly, there's cars like AMG One and F80 with very limited data, where I had to freestyle it. That meant a lot of guess work. So just keep this in mind.

The way this software works, you can't really simulate complex hybrid powertrains or active aero well enough. I mostly worked around this by tampering with longitudinal friction coefficient to improve traction (but inadvertently also affecting braking performance, so I tried not to over do it).
You can poke holes in the numbers all day long. It's supposed to entertain curiosity, not produce relevant data.

Now to explain the results list...

• The number to the left of a track name is a correction factor for all lap times per each track.
  Typically the slower, more twisty tracks need speeding up, and vice versa.
• The number above each car name is the coefficient of friction of tires.
  The simple cars have identical longitudinal and lateral µ. The AWD, AWD hybrid, or airbrake equipped cars have these stats split to manipulate the results.
• Bellow car names is the average lap time for all the tracks, and bellow these there's difference to average reference lap time (production lap record holder or alternative).
• The little black triangle in a corner of some lap time cells signifies an existing real lap time reference.
  A note pops up on mouse hover over the cell, showing details.
  
  

All lap times are color coded:
faster than GT3 (or alternative) reference
faster than production reference
matching prod. ref.
slower than prod. ref.

The 20.8 km Nordschleife is the only track in the list which is calculated using the 24h layout and Nürburgring GP layout, given that it would be near impossible to edit the existing N24 OptimumLap track file into the 20.8 km layout, or create one from scratch.

Onto the stats list...

• power and torque are to the wheels
• car weight includes 100 kg worth of driver and fuel
• Vmax is mostly limited by engine revs, or in rare cases, drag limited

If you want to sort the data, you'll have to make a copy of the spreadsheet on your google drive and go from there.
Please, try resist the urge of telling me this or that stat is wrong (I know it is), unless you actually have accurate data.

Here's the link to the google sheets

 

[Revvd]

I also have to touch on few cars on here...

I had to dial back Dallara Stradale downforce to 66% to match real lap times. I think the factory claims are unrealistic or the downforce balance is out of whack or the tires are too narrow. Either way, the car simply underperforms.

Huracan STO is hardly any faster than Huracan Performante in real life. I didn't try to correct the factory stats, but you can clearly see that something is off about this model.

Senna surprisingly matches well with real lap times. I went into this thinking I would reveal that it doesn't really generates this much downforce. But it turns out the stats might just be true.

You might suspect that, as a Bohema admirer, I am biased and hate Valkyrie. But I'm self aware enough. And to compensate for this I have made the Valkyrie perhaps a bit too powerful trying to match the Silverstone lap time. What I should have done is simulate it without the ERS boost. Then it would fall behind Bohema.

All Ferraris have 220 kg at 250 kph less downforce than is claimed (LaFerrari way less than 220), and yet they match real life performance quite well.

 

[Revvd]

Time for some focus on Praga Bohema

So previously, majority of my simulated Bohema lap times showed faster pace than IRL data:

track	real lap time	simulated	diff.
Monza	1:46.70	1:43	-3.7
Most	1:35.09	1:32	-3.1
Slovakia Ring	2:03.60	2:01	-2.6
Dunsfold	1:09.70	1:08	-1.7
Brno	2:01.60	2:01	-0.6

The reason why I gave Praga the benefit of the doubt and didn't adjust parameters of the car (like I did with Ferrari and Dallara) is because I thought the reference data pool is not solid enough to warrant changes:

• the Monza lap was done by their... let's say less motorsport-active factory driver
• the CZ/SK ones were all set as part of a feature film for Top Gear (one set of tires, time constraints)
• the Dunsfold lap was set by Ben Collins, whose ability was now in doubt (last year the new Stig beat him by two seconds in the same Atom 4R)

The Slovakia Ring lap was the most concerning one. Being their home base, this is where they've done countless hours of testing with prototypes over the years. Even though the factory ace driver overperformed during filming and beat his internal lap record, the simulation was telling me he should have gone 2.6s faster still. That couldn't be right. Something was off.

Then there's Brno. This was the only track where the simulation and reality were a close match. The problem is, earlier this year the track was resurfaced (and gotten faster), and all I had was mostly outdated data. I had to wait for ESET Cup multiclass race weekend to gather solid reference data for GT3 Cup, KTM GT2 and GT3 race cars, which is what determines the track correction factor - that turned out to be 0.98, meaning that Bohema should be capable of 1:57.

1:57 vs 2:01.60 again shows the big discrepancy (-2.6) I was talking about.

I could now edit the specs. The first thing I tried... what I've done to all the Ferraris and the Dallara - was removing 220 kg of downforce at 250 kph. And sure enough, this immediately set the results straight!

Here's how it looks after the fix:

track	real lap time	simulated	diff.
Monza	1:46.70	1:45	-1.7
Most	1:35.09	1:34	-1.1
Slovakia Ring	2:03.60	2:04	MATCH
Dunsfold	1:09.70	1:10	MATCH
Brno	2:01.60	2:01	MATCH

Bar Monza and Most, where I suspect lap times should improve given proper effort, it's a perfect match.

The conclusion is that, just like Ferrari and Dallara, Praga also appears to be inflating their downforce claim. They quote 900-950 kg at 250 kph. I arrived at 730 kg at 250 kph (Cl*A aka SCz of 2.43).

This made the car drop 4 places in the average pace ranking, and it's now sandwiched between F80 and W1 (both being fantasy entries largely made of guesswork). Where the previous iteration of the simulated Praga could hang with GT3 race cars on a few tracks, now it has no chance.
Now that I believe I have an accurate representation of Bohema's track capability, consider me equal amount satisfied and annoyed.

It's not all doom and gloom however, because the new estimated Nordschleife lap time is 6:24 - that's still enough to dethrone the current lap record holder - AMG One. Right... Never gonna happen for the little guys. It's way more likely that Koenigsegg will pay the track a visit, moving the lap record out of reach of everyone else (other than Valkyrie) for the foreseeable future.

But imagine if Christian Gebhardt of Sport Auto did a Supertest of Bohema. Given that he's usually 10s slower than Porsche factory drivers over the shorter 20.6 km layout, what the simulation tells me is that he should be "comfortably" in the 6:30s (20.8 km: 6:34 = 20.6 km: 6:29). That would be insane, considering his fastest lap is 6:52.54 in AMG Black Series.



I'll also briefly touch on Corvette ZR1

Recently savagegeese uploaded a track test of GTD, GT3 RS and ZR1 on Autobahn Club South, where the Corvette beat the Porsche by 1.7s.
It's yet another American track result that is significantly faster than simulation. I don't understand how ZR1 keeps overperforming on NA tracks, but underperforming on Nordschleife. The car drives as if it has ~550 kg of downforce at 250 kph (instead of 308 kg that GM claims it has).

And I know that it doesn't have big underbody diffusers, but am I crazy to think that it looks like it has more downforce than what they say?

Somebody needs to scan the car and run a CFD analysis on it.

BTW, if I adjust the specs in OL to get better match with the US lap times, it tells me that it should be 10s faster on Nordschleife!


Finally, here's an alphabetical list of all cars and tracks:

Spoiler: CAR LIST (57)

ABT XGT
AMG GT Black Series
AMG One
Aston Martin Valkyrie
BAC Mono R
Bugatti Bolide
Bugatti Chiron
Bugatti Chiron Pur Sport
Bugatti Veyron SS
Caparo T1
Corvette Z06 (C7)
Corvette Z06 (C8)
Corvette ZR1 (C7)
Corvette ZR1 (C8)
Corvette ZR1X
Dallara Stradale
Dodge Viper ACR
Ferrari 296 GTB
Ferrari 296 Speciale
Ferrari 488 GT3
Ferrari 488 Pista
Ferrari F80
Ferrari LaFerrari
Ferrari SF90
Ferrari SF90 XX
Ford GT
Ford Mustang GTD
Hennessey Venom F5 Revo Evo
KTM GT-XR
KTM GT2
Koenigsegg Jesko Attack
Koenigsegg Saddair's Spear
Lamborghini Aventador SVJ
Lamborghini Huracan Performante
Lamborghini Huracan STO
Lamborghini Murcielago SV
Lamborghini Revuelto
McLaren 765LT
McLaren F1
McLaren P1
McLaren Senna
McLaren W1
Pagani Huayra BC
Pagani Huayra R Evo
Porsche 918 Spyder
Porsche Carrera GT
Porsche GT2 RS
Porsche GT2 RS MR
Porsche GT3 Cup
Porsche GT3 RS
Porsche GT3 RS Manthey
Porsche GT? RS 992
Porsche Turbo S (992.1)
Porsche Turbo S (992.2)
Praga Bohema
RML GT Hypercar
Xiaomi SU7 Ultra

Spoiler: TRACK LIST (68)

Algarve
Anglesey
Anglesey Coastal
Anneau du Rhin Competition (Sport Auto)
Assen
Autobahn South (savagegeese)
Autocar dry / MIRA
Bahrain
Bilsterberg
Brands Hatch
Brno
Bugatti
COTA
Cadwell Park
Catalunya 2007
Donington
Dunnville (Throttle House)
Dunsfold
Fiorano
Goodwood
Goodwood FoS
Gotland Ring North
Hedge Hollow outer
Hockenheimring
Imola
Laguna Seca
Lausitzring Short
Le Mans 2007
Magny Cours
Magny Cours Club
Mendig AFB (AMuS)
Misano
Monza
Mortefontaine (Automoto)
Mosport
Most
Mount Panorama
Mugello
Nordschleife 20.8 km
Nürburgring 24h
Nürburgring GP
Paul Ricard 1A V2
Paul Ricard 1C V2
Red Bull Ring
Road America
Road Atlanta
Rudskogen
Sachsenring
Salzburgring
Sebring
Sepang
Silverstone GP
Slovakia Ring
Snetterton 300
Sonoma Long
Spa Francorchamps
Suzuka
Sydney MP
Thruxton
Tsukuba
VIR
Vairano Handling
Vairano Pro
Velo Città short (Acelerados)
Watkins Glen
Willow Springs
Zandvoort
Zhejiang

 

[Matski]

The car drives as if it has ~550 kg of downforce at 250 kph (instead of 308 kg that GM claims it has).

I thought I'd seen GM claim 1200lbs, which would be 544kg.

 

[Revvd]

I got this from Road & Track Volume 30:

978 lb at 186 mph (443.6 kg at 300 kph) = ~308 kg at 250 kph

You're right that those 1,200 lb gave me an idea of what to set the simulated car to, in order to get matching lap times. People misquote downforce so often, I could imagine this being the case here.
And this does make the rest of ZR1 lap times on US tracks match within 1s.

What's funny is, even with corrected downforce, it still shows ZR1 overperforming on Autobahn South:

	savagegeese	simulated
GTD	1:25.397	1:25
GT3 RS	1:24.174	1:24
ZR1	1:22.443	1:24

This Britt Casey guy really killed that lap.

 

[Revvd]

Remember this old video, where Salomondrin & Co. went to Thermal Club and set lap times with Veyron SS, Huayra, P1, 918 and LaFerrari?

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

Accept third party cookies

I added the South and Twin Palms Thermal Club layouts in OptimumLap, and here are the simulated results (+ extra 488 GT3 driven by Benjamin Sloss):

Decent results... Other than the P1, where they used P Zero Corsa while my simulated P1 has Trofeo R.

And the track correction factor is 1x - it's always nice when you get the driving line just right!
Here's how that looks btw:

I also added some functionality to the database.

First is a head-to-head view, where you can directly compare two cars, including difference in lap time per 1 km of each track in cell charts:

(mind you, the acceleration results at the bottom can never be accurate due to lack of launch control feature in the software)

And there's also a checklist form to select a combination of different cars and tracks for a dedicated table and chart:

Google Sheets link