Automation vs BeamNG performance - Ellisbury overcompensated?

[Raufbold]

It's no secret that with the previous 1.25 tire grip coefficient, the cars exported to BeamNG were quite OP. Taking Automation test track as a benchmark, even I (being a mediocre driver, and that's me being generous) could almost match track time in BeamNG to that of the Automation (and sometime even go faster).
With the new tire exporter however, it feels like there has been an overcompensation.
I have two cars - one exported pre-Ellisbury, and one after (same body, same fixtures, just slightly different engine and suspension tunings).
The 1/4 mile times on both are comparable to those in Automation, but the Automation test track performances vary drastically.
The pre-Ellisbury one has 2:08 in Automation and in BeamNG I got the 2:13 time, so mere 5 seconds apart. The Ellisbury one however had 2:06 in Automation, but I need to try extremely hard to make the 2:20 time in BeamNG now (over 15 seconds difference).
And it's not just my driving ability - the pre-Ellisbury car 0-100 is 3.4s in Automation and same 3.4s in BeamNG, but the Ellisbury one 0-100 is 3.3s in Automation, but 3.7s in BeamNG.
I noticed that grip is different on different maps though - for example braking distance varies quite a bit between the Gridmap, the Automation test track and the USA West Coast - with the West Coast drag strip being the most grippy and actually matching the Automation braking distances. So maybe that's also a factor?
What have your experiences been so far?

[Slim Jim]

I've done extensive testing with replica cars and the tires are much more realistic in the way they behave. The West Coast drag strip has a friction coefficient equal to that of a VHT-prepped surface so you can't base your acceleration or braking figures on that surface if you're looking to compare to real life magazine test numbers. Practically every automotive publication tests vehicles on normal street surfaces. Also, asphalt and concrete have slightly different friction coefficients as well, so you're right to notice these differences on different maps.

Try different tire pressures as the default ones might work well for most situations, but for specific performance testing absolutely need to be altered. (Go lower for 0-60 and braking, higher for top speed.)

Also, the lap times in the game are assumed to be a professional driver, so differences between the game's time and the real time are to be expected. You'd be surprised how small mistakes can lead to huge time deficits when you're racing around a track.

To put things into perspective as to how unrealistic the tires were previously, I used to de-tune replica cars from their real life counterparts in order to match acceleration and lap times that I found to be more realistic. Now with the new more accurate tire model, no more nerfing is required. :)

It will take some getting used to, but I can assure you the tires are huge step in the right direction in terms of realism.

And while you're at it, feel free to share the ".car" file. I'd like to see how close I can get to those performance targets.

(I should qualify the above statements to reflect that these were replicas built in Patch #7. I'm not sure if there were any changes for the recently released Patch #8).

[Raufbold]

Thanks for the feedback. Sure - try your best: https://drive.google.com/file/d/19daN0W1U0igq0Ed1i2NrIKql3zPd5aYC/view?usp=sharing

So far this is the only car I have modified for the Ellisbury update. It might not be optimal in every department, but I also tried very hard to make it realistic for the Automation world - i.e. be somewhat competitive on the market.

[Slim Jim]

Unfortunately, I can't export it without downloading a bunch of mods. Can you try to make a similar car with just the plain body (no mods)?

[Raufbold]

yep - here it is: https://drive.google.com/file/d/1KfAK_J7vPFhe_g1ZtE-HBWf2z1dWCeg0/view?usp=sharing

[Slim Jim]

Achieved 0-60 in 3.4 seconds repeatedly with launch control set to 2,200 rpm and rear tire pressure dropped to between 16-18 psi on a non-prepped surface.

https://drive.google.com/file/d/16uBbVyStfNijiowPlKYbPwnp_u8CQmi0/view?usp=sharing

Managed a 3.2 second 0-60 run at the drag strip with the same settings. This is all exactly what I would expect from a car with these specifications in real life. If you were getting 0-60 in 3.7 you were launching with too much wheelspin. You can either modulate the throttle manually or adjust the launch control to settings that are appropriate for the surface you're driving on.

Going to try the Automation Test Track next and see what's what.

edit: After 3 laps, I managed a 2:20 and ordinarily I would attempt to go further, but the car is basically uncontrollable at the limit. It may sell well in campaign, but I'm guessing not in any of the track based demographics. LOL With that being said, I think the car has potential. I'm going to make a few alterations and send back the modified ".car" file. Try and it let me know what you think.

[Raufbold]

I drive with my keyboard right now, so throttle modulation is not very easy. I didn't try launch control and tire pressure changes though - was using the car as is.
A bit surprised you couldn't beat the 2:20, because I feel like I really made many minor mistakes and could improve with more practice. The pre-Ellisbury this was a really fast and greatly handling car, but now there's more high-speed understeer and braking distance has gone up quite a bit. It used to have a different suspension config though - with less camber, but 190/265 wheels. After Ellisbury I increased the front wheel width for more grip and reduced oversteer by camber adjustments.
But sure, please let me know what you feel is wrong here and could be improved.

[Raufbold]

oh, and I just realized that by removing fixtures I also removed the wing, so you were basically driving with no aero. No wonder the car behaved weird.
Here's the model with the wing: https://drive.google.com/file/d/1-8E8EKq9UebuVG_pRa8tHAoHtkn2IBqM/view?usp=sharing

[Slim Jim]

I modified your original file, because I didn't see the updated file you sent until now. Here's a screenshot of my best of 3 lap times (2:11 vs. 2:04 in automation): https://drive.google.com/file/d/1fjgWL6GHoLgwS5tT6qbcnIgO6YAvQjfs/view?usp=sharing

The car is much more manageable at the limit. The only issue I found is a bug that is causing the hubs to be too stiff on export and immediately break. You can fix this by going to the Vehicle Config in BeamNG and lowering the hub stiffness front and rear from 100% -> 80%. After that, give it try and let me know what you think. (edit: I reported the bug on discord.)

Download here: https://drive.google.com/file/d/1ukZ5Sq3VYoymr7usEOy8WtZaTOGAdjSG/view?usp=sharing

Originally posted by Raufbold:

But sure, please let me know what you feel is wrong here and could be improved.

* The gearing was too short (numerically high), so I modified it
* The engine was too peaky, so I reduced the cam profile to give it a bit more area under the curve
* The suspension setup was reasonable, but given the car is very rear-weight biased, I altered the flat-ride ratio to about 20%.
* In order to gain more grip, I raised the ride height and altered the sway bars (more weight transfer)
* The ratio between front/rear tire width was too small. So I increased the width slightly at the front and significantly at the rear.
* The brakes were overpowering the tires, so I modified them to be closer to the limit of grip.
* Added aero (I figured there were aero fixture mods you were using that got deleted, so I added vanilla aero fixtures)

FYI: Set the launch control to about 2,700 rpm (+/- depending on the surface you're driving on) and your 0-60 should be in line with Automation's with no alterations (other than fixing the wheel hub stiffness = 80%)

[Raufbold]

yep, that's one hell of a car you have designed - I managed to get 2:17 from the second attempt. Need to see if I can make a version of this with quality sliders being lower - try to keep cost around $50K. And maybe make this one a more high-end version. Thanks!
Interesting bug with the wheels though - wonder what's causing it.

[Slim Jim]

Glad you like! You can use less exotic materials in the engine internals to save on cost - the engine is extremely "overbuilt" for the application. Also, I have +5 techpool across the board in the car/engine designer, and since I'm guessing you exported yours from a campaign your tech pool (and therefore costs) are going to be different. There shouldn't be a noticeable difference in performance with lower quality sliders as the principles of the redesign still hold true. Enjoy!

[Raufbold]

I'm guessing one big contributing factor is those gynormous rear tires ;)
At this point though what we end up doing is trying to bring back that 1.25 grip - just not by default, but by increasing tire width and balancing the suspension. If you think about it though - Lambo Diablo with a similar perfomance (except for top speed) has 245/335, so probably not too big after all

[Slim Jim]

Correct! The tires are modeled after the ones on the (992) Porsche 911 GT3, but in sports compound rather than the semi-slicks it actually comes with.

[Raufbold]

This is what it actually looks like with all the fixtures:
https://drive.google.com/file/d/12MDpnNwfD6WltO8weY2E1wwLa5iWI0ZG/view?usp=sharing
https://drive.google.com/file/d/1ScP2oiDm2dAo86Q_64IzFtjvYoflwzM9/view?usp=sharing

P.S. it actually very similar engineering-wise to the GT3... didn't notice that. It looks like you even modelled the gear ratios to be similar to GT3's :) It has a 100-0 braking distance of just 28m though.... how did they pull it off? Or is it because of semi-slicks being factory spec?

[Slim Jim]

Haha! Never hurts to emulate the best in the business! The gearing is indeed the same as the GT3's and you're correct in assuming the shorter braking distances are due to the semi-slicks being factory spec.

Car looks good with the fixtures on, thanks for sharing!

[Raufbold]

how did you manage to pull off having wider tire allowance? I can only get to max 285 width on rear and 295 on the front..