Ellisbury braking distance

[Raufbold]

There are a few comments in the bugs section about the unrealistically long braking distance after the Ellisbury update.
While I don't believe this discussion belongs into the bugs section, I still want to second that notion.
Here are my thoughts:
Braking distance is affected by many factors, such as brake size and type, vehicle mass, aerodynamics, suspension settings, ABS but most of all tyres - width and compound.
So I might not agree with players who only play with the brakes and complain about braking distance not being optimal.
That being said, it still is quite abysmal.
A typical mid-sized modern car has a 100-0 kph braking distance of 34-40 m (modern Skoda Fabia 1.2 claims to have 34m???). 42 m if it's an SUV. All that while having 185-205 width standard tyres and weighing at around 1500 kg.
In BeamNG this is largely accurately reflected, with 1996 Ibishu Pessima having 205 width standard tyres, weighing 1420 kg and having a 100-0 braking distance of around 40 m.

Yet when I create a car that weighs 1270 kg, also has 205 width medium tyres with ABS and brakes optimized to give maximum possible braking distance, I end up with 47 m in Automation and 49 m in BeamNG. I was barely even able to get 32 m braking distance (in Automation! In BeamNG it's longer) in a 1400 kg modern sports car with aero, full ESC and 315 sports tyre width.

Maybe someone was able to get better stats, but this has been my experience so far and it feels a bit ridiculous.

[Slim Jim]

Built a replica of a 2024 Acura Integra Type-S and the braking distances are in line with reality (attached the .car file for reference). If you're having trouble matching your expectations in the game with reality, I find the most common issue to be people's reluctance to use quality sliders. Case in point: the Acura in question rides on 265/30R-19 Michelin Pilot Sport 4S tires and has very expensive alloy wheels. The wheels and tires alone are about $4,000. 0 quality in Automation has material costs of about $1,300. Logic tells me if my 0 quality sports tires/wheels aren't generating the braking distances I expect to see in real life... crank up the quality sliders until they do.

https://drive.google.com/file/d/1kB1aWwU1_MBpDesicUij6TV8oDjwzhnb/view?usp=sharing

[Raufbold]

Jim buddy :)
You're probably correct, but then the calculations of cost are inflated (again, maybe it's the tech pool that is supposed to be really high). Anyhow - my car in question, a 1987 sedan at 1270 kg actually had even larger tyres - 210 width. Since my thoughts were also around quality, I increased wheel quality by 5 and that resulted in braking distance going down from 47 m to 46 m... It also doesn't have any aero, but come on

P.S. +14 quality on wheels? Wow. Also, doesn't it look weird that this is pretty much the only part (except for bottom end and a bit on the interior) where you had to add quality? Basically, if quality (or tech pool) was overall consistent, either the tyres would be trash or the cars would be extremely expensive... It's probably this consistency issue that is confusing me more than anything else. After all, it's easy to make cool cars with max quality all around.

[Slim Jim]

Hah! :) I know it sounds high, but my goal there was to match the actual wheel and tire cost with the material costs in game. There's a reason why the Type-S is nearly $20,000 more expensive than the base Integra. Not to mention, legacy car makers have had decades to build up their "tech pool" or real-life equivalent. So their actual material costs are probably way less than what you see in my replica.

As far as my fiddling with the quality sliders elsewhere, that was primarily to get the curb weight to match the real car. Not necessarily indicative of any inconsistencies that I'm aware of. For example, I've had to add downforce and use a bit of negative aero quality to get the top speed to match a different replica's real-life equivalent because the base body in automation had too little drag.

I'm happy to build another (1980s) replica where I can find period data for it. It's possible there's an issue with stat progression - that is, the overall grip of the tires as a function of their model year unfairly penalizes older cars vs newer cars. Definitely worth investigating for sure!

[[CAMSO] MrChips]

We're always updating balance for stuff like this; there might be something worth looking at here.

[Killrob]

That is interesting! This would point to longitudinal grip being a little too low, something that easily can be tweaked. The "intended" quality setting is, depending on the category, something like +2 to +5, higher if highly specialized. If braking distances are too long, then what would make sense is to check the longitudinal grip stat of the selected tires. What you also can do is to swap to a higher grip compound to see how much extra you need to get the "correct" braking distance. It could also be down to the loading factor of the tire being off, giving you too little grip when fully loaded up during weight transfer.

Replica builds are really handy to figure out how much tweaking needs to be done :)
Or maybe braking grip in general just needs a little buff.
Certainly something to investigate!
Cheers

[markgy90]

Killrob, please read this!
I'm the one who complained in the update bug section about braking distances. I have made an another experiment about this.
I have tried to recreate the facelift of the second generation Lexus LS400 (produced between 1997-2000).
I have set the tech pool +8 for everything (because Toyota), the year is 1992 (try to simulate the usual 5-6 years of engineering). I have used real life size brakes and tyres. I have raised the quality for the tyres and brakes +8 to match the tech pool. Yet the braking distance still bad. 46.6 m in Automation, 44-45 m in BeamNG.
In real life the 1997 Lexus LS400 can stop 60 mph to 0 in 116 feet which is 35.36 m, that means 100 km/h to 0 is around 37-38 m.
Here is the link for the test:
https://www.motortrend.com/reviews/1998-lexus-ls-400/

Real life 1997-2000 Lexus LS 400:
1765 kg
216 kW at 6000 rpm
407 Nm at 4000 rpm
225/60 VR 16 676 mm diameter
Front brakes: 315 mm vented disc
Rear brakes: 307 mm vented disc
60 mph - 0: 116 ft (35.36 m)

My recreation:
1696 kg (1706 was measured in BeamNG)
216 kW at 5900
395 Nm at 4200
225/60/16 Radial medium compound 675 mm diameter tyres +8 quality
315 mm 4 caliper vented disc front brakes, 305 mm 1 caliper vented disc rear brakes +8 quality
Has ABS (and works in BeamNG)
100 km/h - 0: 46.6 m in Automation, 44-45 m in BeamNG

Also the grip is low (probably playing role in the bad braking distances). Despite using the real world gear ratios the tyres are always spinning in first gear, ruining 0-100 km/h acceleration.
Here are the real world gear ratios:
Final drive ratio: 3.266
I. 3.357
II. 2.18
III. 1.424
IV. 1
V. 0.753
I have set the gear ratios as close as possible to the real world gear ratios in my recreation. Yet the car can't get a grip in 1st gear.

[Slim Jim]

Car and Driver has a more accurate testing procedure for braking. I'm working on a series of 20 replicas across different cars and different years and will be posting the results on Discord once complete. According to Car & Driver (combined with my conversion factor to match speed to distance), the LS400 should do about 41.3 m 100kph-0.

[robert.norgren]

How much difference would it be with one step higher compound tires?
Test all the compounds, with the same settings otherwise. Only change the compound.
Will this give different results? And how much?

[Raufbold]

I had to increase the front camber from 0.3 to 0.5, which resulted in braking distance going from 47.7m to 48m. Changing tyres from medium to sport reduced it to 42.4m

P.S. it is a 1987 2.5m RWD sedan with front longtitudinal engine placement. Wheels are 210/65 R16 medium compound with +2 quality (from tech pool). Weight is 1279 kg 59%/41% front/rear distribution, no aero. Front camber 0.5, front toe 0.15, rear toe 0.25.
Brakes are solid disk - Front: 2-piston 275mm, 121% force; Rear: 2-piston 190mm, 82% force, normal 50 pads, quality +2 (from tech pool).

[Slim Jim]

I completed my comprehensive analysis on braking here: https://discord.com/channels/250008858062225408/1192984456286458037

Feel free to check it out. It contains 20 replicas spanning various decades and tire types, as well as a spreadsheet summarizing each car.

[Raufbold]

so it looks like your point is - not all cars actually have this great braking distance and that we shouldn't really compare 80s cars to modern ones. Basically, if a '96 F355 Spider with sports tyres has a braking distance of 38-39 m, why are you complaining about your 1987 sedan with mediums having 48 m?
Sounds fairly reasonable. I still have my doubts, but I guess I need to look deeper into the cars you made to figure out what I might be missing.
Also om a side note - I like those gearing curves on all cars. Did you copy the ratios from the real cars or did you make those yourself? If those are of your own making, what is the rule of thumb to create such gearing?

[Slim Jim]

Indeed, that is my overall impression. Some publications like to post "hero runs" that aren't really replicable in the real world - but makes them look like "better drivers" or makes the manufacturer look good. Looking at Car & Driver's test methodology I came to the conclusion that their braking tests were overall more representative of what the cars would do in most situations.

The gearing was copied from the real cars where I could easily find the data. If there were any exceptions (or mistakes), I did my best to guess. I think I accidentally gave the McLaren a crossplane crank for instance because I forgot to click the button to switch to flatplane. Little mistakes like that might be present throughout. But overall, I did try to stay true to the real life counterparts especially with respect to curb weight (weight distribution might be off as well on some vehicles), brakes, and tires.

edit: CarWow (YouTube) did a few braking tests that I think we can add to the benchmarks for the giggles. A Bugatti Chiron did 100 km/h to 0 in 33 m. A Lamborghini Urus did 70 mph to 0 in 46 m, which according to my conversion factor equals 36.2 m 100-0. A Mercedes-AMG CLA45 S did 70 mph-0 in 48 m or 37.8 m 100-0 according to my conversion factor. The numbers seem to match my expectations for high performance cars on semi-slick (Bugatti) or sports tires (AMG and Lambo) on a modern car.

[ssgtbaloo]

Since the Ellisbury update went public, I've noticed longer braking times and I've also noticed that it now takes a little effort to induce brake fade. Even very heavy/large cars can be brought under control with a little effort, although, apart from fade, their braking performance is pretty awful compared to before. Are these symptoms related to each other?

[Slim Jim]

OK - let's unpack this.

Longer braking times as compared to... the previous iteration of the game where the tires were far too grippy to be realistic? Yes! Compared to real-life equivalents? They're on par and it's not too difficult to match.

If you find its "easy" to eliminate brake fade then that can probably be explained by 1 of 2 things. Either the pad types real life manufacturers are using are much lower than what most people are used to selecting in game. Or the brakes thermal efficiency need to be nerfed in the same way grip needed to be. (I'm going to guess it's the former)

Pad type is more difficult to compare with real life equivalents because it's a numerical value and not a material type. I'm "guessing" 0 - organic, 25 - ceramic, 50 - semi-metallic (less metal), 75 - semi-metallic (more metal), 100 - fully metallic race pads would be a good rule of thumb.

[Raufbold]

ok, so here's a nail in the coffin of the Ellisbury braking distance question. I did the most reasonable thing and replaced my 210/65 R16 Automation Medium tyres for 205/55 R16 BeamNG Standard tyres in the exported 1987 sedan I was talking about (I used 8-lug hubs with 16x7 Gavril 10-slot wheels). The results (after a few tests on Automation Test Track) are:

1. slightly better handling (seems like suspension setting still does most of the work)
2. 0-100 went from 8.8 to 8.6 (minor improvement) - both are still WAY better than 9.4 Automation value btw - not sure what's going on here, as normally 0-100 times were worse in BeamNG. Car weight is only a couple of kgs lower in BeamNG.
3. 102-0 BRAKING DISTANCE went from 50.2 m to 43.2 m

And that's on narrower tyres. So while BeamNG itself might not be the best representation of the actual real-world cars (maybe they decided to take the declared values and not the Car and Driver ones as their benchmark), the inconsistency is definitely there and it does look like at least in BeamNG we need a slight buff back for our tyres.

P.S. To be fair, the original car has +2 tyre quality (from tech pool) while BeamNG does not differentiate by quality at all. So maybe BeamNG just assumes tyres of maximum quality all around? I was able to approximate the BeamNG braking distance by equipping my sedan with 205/55 R16 Medium tyres with +17 quality (+15 and +2 tech pool) in Automation, then exported it to BeamNG. The results:

1. Lower mass (unintended consequence, but that's what increasing wheel quality does) - 20 kg lighter
2. MUCH better handling. This means that handling probably was originally captured pretty accurately for 'default quality' tyres.
3. 0-100 is around 8.6, so pretty much the same as BeamNG tyres
4. 102-0 braking distance is 43.8 m - again, same as the BeamNG tyres (but the car is 20 kg lighter)

It looks like the default +5 tech pool quality tyres are generally well balanced in terms of handling/acceleration, but are significantly behind in braking.