Sneak Peek: Tire Physics Development

What fillman did give me personally no real hope, but rather showed me again how much of a knowledge gab there is in this community regarding anything tires/physics. Increasing the fidelity of a tire surface and stabilizing the new convex will surely change the driving feel yes, but it doesnt bring it any closer to realism. If that be the solution we had it in the game, but the already present abstract tire model, while sure bound by its technical limits, is closer to reality than most think. Im actually impressed by how it performs with how "nodeless" the tires actually are.

The title "I made the most realistic tyres ever" is also to oversaturated in my mind given that the mod was made on the pure basis of "lets try something and see what happens" instead of on the "i actually know what i do" basis. Just feels off calling it the most realistic yk, especially after looking into a copy of the mod. But sure thats just click bait right

 

I don't care what he called it) It's clear that it's still far from realism, but it just looks and most likely feels better. I don't know what kind of performance loss we're talking about, but the old tire can be kept for AI and trailers.

By the way, for some reason the introduction of real mirrors didn't stop developers due to the loss of productivity

 

Is the hamster ball a test of the new tire model? Its interaction with the surface, weights, etc.

 

bro...

 

I wonder if there will be a big racing update after we get the revamped tire physics, a bit like how 0.27 was a massive off-roading update. I really hope we'll get some older race cars, like 60s Le Mans cars and other closed-wheel race cars from that era. Recreating late 60s/early 70s Targa Florio with such realism would be a dream come true.

I know we'll have to wait a bit longer for tire physics, but I'm already so excited!

 

i could see a large racing theme update happening soon

 

I know they (can’t remember who specifically, maybe the career mode guy?) talked about how they wanted to shift towards motorsports, hence why we got rally content and such.

 

I hope we get weather and Dynamics content soon. And a temp slider

 

Well i hope you guys get your hands on some actual Pirelly DHD(2) or Michelin Pilot Sport GT or Goodyear Eagle F1 Supersport and not just some serial high performance like pilot cups or semislicks. Question 2 will compound tuning will change. Question 3 will there be any visual curve like stribeck or slip angle or dynamic debug data

 

If an architecture change ever happens, tire squaring would be great, as for me this is the most obvious use of the node system.

 

The main problem with Beamng is force feedback and the fact that due to the outdated tire model, you don't feel the car's skid point, and it happens suddenly, which doesn't happen in real life, because for a short moment beforehand, you can feel like your rear axle is about to skid. I speak as someone who has played 2,000 hours of AC and 2,000 hours of Beamng.

 

I have to ask at this point is it an experiment on buoyancy... Portrayed as a toy for the rest of us... I noticed you can pop them. I haven't tested whether altitude causes it to expand or not so...... I'm curious... I feel like I might put something heavy like a block in it and see what water pressure does to it.

 

Yeah, like if you have giant tires with enough air in them, will you be able to float like irl?

 

IR temperature pistol, zip ties, go pro, strap to strut or suspension links to get real time disc temps? Of course this is presuming that track day enthusiast has removed those shields from backside of the discs to get proper cooling air flow.

For me at least tire sound improvements are something I hope to come from this research, currently it is bit hit or miss, I can be sliding at 20% slip angle without audible feedback, which together without G-forces makes experience slightly less than what actual physics can provide. However going to DD wheel most certainly has compensated lack of audio feedback as with a toy wheel speed of FFB just was not there. With DD wheel judging limits of the grip is much easier to me at least.

Currently some stiffer suspensions and lower profile tires tend to have something I call flat spots in FFB, maybe that is realistic, but when I'm driving Scintilla for example, feeling of understeer vs neutral is quite subtle compared to Bolide for example, together without much audible feedback it often leads me to plow straight to the guardrails. Generally with older vehicles that has softer suspension and taller profile tires, I find it much easier to judge between understeer and neutral. Oversteer is much easier to tell apart, imho.

Not sure if research will result improvements to those areas, but of course I am hopeful.

Regarding ABS, currently if you are too fast to go from throttle to brake it tends to get into some bugged out state, similarly with traction control that will forget to check if there is change to conditions if you dance with pedals enough when little wheelspin. While there can be issues with ABS IRL, these issues in BeamNG currently don't behave similar to real world ABS and TC issues, but feel more like if code just jams to some loop, but that is not much related to tire model I guess, however I might not be the person who knows much of this stuff, still it will be interesting to see if improvements to tire model can lead to need to improve some aspects of how electronic driver aids behave.

Data from higher slip angles is difficult to get as temperature increase changes everything with the tire, do you deal with this by measuring how much change there is with heated up tire and then doing math or just try to be quick and fill tire with LN2? It always puzzles me, do tire models include friction changes from temp as well and then temperature effects on top of that or how on earth it would be possible to get high slip angle data without temperature effects already there, it must be very difficult to achieve temperature neutral data.

 

Perhaps a strange, silly question: Where did the money for something like this come from? Unless the testing time was... Well, free as an academic courtesy for lack of a better word, I would imagine it would be overall rather expensive to do something like this. I know talking about money is taboo for a dev studio, but I can't imagine that the volume of sales is high enough for major expenses such as this. Unless Beam.Tech is a paid product in which case they're probably rolling in it, since enterprise software is an incredibly lucrative business.

 

Interesting question curious to know either. To figure it out ourselfs we would have to figure out two things, 1) how much such a feature implementation would cost and 2) how this company could fund it.

1) What the full feature implementation would cost i can only speculate. We gotta consider that this is partially outsourced to a second company (Calspan) who provide the infrastructure and from the looks of it also personal. They dont publicly list pricing, likely because all they offer is tailored towards anyones needs. So we'd have to consider that, then inhouse research costs, dev costs, test costs and a bunch of other things. For a fully feature implementation i would very roughly estimate ~2M €.

2) You mentioned being unsure if BeamNG.tech is a commercial product. It is. They do also offer free licenses to academics but their primary target seem to be clients from the industry. Now they also dont enlist pricings, but ive worked with .tech users in terms of developing for them or to just assist them but i never asked for what their license costs unfortunately. However we dont have to know that, because as a German company BeamNG GmbH is required to file financial reports. And these can be accessed publicly.

According to the unternehmensregister they last filed a report in Feb. 2025 for the year 2023. You can find it here.
Now it has been over a decade since i was teached to read these. So please correct me if i get anything wrong. The company has ~12M in liquids, ~2.5M in Assets and had a net profit of ~4M. All in the year of 2023.

If we now look at the yearly balance history we can see that the company grew year by year until it stalled in 2023. So i figure lets just do a rough estimate by checking player numbers to tell if the game is still selling. Just looking at the steamdb it seems that player numbers are still growing. As such im just gonna say that the company is doing well and that the companies financials havent dipped since 2023.

So to answer your question. If the actual cost for this feature are in any range of my estimation then the company is well able to fund it by itself.

 

Neverless made some points on the matter but I’m just going to say this:

This is a software company first and foremost. The majority of expenses are employee salaries and benefits, and then the office space they rent in Germany. That’s one thing you can say about software is you’re not dealing with excessive commercial properties, expensive machinery, distribution etc. It’s all pretty straightforward with software development and online sales. And the majority of the company’s revenue is from BeamNG the game. It was probably a decent chunk of money to consult with Calspan but it’s an investment for the company. The game has been doing very well and I’m not surprised at all they have the means to do this testing

 

Oh great heavens. Perhaps I underestimated The Beam Team. I had assumed they were just playing backyard ball with standard indie studio funding. I suppose that would explain why they're taking their time making a true and proper simulator instead of rushing a simple video game to market.

 

Wondering if any devs could possibly chime in on these topics:

With the tires themselves going through a big overhaul, will there be a large change to the way tires are selected/types of tires? And will their intended uses or stats (like optimum temps or pressures) be more readily apparent for users? Say there’s a rally stage, and we have a combination of asphalt, gravel and dirt, what level of tire depth are we looking at? Will it be simply: just pick the rally tire or will more unique surfaces and ambient temps play a bigger role?