Thanks to the opportunity spearheaded by Ian Bell, this past Wednesday I was graciously invited as PRC’s foreign relations rep to a private press showing of Project CARS 2, which means we can start opening up a bit more about the upcoming simulator and tell you about everything that was on display at Bandai-Namco’s offices in Frankfurt. I arrived on location around noon and was greeted by two SMS employees, Luis and Marco. They quickly led me to the room where they had a rig setup with a Playseat and a Thurstmaster T300, complimented by a relatively large television running Windows 10.
Marco told me that before lunch would arrive, I could warm myself up with a couple of laps in the Olsberg GRC Lite car at the Daytona infield rallycross track. Straight out of the box the car behaved like I expected a support class rallycross car to behave. If I had to explain the feeling of the car to someone who hasn’t played Project Cars 2 yet, I’d say that it felt very similar to how the rallycross cars drive in Dirt Rally – though much more refined. On dirt or gravel, the car felt almost identical to Dirt Rally, but on tarmac you could really feel that Project Cars 2 is, at its heart, a racing simulator. It didn’t display the weird and wonky tarmac physics of Dirt Rally where the car seemingly has too much grip, but rather felt exactly like you would expect a car with dirt tires would behave on tarmac.
After I’d done a couple of warmup laps, it was time for a proper race against a field of AI opponents. If you don’t know much about Rallycross, then let me quickly explain to you how it works – six cars are spread across two rows, with the race commencing in a standing start fashion. Events are usually short affairs rarely lasting more than five or six laps, of which one requires you to take a slightly longer route around one of the corners of the track – known as the joker lap.
Since Marco was impressed by my speed he disregarded the AI strength setting recommended by Slightly Mad Studios for the press event and set the slider to 100% strength and increased their aggressiveness as well, a slider that many longtime rFactor players will be familiar with. I started the race in last place and had to make my way up the order during the course of five laps. I quickly overtook four of my five opponents, however the AI driving at the front of the pack was giving everything it could, throwing blocks, braking late and trying everything else at his disposal in order to keep me from taking his spot. However, my immersion was affected when I completely forgot to take the mandatory joker lap, and the game did not penalize me for it.
Later in the afternoon, we continued with a proper Honda Civic GRC entry at the Hell Rallycross track, a track many of you will be familiar with from Codemasters’ Dirt Rally. Though the track layout itself was the same as it is in Dirt Rally, I noticed something else now that I had a direct comparison with a different game that has the same track; Even though the layout was the same, the track surface detail was much higher in Project Cars 2. You really felt that you were driving on big chunks of gravel; The wheel vibrated ever so slightly, just like it does when you drive a gravel road in real life. Obviously, this feeling will change on the wheel and force feedback options that you set yourself, but that it’s working this well at this stage of the development of the game is a good sign.
After I’d accustomed myself to the car and track combination, it was time for another race. Since I had beaten the AI easily at 100% strength at the Daytona rallycross track, Marco now set the AI difficulty to 120%, fittingly called “Alien” – though I don’t consider myself one, neither figuratively nor literally. This time I had great difficulty keeping up with the AI since I had started in last place again, and due to a couple of mistakes from my side I didn’t manage to win the race.
Next up was driving the Acura NSX GT3 at the Red Bull Ring in Austria. For the first time, dynamic time changes and 60x time speed up were activated as well, which meant that during the couple of laps I drove to get used to the car and track, the time of day progressed from noon to deep into the night. Headlights turned themselves on automatically on the car and at the track, and due to both lower air and track temperatures you could actually feel the tires producing a little more grip than during day time. This is something the endurance racing guys will love.
During the couple of hotlaps that I completed to familiarize myself with the car, I noticed that the turn-in behavior of the car felt “weird”. It was a feeling that was difficult to describe, but compared to my own experiences playing the alpha version of the game at home, it didn’t feel as smooth as I was used to with these cars. We went into the gameplay options menu and saw that Stability Control was activated, presumably for the more mainstream gaming journalists who would arrive later. I don’t know exactly what it does (apart from making the car feel like shit), but I implore Slightly Mad Studios to explore how stability control has been implemented. Many pad users will undoubtedly toggle on a vast array of assists to come to terms with the driving experience in Project CARS 2, and I felt it was counter-intuitive to what driving assists are supposed to do.
So, after disabling the Stability Control option, it again was time for another race against the AI, this time against a full field of 31 other GT3 cars, similar to what you would experience when driving the Blancpain GT series in real life. Again, I started last and had to make my way up the order during the 10 laps I was given for this event. After I’d gotten into a bit of free air and could turn some laps without having to constantly battle for position, I noticed that the AI somehow had a lot more power coming out of the corners and thus they were able to accelerate much quicker. But even though they arrived with much more speed at the different corners of the track, they could brake five, sometimes even ten or fifteen meters later than I had to in order to make the corner. In turn, their actual cornering speed was somewhat slow; sometimes they were so slow that I bumped into them at the corner apex, even though at turn-in they had a good two car lengths on me. This is something that Slightly Mad Studios definitely have to put some work into until the release of the game, as right now they exhibit behavior traits similar to ISI-powered products, where the line the AI uses conflicts with what the player car is capable of.
What I enjoyed about the AI though, was that thanks to the high aggressiveness setting the AI often dared to try and out-brake me into corners when they were next to me, something which many games don’t do well. Sometimes this led to the AI using me as a brake in order to make the corner, but most of the time it worked out fine. I also didn’t notice any massive pile ups or cars stuck in the gravel pits, which plagues not only the original Project CARS, but also other racing simulators like Assetto Corsa. I eventually made my way up to third before all of the spaghetti fell out of my pocket coming out of turn five, and caused a massive pile up that involved at least five or six cars.
The final car and track combination that we visited was IndyCar at the Long Beach Street Circuit. There was no AI race event configured for this combination, but again we started at daylight and drove until the track was dark. I rarely drive single-seaters in racing simulations since most games get them pretty wrong based on my own real-world experience, but I was positively surprised by how it handled.
What you’re asking yourself the most at this point is probably: “Tell us how the cars drive and feel!”, so let me answer your question. The overall feel, as Austin described a few days ago is very similar to some of the very good cars in rFactor 2; for me that means that it feels very realistic and predictable, but I know there are people out there that really dislike rFactor 2’s physics. The tire behavior is very similar, flat-spotting is simulated at least to the same degree that rFactor 2 simulates it (albeit the force feedback doesn’t try to rip your arms off you once you flatspotted your tire) and PCARS2’s tire model also incorporates an advanced form of tire deformation, as you can see in the video below.
What really makes the biggest difference though in terms of feel compared to other simulators, is the way that the suspension is modeled and how it interacts with the force feedback and the overall feel of the car. As I mentioned with the Rallycross cars earlier, thanks to most of the tracks being laser-scanned, you can feel every little bump on the track. This makes the cars more exciting to drive but also easier to predict, as driving over a bump gives you immediate feedback as to what the car will do next. Obviously all of this is subject to change, but if Slightly Mad Studios keep going in this direction with the development of their physics model, then I’m very optimistic that it will turn out very well and generally be liked by the community.
Another thing that I got to experience was what the team are calling “Live Track 3.0” and how weather affected the track. Track and weather progression was accelerated by a factor of 60 during my time spent playing the press demo, so one lap with one car would feel like like two laps with a full grid of cars putting rubber down on the racing surface. After a couple of loops around the circuit you could not only see a driving line form, but you also felt that you could brake later and accelerate earlier where the rubber was laid down. I also have to mention that unlike rFactor 2’s real road surface, which is a generic uniform strand of rubber, the racing groove begins as two distinct smaller rubber trails (one for each side of the vehicle’s tires) before eventually widening into a bigger, general patch. It looks objectively better than rFactor 2’s real road, and gives you a better idea of the lines people are taking through each corner. An additional feature that I sadly wasn’t allowed to take footage of yet (since the feature isn’t completely finished) is how the dynamic track changes with weather, but I’m sure you’ll see it at some point in the near future
Once I was done with all the car and track combinations that I was allowed to take footage of, we went back to Long Beach, but set the weather options in such a way that it progressed from heavy clouds to a heavy storm, and then back to blistering heat over the course of a couple of laps. Since the build version I was playing didn’t allow for pit stops or a change of tires at the start of a session, I could only use dry slick tires for the complete session with the Acura NSX GT3.
With every lap I turned, it began to rain harder and harder until I could barely go forward anymore due to aquaplaning. And when the weather changed to a scorching heat, a dry line quickly formed where I had driven the lap before while the rest of the track only slowly dried off. According to Slightly Mad Studios, the final version of the game will, on most tracks, have realistic drainage modeled into the track mesh, so once it stops raining water will realistically start to disappear from the driving surface, while taking much longer and in realistic places off the track, such as on grass. Though sim racers who run shorter races won’t be able to see how this plays into an event strategy, the bigger GT3 and prototype leagues will find this to be an extra challenge they’ll have to deal with.
The sound is, overall, pretty decent. Depending on the car it’s already very good, though some cars still experience issues. One thing I spotted during my session at Bandai-Namco was that when you drive under a bridge it sounded like someone is cutting a steel beam with an electric steel saw, though I assume this will be fixed.
In terms of visual fidelity and performance, I can’t comment that much from the event itself since the game didn’t run at full details in order to guarantee perfect performance. I can tell you from my personal experience though that the game looks absolutely beautiful in most cases on full detail as the graphics engine is, after all, an improved version of the one used for the original iteration. Optimization is significantly better than the first game, as during hotlapping on maximum details I never dip below 60fps, and that’s with a 970GTX at a 1440p resolution. This is of course still subject to change as not all graphical features have been implemented into the game yet, but what I just said should give you a rough ballpark figure of what to expect from the final game. Be aware though that the more AI or, when playing online, real players are driving in the same session as you, performance will most likely take a dip as your CPU will be stressed pretty heavily thanks to the complex physics engine that works underneath the cover of Project Cars 2.
All in all I found very few things wrong with the game in its current state (or rather, the build that I played), and if things progress as nicely as they have done so far, I see very little reason to not enjoy this game once it is released. I honestly tried to find as many bad things as I could during the couple of hours that I had to play this game so I couldn’t be accused of shilling for the game, but there just wasn’t that many things wrong with the game. I also did not receive any money or other incentives to write this article other than paid travel to and from the Bandai-Namco Germany headquarters.
[Disclaimer: The poor video quality is down to Austin not having mastered the art of Sony Vegas video rendering yet, since we were allowed to recapture the footage taken at Bandai Namco at home. At the event I was able to play the game at a 4K resolution.]