Devotees of racing games love to throw shade at each other. Xbox versus Playstation, console versus PC, controller versus wheel; you name it, people will argue about it on the Internet. And one of the more common ways to denigrate an opponent in such an argument is to play the purity card.
This inevitably involves some variation of “my game’s better than yours, because mine is a , and yours is just an arcade game.” The implication is that you aren’t hardcore enough because you play something fun and accessible.
It’s not an argument I buy into, but it is one I’ve thought about through the years. If being a faithful simulation is the be-all and end-all of it, then how do consumer games compare to the real thing? Not racing an actual car on an actual track—I answered that one years ago. No, I’m talking about the driver-in-the-loop (DIL) simulators used by professional racing teams—these proprietary setups that move and shake and carry price tags in the hundreds of thousands or even millions. It’s been a tricky question to answer. DIL sims are few and far between, and they tend to be in heavy use doing actual work.
As luck would have it, the nice people at Mazda North America didn’t laugh when I recently asked them if I could visit their sim. In fact, they invited me to see it in action as the team prepared for an upcoming race in the IMSA Weathertech Sportscar Championship. Two of their four drivers were new to the series this year, and they’d be spending a couple of days getting up to speed with a track they’d never been to before. Suddenly I had a chance to see what pro drivers and engineers actually got from spending time in a sim, and to gauge how the whole endeavor differs from even racing games. Plus, if I was really lucky, I might even get to have a go myself…
Testing, testing, one two three…
In racing, as with any other sport, practice makes perfect. Instead of practice, racers call it testing. Testing gives drivers extra miles behind the wheel to hone their craft and engineers extra hours to perfect car setups. But testing isn’t cheap—little in racing is, after all. That’s great when budgets are fat; the fatter the budget, the bigger the test program (and invariably, the better the test program, the better the race results). But those days of easy money are a thing of the past. The seemingly bottomless well of tobacco sponsorship money dried up a couple of decades ago, and time and again we’ve seen that economic downturns and corporate scandals can be the perfect antidote to an OEM’s desire to spend hundreds of millions a year on a motorsport program.
Across the racing world, series have been restricting the amount of permitted in-season testing in the name of budget sustainability. IMSA currently limits this to a maximum of 10 private test days plus a couple of official test events in January and February. That’s not a lot if you’re trying to turn a racecar from an also-ran into a title contender—which is where the DIL sim comes in.
The racing world started to get really serious about simulators in the mid-2000s, and this adoption started in the high-stakes world of Formula 1. Teams like Ferrari and McLaren were no longer allowed to operate their separate test teams, which until then would spend the year lapping tens of thousands of miles at race circuits while spending tens of millions of dollars in the process. Taking a leaf from the aerospace industry, those teams decided to spend that money on developing racing simulators instead.
Up until then, simulation wasn’t unheard of, but it was a tool just for the engineers. They’d input some numbers reflecting different suspension setups, and the software would spit out a theoretical lap times for each based on mathematical models that took into account vehicle dynamics, aerodynamics, and tire performance. But a DIL sim needs to do more since a real human driver provides the lap time by controlling the car in real-time. This meant adding a way for drivers to provide inputs (through a steering wheel and pedals) and receive outputs—graphics, audio, and motion.
The advantages are obvious. There’s no need to hire a track, nor book travel. The weather can’t ruin your plans, and changing a setup is the work of a few keystrokes. There’s no risk of even a skinned knuckle before your driver is back out. Of course, this all only works if you’re able to correlate performance in the sim with real life.
Fancy DIL sims soon became in Formula 1, but before long they were starting to show up in other well-funded racing programs. The one at the Canadian HQ of engineering company Multimatic has been in use since 2011 and is widely acknowledged as one of the best, particularly in sports car racing. (An evolution of this set up, complete with 3D graphics, is now in operation at Ford Performance in North Carolina as well.)
As you’ll note from the images, Multimatic’s sim doesn’t use the hexapod manipulator common to flight simulators or the earlier automotive DILs. While those work fine for aviation, they’re big and heavy and suffer from unavoidable mechanical lag. This isn’t a problem with pilots and flight sims, because the distances to visual cues in that environment are much greater. It’s also less of a problem with drivers of a less than expert skill level, but for the pros it causes motion sickness because the feedback from the inner ear doesn’t match what their eyes are seeing.
Instead, this DIL uses a stratiform arrangement developed by a company in the UK called Ansible Motion. Ansible discovered that giving a driver the right amount and variety of physiological feedback was more important than trying to slavishly recreate the actual motion of a car on a racetrack. This still features six degrees of freedom as the base moves laterally and longitudinally, plus it can rotate (with the cockpit itself also rotating about three axes). But this sim does so fast enough that all those sensory inputs agree with each other. (The other advantage to a stratiform machine over a hexapod is it takes up much less vertical space—see the images in this piece about Ford’s VIRTTEX sim for comparison.)
Multimatic uses physics models developed by its partner VI-grade, and these models contain a lot of optimization of things like tire models done in-house by Multimatic’s technical director of vehicle dynamics Peter Gibbons and lead vehicle dynamics engineer Lars Ogilvie. The graphics, which are projected onto a 160-degree 4m by 1.5m screen, are courtesy of VI-grade and SIM.CO.VR, and the tracks are all based on the same lidar scans that go into titles like or .
All of this runs on a number of workstations, including three fitted with current high-end gaming graphics cards to run the three projectors. (Apologies, I forgot to note the exact spec of those GPUs.) The graphics aren’t as flashy as you’d find on a big budget racing game, but they’re more than good enough for the task at hand. Audio is handle through a set of headphones; I’m sure this is a preferable arrangement to speakers for the engineers who work alongside the drivers when the sim is operating.
The cockpit of the sim is a sturdy spaceframe that replicates the driving position in the actual racecar. In addition to the motion provided by the sim itself, the shoulder belts of the six-point harness tighten under braking to replicate the sensation you get when braking hard on track in a real car. The steering wheel and pedals are basically the same as you’d find in the actual car, as is the Motec electronic bus. The motor that drives the steering is a step above even expensive consumer-grade wheels like those from Fanatec, and the brake pedal actually uses hydraulics (again, just like the real thing).
For the past few months, all this gear has been getting quite a work out as Mazda has been slogging away to make its IMSA prototype racer into a winner.