Racing simulators stopped being a hobbyist curiosity some time ago. Professional motorsport teams use them to learn circuits and test setups, and a parallel competitive scene has grown up around them where people train for the simulator itself rather than for the car it imitates. That shift turned a settings menu into a serious question. Modern direct-drive steering wheels can push back against your hands with real torque, and they can also buzz with a high-frequency rumble meant to stand in for the engine and the road surface. Racers argue endlessly about how much of each to run. Until recently, almost nobody had tested it properly.
The Experiment
A research group studying esports performance put 59 experienced sim racers in the same seat, in the same car, on the same track: a BMW M4 GT3 around Brands Hatch in iRacing. Each drove five-lap runs under four feedback conditions. Nothing at all. Force feedback only, the torque that resists and pulls at the wheel. Tactile rumble only. Both together. Participants were also split by intensity, one group on moderate settings and another on high ones.
The result was cleaner than the forum arguments would suggest. Lap times were fastest with moderate force feedback. Not with the strongest setting available, and not with the wheel switched off. With torque coming back through the rim at a middling level, racers made better corrections, held a more consistent pace, and reacted more quickly when the car started to let go.
The Rumble Did Not Make Anyone Faster
Here is the part that complicates the story. No level of tactile rumble improved lap times. Not on its own, not layered on top of force feedback. On the stopwatch it did nothing at all.
Racers loved it anyway. When asked to rank the conditions, participants consistently rated any setup with feedback above the bare one, regardless of whether it had actually helped them drive. Preference and performance came apart completely. The rumble was not information the drivers could act on. It was atmosphere, and they wanted it.
What Is Actually Moving in the Wheel
A companion study from the same lab went looking for the physical signal underneath all this. Researchers mounted an accelerometer on a direct-drive wheel, had skilled racers run the same car and track across nine combinations of force and tactile settings, and ran the recorded motion through a frequency analysis.
Two bands dominated. Everything between zero and five hertz turned out to be force feedback mixed with the driver's own steering input, which makes sense given that deliberate human movement mostly lives below ten hertz. Everything in the 25 to 30 hertz band was purely tactile: it was completely absent whenever the rumble was switched off, and it carried between 83 and 91 percent of the tactile channel's total energy. Nothing at all was detected above 110 hertz, and between 66 and 99 percent of the wheel's motion ran side to side rather than fore and aft or up and down.
Two findings stood out. Doubling the rumble setting from half to full did not double the signal, it roughly quadrupled it, which means the relationship most racers assume when they nudge a slider is wrong. And turning force feedback up measurably suppressed the 25 to 30 hertz band, suggesting the wheelbase quietly throttles one channel to keep the motor within its limits. Crank the torque and you are, without knowing it, turning the texture down.
Two Channels, Two Audiences in the Hand
The split makes more sense once you look at what the hand is built to detect. Rapidly adapting Meissner afferents in the skin are most sensitive between roughly 10 and 50 hertz and are specialised for sudden movement and slip, which puts that 25 to 30 hertz rumble squarely in their range. Pacinian corpuscles handle the higher band, from about 50 to 300 hertz, peaking around 200. And detection sharpens when the hand is gripping something, so a palm wrapped around a wheel rim is an unusually good receiver.
But force feedback is not really a skin signal. Torque is read by muscles and joints, through the proprioceptive sense that tells you where your limbs are and how hard they are working. That is the slower channel, and it is the one that moved the lap times. The skin got the texture. The muscles got the information.
The Bigger Idea
The useful lesson here is that "feel" is not one thing, and treating it as a single slider to be maximised gets you a worse result. Moderate force feedback beat the strongest setting available, which is a reminder that more intensity is not the same as more signal. And the fact that the rumble improved the experience without improving performance is not a failure of the rumble, it is a description of what it is for.
Haptic technology across gaming, sport, training and wellness keeps running into this same distinction. Some touch cues carry information the body can act on, and those should be judged on whether they change behaviour. Others carry presence, the sense of being somewhere real, and those should be judged on whether the experience holds you. Both are worth building. They just need to be designed, tuned and measured as the separate things they are.