Among that game mode’s defining features: A translucent “ghost car” that represented the current leader in a given trial, allowing other drivers to chase after it or, ideally, pass it on the way to a faster time. Similar features have dotted various other racing video games through the years, from the Gran Turismo series to Forza and F-Zero. And now, we’ve begun to see this Mario Kart feature show up on television broadcasts for real auto races. In the same way it added intrigue to the video game, it brings real excitement to the process of watching a single car drive around a track solo. But it’s not just for fans; this technology is even making an impact on the inner workings of the sport of auto racing itself. Auto racing broadcasters have long struggled to make qualifying trials entertaining. Qualifying involves just a single car on the track at one time, racing against the clock for the best times to determine their starting order in the event’s actual race. “Qualifying, candidly, is boring,” says Gerard J. Hall, founder and CEO of SportsMedia Technology, known in the industry as SMT. “From a TV standpoint, you’ve got a camera taking a picture of this car, it’s in the center of the frame, and you’ve got one announcer saying to the other announcer, ‘Sure looks like he’s going fast, doesn’t it?’ “There’s not really much to compare it to … Until you get something relative to that car to let you know, is it winning? Losing? Is it gaining for the qualifying lap?” Hall and his team have solved this problem; auto racing viewers can now see a visual indicator of the leading qualifying car’s positioning during its qualifying lap, helping them instantly discern whether the current driver is improving on or falling behind the leader’s time. But the ghost car has traveled a long road to find its way onto these broadcasts. Hall and SMT have been involved in auto racing broadcasts and displays for three decades. While working with ESPN Motorsports in the mid-1990s, the company helped pioneer a small five-car leaderboard in the upper left of the screen for full-field race broadcasts. Initially, it was powered by a spotter sitting in the press box with binoculars and furiously typing updates into a laptop. The team’s methods have become more streamlined and high-tech since. Simple picture overlay tools allowed for varying leaderboards on racing broadcasts through the 2000s. These quickly became as synonymous with car racing as score bugs for baseball or football, growing more advanced with time. By the mid-2010s, satellite-based GPS had become advanced enough to consider its use for tracking high-speed race cars with solid accuracy. Hall and SMT began to look at ways to spice up Nascar’s broadcasts of its qualifying runs, which were still relatively bland outside minor additional features like car speed or time split graphics. Nascar offered a few helpful advantages in these early efforts. For one, all its tracks are dimensionally similar or identical: Sloped ovals without any sharp turns. There aren’t many tall buildings or structures near the track that might interfere with GPS signals. And because Nascar vehicles are fairly heavy, adding GPS sensors didn’t cause any balance or speed issues. Even within that context, though, tracking objects moving so quickly is no small feat. SMT developed a proprietary tracking device it calls a “vector” (or “vector box”) to sit on each car. Today’s vector boxes, which have been upgraded multiple times since the original version, combine three tracking formats. The first is GPS, but this is not the same GPS your Google Maps app is using. Vector boxes receive signals not just from the American Global Positioning System but also from Russia’s GLONASS, Europe’s Galileo, and China’s BeiDuo satellites. Hall says the combination allows vector boxes to be connected to anywhere from two to three times as many satellites at one time as standard GPS setups. Inertial measurement unit (IMU) tracking is also used to gauge precise speed and related movement data. Combined with the bulked-up GPS setup, Hall says, SMT can access real-time kinetic metrics and achieve centimeter-level accuracy for both positioning (horizontal and vertical) and car velocity to gauge precisely where the vehicle is on the track at any time. Real-time drive data from the car’s engine control unit for things like throttle, braking, steering, and gear info are also fed into the vector box. The vector can even be battery powered so that it draws no energy from the car itself. Using all of this detail from the vector box, SMT created what Hall believes is the first iteration of the live “digital twin” concept in any sport: An entire 3D re-creation of the cars and track. “Here’s a sporting event I can look at with cameras, and here’s the digital twin that I’ve created this 3D model with a computer, where everything is moving around based on the live data coming off the objects in the real world,” Hall says. “It gives me the ability to do what we call temporal shifting, and that’s really what the ghost car is.” The first Nascar broadcast featuring these tools, which SMT collectively terms its Broadcast Analytics suite, ran on NBC in 2018. The same suite has been used by Fox in each season since. The ghost car has featured prominently as a live companion to qualifying, allowing broadcasters to swap into a live 3D world and show accurate digital renderings of both the active car and the ghost car it’s competing against. The Race Is On Formula 1’s broadcast team has also been playing around with various ways of creating a qualifying ghost car for over a decade. Its first attempts involved retrofitting the tools used to analyze golf swings. But as the F1 organization also moved to GPS-based methods, logistical challenges arose. Unlike Nascar's clean oval of a racetrack, F1 tracks in different locations differ wildly from one another. Plus they are often built within cities, so each course is dotted with tall buildings and crowded with wireless communications that can interfere with GPS signals. “GPS is good in places, but we go to places like Monaco, Baku, Singapore, where all that infrastructure and buildings really make GPS drift,” says Dean Locke, director of broadcast media and digital for Formula 1. F1 cars are also much lighter than their Nascar counterparts, necessitating lower-weight sensors. Its broadcast teams struggled to put together full laps of ghost car or related overlay features as a result. To get around these limitations, F1 built several internal prediction models based on past seasons’ GPS data and lidar scans of each track. It then integrated those models into an in-house application that overlayed a synchronized video feed of those same past qualifying trials, then compared the two data inputs side by side. By first manually identifying discrepancies where GPS data from qualifying clearly was out of whack based on video footage, the team was able to train its models to spot anomalies and correct them. F1’s own ghost car feature launched across its broadcasters for the 2025 season. It features both an in-car driver view and a “helicopter” view above the car—especially helpful for qualifying laps where the current driver is ahead of the ghost car, since the ghost car would be invisible during in-car views in these situations. A drawback, though: F1 can’t yet show its ghost car on live broadcasts. Because the team’s process still requires that final bit of manual fine-tuning to ensure the translucent ghost car is positioned accuracy, Locke says it takes around 90 minutes to pass the ghost car overlay to broadcast partners after each qualifying run. The team’s goal is a 30-minute turnaround to facilitate quicker post-qualifying analysis. Still, Locke says, ghost car broadcasts have been immensely popular on F1’s social channels. F1’s single-car competitors at IndyCar, meanwhile, have pushed the envelope further this season as Fox, the home of Nascar, has taken over from NBC as the IndyCar series’ broadcaster, bringing SMT’s vector box technology with it. Starting with Indy 500 qualifying earlier in May, a fully live ghost car animation has been used as a superimposed overlay for Fox’s in-car driver camera. This isn’t a 3D rendering, like SMT’s prior Nascar ghost cars, which swap into a separate digital view (either the full screen or smaller box format); it’s the actual, real-life driver camera from the broadcast with a centimeter-accurate ghost car overlayed onto the screen. Laps of Judgment Ghost cars are still in their relative infancy, especially for truly live broadcasts. Hall sees the uses of this underlying technology quickly expanding as broadcasters become more familiar with them—including for full-field races in addition to qualifiers. “Maybe you want to compare tire wear,” Hall posits. “Let me show you this car now running live, and let me show you this car five laps ago where it was running on different tires.” Hall also suggests the idea of multiple ghost cars for qualifying; one for the pole sitter, but perhaps others for bubble positions or other relevant spots. It's easy to see how the ghost cars zipping around on your TV could just be the beginning of the way this ultra-specific car data will be used. Just look at Nascar; SMT received interest from multiple teams within less than a year of pioneering its vector box. The teams wanted to use data from the box for tactical purposes like postrace analysis or comparisons to competitors, and SMT developed a Team Analytics application. Demand was so high that in 2018, Nascar opened this data up to allow all teams full access. They can view everything from car positioning to gear and throttle data, not just for their own drivers but for the whole field. Team Analytics has a feature that allows multiple ghost car overlays of prior races, allowing crews to compare driving lines or turn speeds. “Now at every pit stall, it is the center application,” Hall says. “Because the teams, the crew chiefs, they want to know: How is my car doing relative to other cars? And because it’s not just positioning on the track, it’s also gears, throttle, et cetera, different drivers can compare their performance against” each other. It's likely only a matter of time before these features trickle into IndyCar and F1 teams. Locke says those teams “will use absolutely anything they can to get an improvement over another team,” though he notes that F1 teams already collect a wealth of internal data on their vehicles. Who’d have thought a funky Mario Kart feature used for bragging rights among friends could have such an influence in the world’s highest-stakes car races?