By Federico Bertelli
Another Grand Slam is over. As customary, the days following the event is a good time to catch our breath.
Ubitennis brings to our readers an interview with Anthony Taylor, Head of Product at Tennis Data Innovation (TDI). TDI is a branch of the ATP Tour. In our exclusive Q&A we discuss what Tennis Data Innovation is and why it is becoming a key component for every player.
UBITENNIS: First of all, thank you for your time, Anthony. Let’s start with a brief overview about TDI and your role there.
AT: It’s a pleasure to discuss TDI’s role in the tennis world with Ubitennis. I am the head of product at TDI, a role that encompasses promotion and development of initiatives by TDI. TDI, a joint venture between the ATP Tour and ATP Media created in 2020, is responsible for collecting, managing & commercialising data & streaming across all ATP events from Challengers to Masters 1000s. As for data, we collect it from various sources, including chair umpire data and optical tracking (like Hawk-Eye data). We aggregate all available ATP-level data points and derive informative products for the market from them, operating in the realms of betting, media, and player performance.
UBITENNIS: You’ve given us a great overview of your operations. Speaking of data analysis, just to be clear: Can we say TDI is now the sole official entry point for ATP level data point registration?
AT: Yes, TDI is the single source of truth for ATP event data. We collect data point by point from the chair umpire, what we call “level 1” data. But we don’t stop there. We collect deeper data to create advanced, easily understandable metrics that can enhance fan engagement and be used for player analysis. This approach is already established in other sports, like the NBA or Premier League football. We use various data sources beyond chair umpire data, including optical detections like Hawkeye and other providers. These provide spatial tracking of ball movement, player positioning, and shot physics (speed, spin, net height, etc.). With this comprehensive data, we fuel a range of metrics like winners, unforced errors, shot distribution (forehand/backhand, rally length, net points won), and even game pattern analysis (offensive and defensive shots, counterattack exchanges).
We also focus on synthesizing these metrics into significant, easily interpreted numbers that reflect shot quality. We’re showcasing infographics on our X and Instagram accounts that do just this: offering comparative indices and ratings for shots on a scale of 0 to 10. The aim? To quantify how well a player is performing in a match compared to themselves, their opponent, and the tour average.
Another interesting metric is on-court attitude: is a player predominantly offensive or defensive? Here too, we offer a condensed representation based on speed, spin, and shot depth. And we assess how well a player exploits offensive situations or manages defense. These insights are distributed through ATP media, enabling ATP TV commentators to provide advanced insights during broadcasts. We also make these metrics available to players via ATP TennisIQ, our player performance portal.
UBITENNIS: You mentioned an analytics platform available to players, Tennis IQ. I’m curious to know more about what it offers to players and how they’re using it.
AT: We have staff at tournaments providing support and guidance on using the platform, and the feedback has been positive. For instance, data expert Mike James, who works with Holger Rune, has utilized it. The idea is to democratize data analysis access, which often involves high costs. Now, players can download raw Hawkeye data for their data analysis teams or directly use the metrics and insights we provide on Tennis IQ. This opens access to certain types of information for everyone. There are service providers who continue to offer more refined analyses, but in any case, we’ve freed access to ATP data for player performance purposes. Our next step is to link game footage tags to shots, allowing coaches to recall video of each shot type.
UBITENNIS: Speaking of data’s role in engaging the public, what do you think is in store for the media? Do you have a strategy to standardize the use of new metrics, like expected goals in football?
AT: Certainly, this is something we’re very aware of. Looking at American professional sports (like the NBA, NFL, MLB) and European football, statistics flow through media and betting spaces and are easily consumed by the public. It’s part of the storytelling. You can’t flood people with statistics, but a few targeted metrics can lay the foundation for a narrative that highlights new aspects or objectively explains trends. After all, our name signifies our ambition; we’re TENNIS Data Innovation, not just ATP Innovation.
We want to help elevate the use of data & analytics across the sport. ATP Media commentators can also request on-demand statistics on dimensions they deem relevant; for example, analyzing Djokovic’s return performance across sets. We think it’s important to make these tools available to other broadcasters to elevate the entire experience. From an editorial perspective, it’s useful to provide these insights for constructing stories across the media ecosystem. Here too, we’re thinking not just about making tools available, but also about their ease of use and flexibility, possibly allowing for ad-hoc information requests through parametric query interfaces. The idea is to enable journalists to validate their story about a particular match and add depth to the narrative by highlighting and quantifying game patterns that are difficult to discern and quantify by eye. For example, we’ve presented analyses on social media about Sinner’s serve improvement in the second half of 2023 through succinct infographics.
Source: X account ATP_insights
The goal is to move beyond elementary statistics like break points and first serves. We want to show the real reasons why a player performs a certain way or why a match has gone in a particular direction.
UBITENNIS: The objectives are clear, but what would you say is the strategy for implementing this ambitious plan?
AT: At TDI, as you mentioned, we have social media accounts where we develop our analyses, which serve as laboratories to experiment with new ideas and gauge reactions to new metrics we propose. This testing ground helps us introduce these new ideas to ATP media, incorporating these nuances and perspectives into the content of ATP and ATP TV accounts. These aren’t for everyone; the idea is that we reach the public through commentaries, regardless of the platform (pay TV, free-to-air TV, OTT, etc.). The goal is to work closely with ATP media, where we have a significant fan base. It’s an organic strategy that touches on all points of contact, aiming to incorporate our insights into the narrative.
UBITENNIS: Moving to the betting space, what developments do you foresee and what metrics might bettors be looking at in the coming years? What could be the killer analytic in tennis?
AT: We work closely with our partner in this space, Sportradar, to surface additional statistics for betting clients. We’re developing new products to provide more in-depth insights. In my opinion, performance rating, which synthesizes overall player performance into a single number, could stand out. For instance, in the Turin final between Sinner and Djokovic, Djokovic played an outstanding match: according to our performance rating, which ranges from 0 to 10, he recorded the highest value ever in an ATP match since we began our measurements.
This validates the quality of the match and demonstrates that Djokovic’s victory was more due to his own merits than his opponent’s underperformance. Another example: Rune versus Djokovic in Turin. Rune played a high-level first set, and Djokovic was slightly below his level. However, when Djokovic raised the bar, Rune’s level began to drop, indicating his level was dropping. Finally, “momentum” could be another valuable metric for betting, as it provides immediate trend indicators for live betting.
UBITENNIS: Lastly, I’ve noticed that for some tournaments, ATP live scores provide richer statistics than others, particularly for ATP 250 and 500 clay tournaments where Hawkeye might not always be implemented. Can you confirm this?
AT: We work closely with ATP tournaments. Historically, ATP data was presented by Infosys, and where Hawkeye was present, the statistics are complete. Any event where Hawkeye wasn’t present, Infosys were unable to present full statistics. However, I can say that we are now collecting richer data across all ATP tournaments. In 2024, we aim to achieve uniformity in data collection and analysis for all ATP events.
