"Esports," where games are contested as serious competitions, now boasts worldwide popularity, with prize money soaring and professional gamers becoming household names.

Bringing revolutionary change to this world of esports is TsuCre, a team of graduate students from Kyoto University and Tohoku University who have joined forces to develop products for gamers.

They are developing an AI analysis and coaching platform for gamers. In simple terms, it's a service that analyzes gameplay footage and provides advice on how to improve skills based on that data.

Yuki Oya, the representative of TsuCre, is himself a gamer with over 4,000 hours of playtime. From this standpoint, he saw the diverse, worldwide potential of esports, leading him to form TsuCre and begin development. His goal is to elevate esports to the same status as traditional sports. We spoke with him about the potential for Thumaster, the first step in this endeavor, to change the industry.

Making Advanced Analysis and Coaching Accessible to Everyone

Could you tell us about the service you are currently developing?

Oya: "Thumaster" is a service for players of first-person shooter (FPS) games. By simply uploading their gameplay videos and device inputs, it automatically analyzes metrics like in-game accuracy, precision by weapon, accuracy by distance, and aiming stability. Based on these results, the service uses AI to provide feedback for improvement tailored to the player's challenges.

So, to put it another way, for players who want to get better, it provides advice based on their gameplay videos, like, "Your long-range accuracy is high, but your short-range is low. You should practice more at close quarters." Is that the right idea?

Oya: That's right. We're providing an environment where anyone can receive the kind of advanced analysis and coaching that was previously only available to pro teams and a handful of top players, and they can get it quickly.

In modern traditional sports, like baseball and soccer, a wide range of data is analyzed, from matches to individual skills, which is then used to improve technique, win rates, or for scouting players. Is a similar movement now emerging in the world of esports?

Oya: Yes. In baseball and soccer, a culture of data-driven development and evaluation has raised the level of competition and, furthermore, has helped to increase the appeal for both players and spectators. However, while esports has grown rapidly, it has lagged behind in this area. While pro teams and some top-tier teams have analysis tools and dedicated coaches, many players rely on feel-based practice, leading to inconsistencies in their growth speed and direction.

Using the earlier example, it seems there could be cases where a player's perception and actual results are misaligned—like they thought they were good at long-range, but were actually better at short-range—and they wouldn't even notice.

Oya: This can also be attributed to a lack of objective evaluation criteria. That's why in the player transfer market, past achievements and impressions are heavily weighted, and it's not uncommon for players with great potential but few achievements to be improperly evaluated.

In baseball and soccer, data is now used to objectively evaluate players in the transfer market and for acquiring new talent, leading to the discovery of "hidden gems." If Thumaster becomes widespread, it seems a similar trend could emerge in esports.

Oya: That's exactly right.

The Gamer's Desire "To Get Better" Was the Catalyst for Development

To begin with, what led to the development of Thumaster?

Oya: Originally, TsuCre was a team formed to develop a gaming controller that anyone could easily customize and repair. I was a hardcore gamer myself, and I was frustrated that my controllers broke so quickly... I wanted to solve that problem.

So it sounds like it started more as a hobby than a business (laughs).

Oya: However, reading the reviews for gaming controllers on Amazon, I saw that many players had the same frustrations, so I felt there might be a demand for it. Actually, Thumaster was conceived as a sales tool for that controller. The idea was that by analyzing gameplay with Thumaster, you could understand your own habits and choose the right controller.

And that ended up becoming your main business.

Oya: For various reasons, we had to stop the controller development. After it came to a complete halt, I was thinking about what to do next, and I decided to just play games with other gamers. While playing with streamers and VTubers, I saw them practicing earnestly for scrims (competitive practice matches) and community tournaments, all while having fun. That's when I realized the demand from gamers "to get better" was greater than I thought. Come to think of it, I'm a gamer myself, and of the 4,000 hours I've played, 2,000 weren't "good hours"...

When you say they weren't "good hours," do you mean they weren't fun?

Oya: It's not that they weren't fun, but in any game, players start out toddling and get beaten by enemies right away, right? So I wanted to get better and studied by watching YouTube, but I just wasn't improving. There are two ways to "get better." At first, I thought it was external factors and focused on choosing a controller, but internal factors—the gameplay itself—are also huge. That's when I felt a service like Thumaster could evolve the gaming experience.

Since esports is a digital field, one would imagine that data analysis is quite advanced. Was that not actually the case?

Oya: Startups in the U.S. have recently started emerging in this space. However, they mainly focus on the MOBA (Multiplayer Online Battle Arena) genre, where analysis centers on character movement and tactics. There was no data analysis for the physical skill aspect of FPS games, so I felt "we can compete here," "this might work."

What do you mean by "physical" in this context?

Oya: It has two meanings. One is how and how much you move the controller or mouse. The other is how many of your shots hit the enemy. Right now, in Thumaster's development, we're focusing on the latter.

I see. In any case, it's clear that TsuCre, and Thumaster, didn't start as a business-first venture, but from the frustrations, desires, and problem-solving of a true gamer like yourself.

Oya: In addition to that, since my university days, I've always thought entrepreneurs were cool. I had this vague desire to "live a vivid life," and that has also led me to where I am today.

Muzzle Flash, On-Screen Text... The Obstacles Hindering Image Analysis

What motivated you to apply for the ICT Startup League?

Oya: One reason was funding. Also, all the members of TsuCre are under 24, a very young team, so we lack experience on the technical side. Having someone to support us in that area was also a big factor.

What kind of technical challenges are you facing specifically?

Oya: First is improving the data analysis accuracy. For example, in game analysis, estimating the distance between yourself and an enemy is surprisingly difficult. Autonomous driving technology requires similar estimations, but it analyzes multiple screens from multiple cameras. However, with games, you basically have to analyze and estimate from a single screen. That's what makes it difficult. And since enemies move violently and quickly, you also need to track them.

So just recognizing an enemy as an enemy is a challenge in itself.

Oya: Furthermore, in games, the muzzle flash from your own shots—the spark when you fire a gun—is so large that it momentarily hides the enemy. Because of that, the system stops recognizing the enemy. As a player, a bigger, stronger spark feels good, but from an analysis perspective, it makes me want to say, "We don't need that much muzzle flash!" (laughs).

I see. There are a lot of struggles that aren't apparent from the outside.

Oya: Also, when we analyze a VTuber's game stream video, we have to remove on-screen text and other overlays from the gameplay screen to analyze it properly, which is also quite a hassle.

And the service probably wouldn't catch on without the ease of just uploading a stream video.

Oya: At the ICT Startup League, we're really hoping to meet people who are skilled at solving these kinds of problems. We also have ideas for businesses other than Thumaster, so we'd like to be introduced to people involved in those fields as well.

The Innovation That "New Metrics" Will Bring to the Esports World

Finally, could you tell us about your future business development and the company's vision?

Oya: In the short term, the alpha version of Thumaster is nearing completion, so we want to have pro gamers and streamers use it and give us feedback. Furthermore, we plan to complete a beta version with additional features early next year, and depending on customer reaction, we'd like to raise more funds for growth.

And what about the long term?

Oya: I want to build a track record that makes customers think, "If you want to get better at games, use Thumaster." Then, the deployment of Thumaster will also change.

In what way?

Oya: Thumaster is a B2C business for players who "want to get better at games," but its content can also be applied to B2B. As I mentioned at the beginning, I want to bring data-driven player evaluation and scouting into the world of esports and make it the standard. If we can use our analytical data to propose that "if you acquire this player, this team will become like this," I believe TsuCre can also take on player agent duties.

That's just like modern baseball and soccer.

Oya: In fact, when you talk to esports team managers, they say they are currently facing challenges with scouting and analysis. Scouting is often done through ad-hoc, almost analog methods, with someone in charge gathering information on platforms like X and making contact via DMs. The work of analysts is also a reality of each person doing it bit by bit with sheer manpower. In that respect, baseball and soccer often have measurement software and data available to the public, which makes it easier to identify and acquire players who seem to be a good match for the team, and to predict their performance if acquired. The challenge in esports is the lack of that initial screening data. We have the potential to provide that, so I see great potential in B2B as well.

It sounds like you could change the conventions of the esports world on a global scale.

Oya: TsuCre's goal is to elevate esports to the same status as traditional sports. To do that, we need metrics that convey the greatness of players not just to those involved, but to the general public as well. In baseball, if you say "Triple Crown winner," most people know how amazing that is. If esports also had such metrics to visualize performance, it would be easier to convey the players' greatness, which would lead to better evaluation and compensation.

The data that Thumaster analyzes seems like it could fit that bill.

Oya: Yes. Furthermore, we are also thinking of creating and popularizing "new metrics." Esports have become popular worldwide and prize money has increased, but when I talk to players, they say there's still not much of a dream. Even players who regularly compete in world tournaments might only earn around 200,000 yen a month. Their careers are short, from about 16 to 25, so I want them to be able to earn as much as a top-tier baseball or soccer player during their active years—enough to live comfortably after retirement or to pursue a new path.

Data analysis changing the industry itself. Your work at TsuCre holds a grand vision.

Oya: With Thumaster, esports athletes, just like traditional sports athletes, will be able to track their growth with numbers and optimize their training plans. At the same time, as objective data accumulates, teams and tournament organizers will be able to evaluate players fairly and conduct transfers and contracts in a data-driven manner. In turn, the value and growth stories of players will be communicated more clearly to fans and sponsors. I believe all of this will lead to an increase in the overall appeal of esports as a competition, and I want to continue working on this business with that in mind.