Recently, Evil Geniuses (often shortened to “EG”) has enjoyed its share of competitive success. Earlier this year, the organization’s “League of Legends” team won the League of Legends Championship Series Spring Split. Its “Valorant” squad made a heroic run in the esport’s North American playoffs and group stage, which included an upset win over the region’s dominant team, OpTic. And the organization also recently undertook an unorthodox approach to roster-building, hiring a 15-player roster for “Counter-Strike.”
In an interview, LaPointe Jameson said she wants to build an organization that knows the exact elements to building the best rosters for “Counter Strike” or “League of Legends” — and, to her, that starts with running the numbers, finding the signals that help the coaching staff identify and invest in talent with the right training programs.
“I could totally buy the best ‘Counter-Strike’ team tomorrow. It’s a money function,” LaPointe Jameson said. “But that doesn’t meant I know how to build a good ‘Counter-Strike’ operation. We will figure out how to build a good Counter-Strike operation with this.”
The Washington Post spoke with LaPointe Jameson about EG’s master plan — maybe mastermind plan? — to bring big data to esports.
The following interview has been edited for length and clarity.
Launcher: Evil Geniuses has been using in-game data to build rosters for different teams for about two years now. Since you’ve started taking this approach, what have you learned?
LaPointe Jameson: It started as a means to help buffer our recruiting pipeline. I’ve been with EG three years now. When I came in, I asked the incumbent leaders: “Hey, how do we add to the team? How do we replace a player?” It was the same answer over and over. “Ask the coach. They probably know someone.” But that’s only a piece of the puzzle.
And so we’ve been using data. We have just as many analysts and engineers as we do coaches, with a mixture of scouts that have more of a quant [quantitative analyst] background, to help us find the markers of what makes a good player. Some of those might be in-game metrics, some of those might be bioinformatics and some of them are more experimental as we figure out what are good markers of best-in-class champions across all different titles.
“We have just as many analysts and engineers as we do coaches.”
— Nicole LaPointe Jameson, chief executive of Evil Geniuses
When you sought out data analysts, was that a hard sell, or was it easy to convince people to come to EG rather than some big financial firm?
LaPointe Jameson: It’s an astute question. I was lucky on two ends. The first is our headquarters is in Seattle, which is not accidental in terms of the type of talent we knew we wanted to bring on, that was nonendemic to the esports scene, which in the U.S. skews heavily toward Los Angeles.
It’s a joke, and it’s maybe rude to even say, but we love the Microsoft, Amazon, Tesla burnouts tired of being a cog in a machine who want high agency in an innovative field. And that tends to be our engineer profile that you see today.
I was also really lucky, though, and I would say a lot of our success is based on: We got the unicorn. We have a classically trained Tesla and Google data scientist turned “Counter-Strike” coach Soham “valens” Chowdhury, who is our director of athletics for “Counter-Strike,” who was able to be an anchor voice and had a long-term belief, like myself, in “Yes, we should explore doing this differently.” He’s actually been one of my longest tenured employees since I came into EG, and helped us really build our data operation and find best-in-class engineers to follow him. So a mixture of right place, right people.
How has this data-driven approach failed in those two years? What lessons have you learned in building this out for the first time?
LaPointe Jameson: Oh, where do I — there are too many failures. Where do I even start? You always learn from failure, and at EG we try to fail fast and move on quickly and pivot and grow. We’ve had a few. I’d say the first was that we overestimated the cultural willingness of coaches and players to utilize data for decision-making, especially veterans. This problem is not unique to us — baseball, Formula One — you see this generational shift of the adaptation to data and two-way telemetry of decision-making that has a cultural nuance.
It wasn’t enough to have the most brilliant engineer. We had to have coaches and players who knew very clearly what they were getting into and that we were purposely doing things differently than they’ve likely done everywhere else in the ecosystem.
When you’re talking about whether coaches and players are receptive to the findings and takeaways, what does that look like? What are the findings from the data that suggest players should do something differently?
LaPointe Jameson: It definitely varies on application. A good tactical example is “League of Legends.” We have tools that can help stimulate pick-ban probability and how we should draft. That’s not how anyone currently does pick-ban.
Today, you rely on a coach’s intuition … [esports teams] are completely reliant on a coach to be informed, whether subconsciously or consciously, and to make these decisions. But now, we’re adding data, where our coaches get booklets of “Hey, here’s what they played, here’s scrim results, here’s what they’re picking and banning.” It gives more information for us to make an informed decision than just one source of information.
I would say culturally, for coaches, it can feel like it’s a slight on someone’s capabilities, right? It’s like, “Hey, our computer can do some of the stuff that you used to do.” But we’ve been able to show results and help make the coach’s life easier, where now they’re not spending 12 hours a week manually pulling information to help make those draft decisions. … But it takes a special coach to be open-minded to thinking that way.
So, where does the data stop and the intuition begin at EG? Have you guys already gotten to the limits of insights from big data?
LaPointe Jameson: Esports, generally, including EG, is still marred and mired with a lot of manual and key-man risk dependent dynamics that influence decision-making, whether it’s what player signs on, to when we call tactical timeouts in games. We haven’t even scratched the surface in potential for where we can get smarter.
Another good example of where we’re looking five, 10 years from now is how the definition of who can be an esports coach expands if in-game decision-making can be more automated and standardized. We can focus on other historically overlooked areas, which is what our current director of performance is really focused around: leadership skills, how to build team culture, how to manage resiliency. It’s hard to find the unicorn that can be all facets of a traditional sports coach, a people leader and tactical-positional knowledge. And that’s why data partnerships, like our HPE partnership, are really exciting for me. It gives us stepwise development in our professionalization of this, as well as opportunities to scale.
“We haven’t even scratched the surface in potential for where we can get smarter.”
— LaPointe Jameson
What is HPE providing EG to help the organization collect and analyze esports analytics?
LaPointe Jameson: HPE is fantastic. They have a product called GreenLake, which is a cloud-based AI machine-learning infrastructure that they are helping us with. And their mission is to bring edge to cloud. And what that means is to have data access and availability and decision-making in a real-time sense available in areas that push the edge of what could have been or wasn’t there before.
A great example is they have a partnership with Formula One where they have data centers in the garage that can take information from the car when it rolls in to help make informed insights around car performance. That is completely new territory in the world of esports where, right now, our data is printed booklets that our coaches bring to the stage. Imagine if we had access to tablets or live data feeds … Even if we’re not in our office or our coaches aren’t at their desk.
So, you’re imaging a world where a coach could be reviewing a tablet during a match that shows real-time data predicting the percentage likelihood that EG is going to win a match?
LaPointe Jameson: Yes, live odds, predictive analytics, even in-game economies.
In those initial steps toward big data, how did you show coaches and staff this is something that saves you time?
LaPointe Jameson: I think people underestimate the effort that has to go in. We can be as radical as we want and say “This is the way things are.” But if it can’t be implemented or utilized, it means nothing.
What problems are my engineers — coming from great places like Microsoft, like Indeed, but not from esports — solving for coaches? That requires two-way dialogue. … We’re really thoughtful around making sure people come in eyes wide-open to the opportunity and what we’re trying to do.
I could totally buy the best “Counter-Strike” team tomorrow. It’s a money function. But that doesn’t meant I know how to build a good “Counter-Strike” operation. We will figure out how to build a good Counter-Strike operation with this, but that’s not going to be a two-week exercise.
We’re looking at a different time horizon than much of esports today, which is: “We’re going to just buy the best player, just buy what I want now.” That’s why we have outsize salaries for our revenue streams. … We are trying to really build a long-term operation so Evil Geniuses are still here 15 years from now, known for and continuing to be known for winning the right way, in a thoughtful way.
You’ve outperformed expectations in both “League” and “Valorant” recently. How much would you attribute that success to your data-centric philosophy?
LaPointe Jameson: It’s very critical. We would have never found Danny and Jojo [two “League of Legends” players for EG] without this lens of how we want to do things differently. … So, this team is constructed quite literally as our true test and learn. We brought people up through our systems and we started them and the proof is in the pudding.
Where do you want to take this approach to roster building and competition? What is the big data solution three to five years from now for EG?
LaPointe Jameson: With certain coaches or sports franchises in traditional sports, the caliber of the training and experience you get with them is known. Today in esports, the underbelly — and the bad part — is it’s all kind of random based on who is spending the most. But if I could, my legacy at EG could be that here you would know that you’re going to be trained to be a champion.
Right now, esports operates in fine-tune optimization or mitigation of problems becoming an issue, versus here’s someone that has the markers of success that we can make successful. That would be exciting: An actual traditional athletics program in esports that also incorporates wellness, coaching, how to play as a team. That has to be done through data though, to be efficient and to scale.
