40 GENIUS INTERVIEW Sabina Hemmi
EVEN FROM A REALLY YOUNG AGE, IF YOU SHOWED ME A NUMBER, MY FIRST QUESTIONWAS, ‘WHERE DID THAT NUMBER COME FROM?’
“This feels really good. This feels really bad”; but they couldn’t validate that with data. FEED: Did you ever feel at a loss with what to do with all that data? SABINA HEMMI: It’s funny. At the time, I didn’t think of myself as a data person. Now, obviously, I do. I’ve established myself working with data for a while. When I see data I just see opportunity. It was never like we saw new data and we thought, “I don’t know what this is good for.” It was always more that we saw data and had ideas. Back then people were asking, “What is actually good? How do you measure how somebody is good? How do you quantify that? Is it better for you to play a character that you’re personally really good on, or is it better for you to play a character that is the most strong right now?” I think as I’ve gotten more intertwined with the data, and have made it a career, I have realised that I’ve always been a data person. I just didn’t think of myself as one.
visualisation way. Numbers and data are just a common way that I use to express things. I would rather express how I feel emotionally by showing you a graph of what an emotional journey looks like. That’s an easier way for me to describe or present an idea, then trying to use words or trying to show you a single number. I think data visualisation is an extremely powerful communication tool. FEED: How much does data match up with instinct when you look at how gamers actually play? How good a judge are people of their own skills? SABINA HEMMI: That’s a hard question, because now everybody instantaneously has the data. It’s hard to measure the difference because people are perpetually validating their instincts and validating their ideas naturally as they look at data, especially if they are analytical, smart game players. When we were first uncovering pro player data, one of the things I saw that I found really interesting was that there would be a new player on the scene and that player would be really famous for playing a specific character. But what we would see when looking at the data, is that they would be really bad at playing against that character in the same way. You wonder
GRASPING DATA The Elo websites collect and offer up massive amounts of game data for Overwatch, Dota 2, Fornite and Artifact
why would they be bad at that? I have a theory that they are so good with the character, that they know how they would counter everything. They’re so focused on it, that they almost become a little scared of the capabilities of the character. For some pro players, that’s a hard thing. It’s their signature character and they are actually awful at playing against it, when they should be good because they know that character very well. SABINA HEMMI: I think the digestion of the data is somewhat similar. But the data is less accessible in traditional sports. They have developed great data collection mechanisms, but a lot of them have taken years and years to develop these heavy data visualisations. They do things like measure a player’s body movements and position and things like that in an automated way, and that just takes a ton of technology to do. Not to say that esports doesn’t take tech, but the initial accessibility is a lot easier for us. FEED: How does data collection in esports compare with that in traditional sports?
FEED: When did you “wake up to data”?
SABINA HEMMI: Even from a really young age, if you showed me a number, my first question was, “Where did that number come from?” Which to me is the mark of a data person. I also think in a data
feedzinesocial feedmagazine.tv
Powered by FlippingBook