Michael Schuckers is the Charles A. Dana Professor of Statistics at St. Lawrence University in Canton, NY. An applied statistician he has received funding from the US National Science Foundation, the US Department of Defense and the US Department of Homeland Security. He is the author of over three dozen publications including Computational Methods for Biometric Authentication (Springer, 2010). Additionally, Schuckers has done work in sports analytics particularly ice hockey including consulting with a MLB team and an NHL team. For his work in this area, he was named a American Statistical Association's Section on Statistics in Sports "Significant Contributor".
+ Full Transcript
Tarran: So I’ll do a sort of introduction basically. So I’m Brian Tarran and I’m with Significance Magazine here today live from Denver, Colorado at JSM 2019, and I’m with Michael Schuckers of St. Lawrence University. Hi Michael.
Schuckers: Hi Brian.
Tarran: And we’re talking about e-sports today. Specifically a game called League of Legends, but I think we should step back a little bit about yourself and your background and what you do in your day job.
Schuckers: Sure, so I’m a professor at St. Lawrence University. It’s a small liberal arts college, one of these American sorts of things. It’s about 2,400 students, so mostly my job is to teach those students and prepare them. I do a little bit in terms of research. I’ve done a good bit on sports, primarily ice hockey in the last ten years. And I’ve got into the e-sports because I’ve got some colleagues that are interested and I’ve got some teenage kids.
Tarran: Well, e-sports is really massive now, it’s really growing. I see that in terms of ESPN running competitions. We recently had an intern in the office who had the- it was counter strike Go Well Championship on the screen and he was talking about the hundreds of thousands of pounds that were at stake for winning this, it’s a real growth area.
Schuckers: Yes, and I think in terms of some of the championships of some of these games like League of Legends, you’re talking about 100 million people playing in any given month, which is just massive. I happened to be touring around Colorado last week and was in a Walmart, of all places, and they had a small e-sports arena that you could rent and bring all your friends in and play against them and fully outfitted with the latest gaming gear.
Tarran: That’s amazing. So we’re familiar with statistical analysis of sports data, has e-sports fallen a bit behind? Have they not had that sort of statistical analysis behind the games yet? Or are they almost sort of keeping pace with what’s going on in the wider sports area, because of the fact that it’s all online, and there’s probably a huge amount of data that you can take from these games from the servers from …
Schuckers: Yeah, I think certainly in the academic space and even online that the statistical analysis of esports is probably not where a lot of the other physical sports are. And I think part of that is the data- it’s unclear actually, what should even be collected- and the companies that are producing these games are not making it easy to collect that data. And so, I think that we’re a little bit behind, but I think given the amount of interest there, it’s just a matter of time before it catches up.
Tarran: So you’re presenting at JSM on Monday. And your talk is focused on a game called League of Legends. Now, I’ve played computer games and I’m aware of League of Legends, I have no idea how it works. It seems a very complicated system. Tell us what you know of the game.
Schuckers: Sure, it’s basically- in the form that we’re analyzing- it’s a five-on-five game. There are basically three attack paths. It’s essentially- you’re trying to destroy your opponent’s tower. You have these three main paths, and then this sort of jungle in the middle. And because of that structure the champions, which are the players, tend to have fairly defined roles. And we were specifically looking at this bottom path where generally you have two players. So of your five you usually have one on the top, one in the middle, one in the jungle, and then two at the bottom?
Tarran: So these are the defenders essentially.
Schuckers: Yeah, the defenders are attackers, as it may be, so you’re going five-on-five, and looking at, were their pairs of players in that bottom path that worked well together?
Tarran: And when you say pair of players you’re not meaning the characteristics of individual players, you’re meaning the champions of, I mean so each character has a set of stats that – maybe it’s good at dealing damage, or absorbing damage-
Schuckers: Right. That’s the sort of thing, right. Are they hand to hand or are they sort of distance? Do they deal a lot of damage, but can only fire irregularly? Those sorts of things.
Tarran: So what were the questions that you were looking to answer when you –
Schuckers: Yeah, so the primary question- and I worked with two colleagues at St. Lawrence University- Chung Su Lee and Ivan Rambler, the primary questions were in that bottom area where you have two players, were there pairs that worked better than others? And we certainly found that there were. I think one of the other big findings is that because of the structure of the game and because of the nature of the battle and the players and the teams that you’re going against, the actual humans- the players aren’t really good at knowing which champions go well together. There seems to be a lot of evidence- so the way that teams choose champions is sort of a draft. So, you pick one and then that champion’s ineligible for the other team, etcetera. And yeah, they don’t seem to be very good at distinguishing and making out which are the best pairs.
Tarran: So would the assumption be that players are kind of choosing their favorites rather than optimizing basically for team strength?
Schuckers: Yeah, and I think part of it is that the feedback is very noisy. You’re getting feedback that is dependent on the quality of your teammates, the quality of your opponent, and being able to distinguish that from the quality of the champion, or the characteristics of the champion, that are important.
Tarran: So what were the pairings that you found worked best for surviving longest? Winning the game?
Schuckers: I think the general trend- and we found a good number, but the pairings that seemed to work best together were champions that have the ability to stun and champions that have the ability to rapidly fire. Although that’s just sort of a general trend among those that we found that was really, what we call symbiotic, which is you’re getting a good effect from both.
Tarran: And did you see- were these pairs used quite a lot in terms of actual competitions? Or where they’ve got an unusual pairing?
Schuckers: They seem to be not any more selective than other pairings.
Tarran: Right, okay.
Schuckers: Certainly, we have a lot- we have millions of matches that we were able to analyze the outcome of and so certainly used a lot but I don’t think any more proportionally than some of the other pairings.
Tarran: But this is, of course, an early study?
Schuckers: Yes, very early.
Tarran: I guess there’s much more than you want to see. You talked earlier about having- obviously you’ve got the champions interacting and the different skill levels of players- I guess there are ways you’re going to want at modeling that and trying to remove some of those effects to figure out the best strategy?
Schuckers: Right, and so in this, we were basically just looking at the champions used, not necessarily the humans and their skills. We’d like to be able to tease that out. The other thing is this data was originally collected in 2015, they’ve added more champions, they’re always tweaking the champions, and so yeah, I’d like to continue that, and see what other things we can do to improve.
Tarran: Have you had much interest from the company- is it Riot Games that makes League of Legends I think?
Schuckers: That sounds right…
Tarran: We can skip that.
Schuckers: Yeah, it’s Riot Games.
Tarran: Okay so, do you have much interest from Riot in what you’ve done?
Schuckers: Yeah, we’ve not heard anything from them. I’m sure they’re busy doing other things.
Tarran: I can imagine. But you are right, the balancing the fact that these games are online they’re always constantly being updated and, I think, from what I know of the way these games work is that the developer’s always trying to find a way so that the game is balanced, so that you can always have whatever combination of champions against whatever combination, and ultimately it should come down to player skill, rather than character stats, that make the difference.
Schuckers: Right, and I think that if there’s going to be any interest from Riot Games or other game-makers, it’s this idea that you mention of balance, and trying to balance out the characters.
Tarran: So they may very well use your findings to then tweak the characters and be validated next time.
Tarran: Well, that’s great. Keep doing it and I’m sure at least the players if they can latch on to this information enough, will get a competitive advantage the next time it comes around and maybe are in line for those hundreds of thousands of dollars in prizes that are up for grabs.
Tarran: Well, thank you, Michael. Thanks for talking to us about- I’ll stop there. Thank you, Michael. Thanks for talking to us about your work today, and I hope you enjoy the rest of JSM.
Schuckers: Thank you, Brian, good to be with you.
Brian Tarran: My name is Brian Tarran and I’m the editor of Significance Magazine. Find us online at significancemagazine.com. For this special JSM series of podcasts we’re collaborating with Stats and Stories. Stats and Stories is a partnership between Miami University’s Departments of Statics, and Media, Journalism and Film and the American Statistical Association. Follow us on Twitter, Apple podcasts or other places where you can find podcasts. If you’d like to share your thoughts on our program send your email to Statsandstories@miamioh.edu, or check us out at Statsandstories.net. And be sure to listen for future editions of Stats and Stories, where we explore the statistics behind the stories and the stories behind the statistics.