Wednesday, April 24, 2024

Findings from the Pro Lab: Cognitive Profiles

Written by:
Bonnie Qu
Team Liquid Crest Logo Light Version
Written by:
Team Liquid Crest Logo Light Version
Edited by:
Team Liquid Crest Logo Light Version
Graphic design by:
Wednesday, April 24, 2024

Findings from the Pro Lab: Cognitive Profiles

Team Liquid Crest Logo Light Version
Written by:
Team Liquid Crest Logo Light Version
Edited by:
Team Liquid Crest Logo Light Version
Graphic design by:
Wednesday, April 24, 2024

Findings from the Pro Lab: Cognitive Profiles

Team Liquid Crest Logo Light Version
Written by:
Team Liquid Crest Logo Light Version
Edited by:
Team Liquid Crest Logo Light Version
Graphic design by:

Humans are obsessed with drawing connections. In this world that so often seems ruled by random chance and coincidence, we naturally gravitate towards any sort of explanation that allows us to believe that we have some control over the moving pieces of our lives, however meager. Basically, we’re always looking for a way to make it all make sense.

Esports and gaming are no different. We’re constantly looking for patterns and reasons, whether it’s drawing on your opponent’s past moves to predict what’s next or coping with an in-game loss where you definitely didn’t just get outplayed. But seldom do we get the chance to draw larger conclusions about gaming based on real data from professional gamers — until now, thanks to the Pro Lab!

The Pro Lab is something we developed in collaboration with Alienware, with the goal of both improving esports performance and understanding it. The two go hand in hand, in fact; by using the data we gather from our professional athletes, we’re able to help them better understand which areas are their strengths and which ones need work, which leads to more efficient and focused practice.

We aren’t just doing this for the athletes, though. We’ve also been able to extrapolate some interesting conclusions about cognitive behaviors when it comes to gaming as a whole, like how IGLs fundamentally think differently than their teammates, or why MOBA mid laners have slower reaction times than supports do. We’re sharing them today in the hopes that they’ll confirm what you’ve always suspected — that yes, you were born to main Yasuo, no matter what the haters say.

Different roles = different brain functions

It feels fairly obvious that different roles within the same game require different skill sets and abilities. In a MOBA, the five roles are very clearly defined — top laner, jungler, mid laner, bot laner, and support. As a whole, MOBA players have higher working memory capacity, which can broadly be defined as one’s ability to hold information in their brain and make decisions based on that information, regardless of any other potentially distracting stimuli. Broken down into roles, however, players of different roles exhibit cognitive behaviors that reflect their differing approaches to the game.

As a preliminary note: all of the data below is taken from players at the very top of their respective games, which means we’re working with narrow margins. Esports athletes consistently display cognitive function far above that of the average person, so even seemingly unfavorable comparisons are shown within the context of these subjects performing well above the level of the regular population. Keep that in mind as you look at certain comparisons and rankings; normally, the difference is so small that you wouldn’t be able to notice, unless you were specifically looking.

Take mid laners as an example. Mid lane, by definition, is the one with the closest proximity to every other lane, and therefore is capable of the most flexibility when it comes to executing gameplans or making decisions on the fly. According to Pro Lab data, mid laners exhibited the highest average foresight. In other words, mid laners are the best with short- and medium-term predictions, and figuring out how things are going to proceed in any given situation. This is a hugely beneficial skill for a role that needs to constantly be adjusting based on what’s needed of them, whether it’s helping out in the jungle or setting up a gank.

On the other hand, mid laners are slightly behind when it comes to precision, a cognitive function that reflects a player’s ability to be accurate in their decision-making. Support players are the best when it comes to this, since choosing the right opportunities to initiate team fights is one of the support role’s most crucial responsibilities. Mid laners generally displayed lower precision because they’re not usually responsible for initiating — in most metas, they tend to follow up, lay down damage, and look for angles within a teamfight, which is a role with its own merits. There is such a thing as too many cooks, after all.

Support spotlight: CoreJJ vs. Boxi

Even among players of the same role, you can find significant differences in cognitive behavior. Below is a comparison of some of the data collected from League of Legends support CoreJJ and Dota 2 support Boxi, both high achievers in their respective games and certified MOBA veterans.

We can see from this graph that Boxi significantly outscores CoreJJ when it comes to automatic response, since he’s a little more of a trigger-happy support. Conversely, CoreJJ has higher response inhibition, which is how measured and deliberately constructed an action is. A player with high response inhibition is generally making informed decisions, whilst a player with low response inhibition is mainly reacting based on instinct. 

Both have high precision, though when it comes to their actual play styles this manifests itself in slightly different ways: CoreJJ is a thinker, while Boxi is a doer. Though many of their stats are similar, their approaches diverge in notable ways, despite both playing the same role. This likely has to do with the fact that Dota and League of Legends have core differences in how they flow; League generally has a more structured laning phase and revolves heavily around setting up fights to take key objectives, so naturally its supports are more cerebral. Conversely, Dota 2 games see more kills and a more active early game, which favors a more instinctual play style.

Initiating across genres

Even though MOBA supports are often the ones initiating fights, the “initiator” role doesn’t function the same way across genres. Interestingly, when we compare MOBA supports with FPS players, we find that their cognitive profiles most strongly resemble those of AWPers and fraggers in FPS games. MOBA supports scored highly in deliberate response time — reaction time, in essence — and low in response inhibition. Given that AWPers often have to react and decide on a shot very quickly, it adds up that they’d score high in deliberate response time.

Of course, we’re talking about players at the highest echelons of their respective games here, so the margins are pretty narrow. But since supports in MOBA games have the responsibility of seizing opportunities, it stands to reason that their reactions are similar to those of FPS fraggers, whose ability to react to things near-instantly and on instinct is their greatest strength.

In FPSes, you can divide the six broad FPS roles into two categories: the ones who make measured decisions based on prior information, and those who follow their guts and react based on intuition. The former category includes initiators, AWPers, and supports, whereas the latter includes fraggers, flexes, and lurkers. Examples of response inhibition in an FPS might be holding an angle that you think an enemy player will cross, or clearing a corner where you think someone may be hiding, rather than relying on the spray-and-pray to clear out hidden opponents. Basically, any decision that you act on based on prior information or logical deduction is a display of response inhibition — something that our calculated decision-makers excel at.

Meanwhile, those whose play revolves around intuition displayed low response inhibition. This isn’t a surprise, really; all three are fragging roles, with a focus on shooting the enemy before the enemy shoots them. They’re much more used to adjusting on the fly and not being given as much information to work with. Incidentally, that’s why these players are also typically the most interesting to watch — sometimes they crash and burn, but they also have the highest potential to pull off huge and unexpected plays.

How IGLs think

Outside of the in-game FPS roles, there’s one more that’s of crucial importance to any team: the in-game leader (“IGL”). In FPSes, the IGL is the one who usually makes calls and sets up plays, which is a learned skill set. Under the surface, however, there’s more going on with these players than just pure decision-making.

It’s not uncommon for IGLs to be fragging with the rest of their team, but there’s another dimension to their thinking, too. IGLs have higher working memory capacity as a whole, since they have to keep track of the flow of each round and make decisions based on that. They’re not just good at storing information, however; they also scored higher on information filtering, which allows them to filter out irrelevant information. This allows them to be more productive in their decision-making and lets them recognize crucial opportunities.

IGLs are generally more cerebral players, so they scored higher on prediction and foresight, too. They’re better at making short-term predictions and have a better idea of how certain situations will end up playing out, so they’re thinking about the future a lot more than their teammates.

There are certain intangibles when it comes to who makes a good IGL, of course, and data can’t ever tell the full story. Those who excel at IGLs typically have strong social skills, too, and have an affinity for telling others what to do. But knowing which cognitive skills IGLs typically excel in might help you identify improvement areas, if you’re looking to hone your craft.

Smash spotlight: Riddles vs. Dabuz

To cap off these findings, let’s take a look at the world of fighting games. These are pretty different to MOBAs and FPSes because they’re 1v1 games, so players aren’t operating within the context of a team. Two of our best, Dabuz and Riddles, have such obviously different play styles that you don’t need any Big Brain Data to spot it. What the brain data does is show us exactly how the players differ and give us the key areas of brain function that their characters and play styles build up. Plus, it can also give us a glance into what roles they might play best in an FPS or MOBA! 

Dabuz is best-known for playing the characters Rosalina & Luma and Olimar, two characters who require a lot of mental juggling and information storage. Rosalina & Luma is, in fighting game terms, a “puppet character.” Puppet characters control another entity on the screen and use them to hit opponents, take space, and set up defensive zones. Rosalina does this by maneuvering her Luma around the stage and inputting moves that her Luma will match from its position. So, Dabuz needs to be tracking not only where Rosa is and can be, but where Luma is and can be — and what options both Rosa and Luma have available to them at that moment.

Similarly, Olimar’s play style revolves around using different combinations of Pikmin to fight. He doesn’t control the Pikmin themselves so he is more a traditional zoner than a puppet character, meaning he treats Pikmin like projectiles to keep opponents away. (Olimar once again oppresses the innocent carrot people.) Olimar takes a lot of brain power to pilot because the player has to structure their game plan around Olimar’s Pikmin lineup — AKA which Pikmin he has on the board and which Pikmin are up next. You’ll regularly see Dabuz throw some Pikmin off stage to cycle into a better lineup. When Pikmin hit the opponent, they latch on and not only build damage, but they delay and change how long the opponent’s moves last, remapping punish windows in each match up. 

Riddles, on the other hand, is best-known for his Kazuya and Terry, two characters who are built around finding opportunities to combo and deal high damage. Both characters can light an opponent up for making a mistake, but they take a great deal of speed and reaction to pilot properly. They’re both large characters with bad defensive options, so if you drop your combo or miss your window of opportunity, you’ll be the one taking lots of damage. Not to mention, both have decently difficult combos to hit when you’re playing at the highest level. Their combos even have different pathing — responses you make mid-combo — based on what your opponent does while they’re being hit. So Riddles often needs to make quick decisions even after the hit. 

The Pro Lab findings line up pretty much perfectly with our two Smashers’ preferred characters Dabuz’s higher working memory capacity, information filtering, and information overload resist — his ability to pick out the correct information and reorient himself based on that — align with how his signature characters require him to consistently keep track of multiple moving pieces. Riddles’ lower response inhibition and slightly higher deliberate and automatic response reflect his more reactive play, his ability to capitalize on small openings to make a big impact, and the speed at which he can make the right decision.

These skills could have easily been developed and honed by the way both of these competitors entered into Smash as well. Brawl Olimar, the first character Dabuz played, was even more complex given he could have six Pikmin on the field. Riddles’s first main, Smash 4 Ryu, operated in a pretty similar way to Terry or Kazuya, capable of blowing an opponent up on an opening but also being exploited for failing that. Going even deeper, Dabuz got into Brawl out of his love for Pikmin — a game that probably does train similar skills to playing Olimar in Smash — and Riddles has played Tekken and traditional fighters that would train him for Kazuya or Terry.

Extending into Ultimate, the profiles offer some insight into why both of these pros picked the characters they did — and have both overperformed on them. (Dabuz is far and away the best Rosa and Riddles is easily the best Terry and likely the best Kazuya.) In Smash, Dabuz and Riddles have both been criticized for their character picks and not selecting something stronger. But when you look at their cognitive profiles, they really do line up well with their chosen mains. Whether that’s due to training their own brains for their mains, or some kind of fit that built up before Ultimate released, the profile could help explain why they feel so comfortable with their characters and less on others.

It’s also interesting to think about where both players would go in a team game. Riddles’s deliberate response time suggest the makings of a strong AWPer in CS, while Dabuz’s informational processing abilities imply a good IGL. Dabuz’s response inhibition might make him the better support in League, but Riddles’s deliberate and automatic responses could make him the better support in Dota 2. 

But it’s very important to remember that the data never tells the full story. Practice, experience, nurture, there are a lot of other factors that would sooner determine where Riddles or Dabuz fit in an FPS or Moba team. Still, the data does show how games like Smash can accommodate all kinds of different players. And even if no data could tell us for sure, it’s fun to imagine the timeline where Dabuz is the world’s best Valorant IGL, telling Riddles to go kill.

Conclusion, and considerations

Thinking about all this, you might be thinking about how to apply the data to yourself. You might even be asking if you were destined to play the roles you play. Once we find a role we like in one game, we often gravitate to similar roles in every other game too. Sometimes, a character or role seems to click immediately, whilst others remain inscrutable no matter how much you play them.

But it’s important to remember that we aren’t born with our cognitive profiles, nor are they resolutely unchangeable. For a long time, it was assumed that “perfect pitch” was something people were born with. Then music instructor Kazuko Eguchi developed a method to teach it to children, with an alleged success rate of nearly 100% when taught to children under the age of four. It’s the same story with language learning; learning a language before age 10 is your best bet for achieving native fluency, but language learning proficiency persists until you’re 18, then tapers off after that. The human mind is pliable — your cognitive profile is shaped by your upbringing and the stimuli you’re exposed to, and can change as you grow older too.

So, are we born to play our preferred roles? It’s probably a mix of both nature and nurture. Our data findings aren’t meant to help you come to a definitive conclusion on which roles suit which people; rather, they show how the way we play games shapes our cognitive behaviors — and vice versa.

Share article

Thanks for reading!

Though the findings came from pro players, we hope they can help you in your own gaming journeys. Go forth and conquer those lobbies - and take some points with you for good measure.
Your brain wants to click this button

Related articles

Team Liquid

Liquid Community Spotlight: Huntyy

This month, our Liquid Community Spotlight falls on Huntyy, a longtime Team Liquid CS fan and our resident sloth artist.

Team Liquid

Total Football: Team Liquid taps Team Gullit for return to EA FC

Team Liquid partners with Team Gullit to bring EA FC onto the Team Liquid roster

Team Liquid

Liquid Community Spotlight: Kandra

For this month's Community Spotlight, we're talking to Kandra, who's well-known on Reddit for his excessive intake of Team Liquid copium.

Team Liquid

Introducing MyBlue

The next step in the ultimate Team Liquid fan experience is here! Find out more about what we have planned in the near future, and beyond...

Team Liquid

How to create a nonbinary character

Nonbinary representation is a new concept, and one that we've begun seeing in games more and more. But what does it take to create a nonbinary character?

Team Liquid

Pets of Team Liquid

Come meet the friendly pets who REALLY run Team Liquid. Coming soon to shelves (not really).