How Psychology Can Help You Win More Games in Among Us

by George Karapetsas

Your crewmates will be putty in your hands


Although it’s been around for more than two years now, it’s only been in the past couple of months that InnerSloth’s space-themed social deduction game, Among Us, has exploded into popularity, topping Twitch charts on an almost-daily basis and getting a lot more attention over on YouTube. Based on the real-life party game Mafia and the product of a three-person team, the game released to very little fanfare (or even attention, really), but its recent surge in popularity — it’s gone from having “fewer than 30-50 players per hour” at release to “70,000-110,000 per hour” toward the end of August — has seen InnerSloth cancel the sequel it was planning to release so that it could focus on improving its newfound cultural phenomenon. 

Among Us sees a group of up to ten crew members working together to complete a series of tasks to prepare a space station for launch, all the while trying to uncover the identity of an impostor (or impostors), who goes around the station attempting to kill the crewmates. Whenever a member of the crew finds and reports a dead body, the players can use the in-game messaging function — or better yet, a Discord call — to try and figure out who the killer amongst them (haha, get it?) might be. The crew then votes on who they believe is the impostor, and the person with the most votes gets removed from the game whether they were the killer or not. The crew members win the game if they correctly expose the impostors or complete the tasks set out to them; the impostor or impostors win if they manage to kill everyone. Given that it’ll only run you AU$7.50 on Steam and it’s free on mobile, it’s easy to access and easy to learn… but very difficult to master.

The state of Victoria is in its second lengthy lockdown period, so Among Us has become the arena in which my friends and I have our weekly game nights. Instead of doing anything that vaguely resembles actual productivity outside of those game nights, though, I’ve focused my attention on ensuring that I win every single game for the foreseeable future (if you’re reading this, Luke, go to hell). The internet is full of guides on how to best approach being a crewmate or the impostor, but I’ve got something even better than that: I’ve got psychology, and even though I know my own friends will see this, I’m going to share the techniques I’ve been using because they’re sure to increase your win percentage — when they do, remember you found them on Doublejump.  

Among Us

Change the narrative:

At its core, Among Us is about witness testimony and memory. Fortunately or unfortunately, human memory is incredibly easy to manipulate; thousands of studies have consistently shown that it’s possible to implant ideas in participants’ memories simply by altering the questions and statements you use when exploring them. One of the most famous examples of this is that of Loftus and Palmer (1974), who showed their participants a video of a car accident and then gave them a questionnaire, with one of the questions asking how fast the cars were going when they “hit” each other. Subsequent groups received the same questionnaire, but with the verb “hit” switched out for words like “bumped”, “collided”, “contacted” and “smashed”. Loftus and Palmer found that the choice of verb in the questionnaire appeared to influence the participants’ perception of how quickly the car was moving: those who saw the verb “contacted”, for instance, estimated that the cars were moving at 51.2km/h, while those who saw the word “smashed” estimated that they were doing 65.7km/h — a considerable difference. 

That in itself may not seem useful in this context, but the experiment had a second phase to it. A week later, Loftus and Palmer followed up with the participants, asking them a single question: “did you see any broken glass?” The video itself did not show any broken glass (as surprising as that may be in itself), but the participants who had the verb “smashed” in their questionnaire a week earlier —  as opposed to something like “hit” or “collided with” — reported that they had seen broken glass. This suggests that the verbs themselves had at least some impact in modifying and distorting the participants’ memories; of course, the one-week delay may have played a part as well, but in a high-stress situation like a game of Among Us, memories are more prone to lapses and therefore, more prone to manipulation. That brings us to how we can apply this theory in Among Us

Among Us

In this hypothetical, we’re in a game where Blue has called a meeting in which they’re casting suspicion on their crewmate, Orange, but Red is the actual impostor. Some guides would recommend that Red just stay quiet and let Blue and Orange argue with one another, eating up the discussion time, not arousing any suspicion and potentially letting either Blue or Orange say something that gets them suspected. That’s a perfectly valid strategy, for sure, but there’s a more “aggressive” option: Red could align themselves with a crewmate, let’s say Green, while completing the tasks, at the same time keeping an eye on other players. 

In the meeting, Red could then ask Green a leading question like “how fast do you think Orange/Blue was going when they darted past us?” or “how long do you think Orange/Blue was waiting when they snuck in behind us?” These types of questions imply that Orange or Blue has done a certain thing, but because that intention isn’t overtly obvious, they will often plant the idea in Green’s mind — and Green’s corroboration of that question will plant the thought in the other crewmates’ minds. Red can even repeat the strategy later on in the game because a crewmate won’t typically notice a well-presented leading question unless they’re looking for it; as a crewmate, the best way to pick this strategy out is simply to be looking out for it. Additionally, keep your own questions more general, asking things like “what can you tell us?” rather than “did you see when X did Y?”, as this reduces the sway that leading questions can have. 


Tell them sweet little lies:

Leading questions are hard to pull off without getting caught out, both in Among Us and in real-world applications. What isn’t so difficult to pull off, though, is simple misinformation, and it just so happens that Johnson and Seifert (1994), among many others, have shown that humans have a very difficult time negating the influence of misinformation, even if it’s retracted immediately afterward. Johnson and Seifert split their participants into three groups and told each group a story about a warehouse fire caused by short-circuiting in a storage closet, which differed ever-so-slightly between the three groups: they told the first group (the control group) that the closet where the fire started was empty; they told the second group that there were flammable items in the closet, and then retracted that information immediately; and they told the third group that there were flammable items in the closet but waiting until much later on to retract that information. After telling the story to each group, Johnson and Seifert then had them complete a questionnaire about the fire, testing what they thought caused it. The control group attributed the fire to the faulty wiring, while the other two groups both attributed it to the flammable items — despite the fact that that information had been retracted. 

Applying this principle to Among Us is quite simple. In this hypothetical game, Blue has reported finding a body and everyone is asking the customary who and where questions, except Cyan isn’t talking. The impostor, good old Red, can call attention to Cyan’s silence, casting an inkling of suspicion on them (but not enough to make it obvious) and immediately follow up with a “just kidding”. Alternatively, Red can use the same logic to suggest that Blue self-reported finding the body, watch Blue get flustered, act out and create a self-fulfilling prophecy that leads everyone to suspect them… and then take it back. If Blue still hasn’t been thrown out later in the game, Red can repeat the tactic and/or call attention back to Blue’s actions throughout the game. Here’s an example of how this strategy played out in one of my own games; in this context, Cyan is adamant that they saw Red kill Purple, but the other impostor and myself were able to cast doubt on both players. In the following rounds, Red and then Cyan got voted off, which allowed Orange and I to win the game. 

There’s an important caveat with this strategy, though: it’ll only work if you’ve already earned a couple of other players’ trust — Among Us motivates players to question everything, which is one of the key ways to negate the effects of misinformation, so you’re going to be better off if fewer people have a reason to question your lies. Orange and I got lucky in the example above, because Cyan and Red were motivated to ignore facts and accuse each other, which allowed us to manipulate poor Brown. Additionally, it’s important to note that if you’re on the receiving end of this strategy, just saying that you’re not guilty and putting your word against someone who’s gotten a few crewmates to trust them is not going to cut it; people want an explanation, so the best defence against misinformation is to stick with another crewmate at all times so that they can corroborate your side. Alternatively, just keep your cool if someone accuses you, because you know the truth and they’ll slip up first. Whatever you do, just don’t pull a Red. That’ll just make you look more suspicious.


Take advantage of cognitive modelling: 

Of course, impostors can also win by simply convincing everyone else that their arguments are true, and there are two effective, somewhat foolproof methods through which to achieve that — and I believe I know when it’s best to use each one. Social psychology suggests that the way an individual presents an argument has an influence on attitude and behavioural change. The Elaboration Likelihood Model proposes that there are two major ways in which individuals consider the content of a message: the central route, which considers a lot more cognitive information like the quality of the argument, individual motivation and how important the decision is; and the peripheral route, which comes into play when the individual isn’t motivated or doesn’t have enough time to engage with the message. That peripheral route is also subject to mental heuristics, such as the quantity of arguments. 

Among Us

A schematic diagram of the Elaboration Likelihood Model.

Throughout my time with Among Us, I’ve noticed that the later it gets in the game, the less likely the crew is to throw someone out. This makes sense when you consider the game’s mechanics. Early on in the game, the crew outnumbers the impostor(s) by quite a bit, and its chances of winning increase dramatically if they manage to eject an impostor early. On the other hand, as the game wears on and there are fewer people left, people are less confident in their ideas because the wrong decision could mean an instant game over. When considering this, the impostor’s optimal play becomes obvious — cast as much doubt as possible into other players’ minds and encourage early ejections whenever they can. Early in the game, people don’t have a lot of motivation to critically engage with any evidence presented for a number of reasons: it’s easy enough to just pop into another game if they make a mistake early on and get ejected; there are too many crewmates talking over one another so it’s difficult to make a coherent point; and, of course, the time limit forces everyone into quick decisions. The impostor(s) can therefore take advantage of the early game, overwhelm the other crewmates with argument quantity rather than quality, and get a few players out the airlock.

Of course, the simplest strategy is also the one that fails the easiest. As you get further into the game, crewmates have more of an opportunity to contribute to the discussion because there are fewer people involved in it, and they’re more motivated to engage with arguments critically because they can see their win in sight. If you’re the impostor and you get to this point, the crew will likely be more motivated to create a strategy — like a buddy system — to limit the impostor’s killing opportunities, so you should focus on giving yourself an alibi and relying on sabotage to separate members of the crew. If that doesn’t help and you can feel an ejection coming on, you shouldn’t try to argue your way out of it: the crew is smarter towards the endgame, so a more optimal play would be to suggest that you aim for a “task win” instead, citing what happens if the crew makes the wrong call. 

This strategy is based on cognitive modelling, so the best way to combat the strategy would be to have read this article and know about these processing models — that way, you can call them out and cast suspicion where it belongs. Failing that, just don’t entertain baseless accusations like “Red seems sus” without asking why Red seems sus. It might seem like common sense, but one only needs to spend an hour playing Among Us to know just how successful this strategy can be for an impostor. 

Among Us

Pay it forward:

If you’re planning to employ these strategies the next time you’re cast as the impostor, bear in mind that they all rely on the crew actually trusting you. You’re going to do a whole lot of talking during each discussion, but it’s going to mean nothing if the crew can’t account for your whereabouts in the round; since you’re going to be killing some people, you’re going to be unaccounted for at times, so you need the crew to trust you to be alone. For that, the crew needs to like you. My favourite way to earn the crew’s trust is by helping them — not in an overt sense, of course, because that would be counterintuitive, but in a more subtle way. 

Recently, I played a game where myself and Black were the impostors. I didn’t kill anyone when the game started, instead opting to walk with Brown and fake doing some of my tasks. When the crew found the first body, I vouched for Brown, who immediately vouched for me as well. Perfect. Pink was arguing that Black was the killer, but a little bit of misinformation caused Brown to doubt Pink’s testimony, leading to a skipped vote. Immediately after the skipped vote, Black goes ahead and kills Pink, and Brown says we should vote Black out the airlock. impostors typically don’t want to vote their fellow impostors out unless it would be suspicious if they didn’t — it wouldn’t have been suspicious in this situation, but I’d gained Brown’s trust and I wanted to keep it, so I backed them up. Black then disconnected from the game, leaving four players: myself, Blue, Lime and my good mate Brown. Lime picked up on the fact that Blue wasn’t active in the chat, so I used that to cast doubt away from myself and onto Blue. In hindsight, Blue was just away from the keyboard, but by this point all I needed to do was slip away, kill Lime and blame it on Blue. After that, well, see for yourself:

The trust that I earned from Brown in that game is an example of the reciprocal norm; people feel obligated to return a favour to someone who does them a favour. In this example, Brown felt obligated to vouch for me solely because I vouched for them. Then, when I agreed with Brown about voting Black out, they viewed me as less of a suspect because, in their mind, it would be silly for me to halve my chance of winning in that way. The more players the impostor can successfully get to trust them, the less of a suspect they’ll appear to be as the round progresses. Don’t overdo it, though — remember, you still need to kill to win. 

As a crew member in this situation, all you really need to do to beat this technique is avoid reciprocity: old Brown was blinded by their obligation to return my favour, but if they’d paid closer attention to what I was doing rather than just where I was, they would have known that I wasn’t actually doing very much, but they were blinded by their obligation to reciprocate. Marry yourself to the facts — only vouch for a player when you know for a fact that they’re not the killer — and you’ll be able to avoid getting played for a fool like poor Brown (Brown, if you’re reading this… sorry mate).


Ultimately, it would be impossible to break down all of the optimal decisions to make in Among Us, but reading this Herculean breakdown of social and cognitive psychology should give you a pretty significant edge when it comes time for a meeting. Sure, you may start to feel a little bad for messing with people’s minds like that, but just remember, if you’re the impostor… that’s the name of the game. 

Good luck in your future game nights, dear reader — may nobody suspect a thing. 


This article was originally published on Doublejump. If you enjoyed what you’ve read, you can support the site further by following us on social media, becoming a Patron, and/or purchasing some merchandise!

1 comment

Doublejump Digest: October 12, 2020 - Doublejump October 12, 2020 - 8:52 pm

[…] delivered a psychological masterclass in manipulation and deception that he promises will help you win more games in Among […]

Reply

Leave a Comment!