Marshmallow Test

I really wish game developers would just let us eat the damn marshmallows already.

If you have never heard of the test before:

The Stanford marshmallow experiment was a series of studies on delayed gratification in the late 1960s and early 1970s led by psychologist Walter Mischel, then a professor at Stanford University.[1] In these studies, a child was offered a choice between one small reward provided immediately or two small rewards if they waited for a short period, approximately 15 minutes, during which the tester left the room and then returned. (Wiki)

I have been playing Prey lately, and noticed it does something similar. Over the course of gameplay, you accumulate a number of Neuromods, which are essentially skill points. At the beginning, you can only assign these points in “traditional” skills, such as hacking, increased weapon damage, more inventory space, and so on. A few more hours of gameplay later, you will be able to invest points in “alien” skills, like Kinetic Blast, short-term mind control, flame traps, etc. The game warns you though, that if you start gaining alien skills, the security system (e.g. turrets) in the space station will start registering you as an alien. It might also affect which ending you receive, although I have resisted looking at spoilers for that.

That is basically the marshmallow test. You can either be rewarded with fun new toys now… or you can abstain and be “rewarded” with a better ending later.

Prey is nowhere near the worst offender here. I have also been playing through the DLC of Dishonored off and on, and it’s a thousand times worse. In Dishonored, killing people (instead of knocking them out) increases the “chaos” of the city, which not only leads to a bad ending, it also makes the game harder by spawning swarms of rats that attack you on sight (and are immune to typical assassination skills). Which would be somewhat fine, if it were not for the fact that damn near 95% of the abilities and skills you unlock through gameplay revolve around killing people.

Life is full of delayed gratification. Most of us spend ~40 hours a week doing something we’d prefer not doing, in order to receive money weeks from now to finance the things we actually do want to do. Delaying our already-delayed gratification is some Inception-style nonsense.

Now, I do not necessarily have an issue with the best endings being difficult to achieve, or the existence of Achievements, or even just choice in general. What I have an issue with is a game that gives you a carrot and then beats you with a stick for eating it. The original Deus Ex made you choose between invisibility to humans and invisibility to robots. That’s a good choice! Note how the designers didn’t give you access to invisibility and then tell you there would be dire consequences to using it. That would be dumb.

Do not make your players choose between Fun and No Fun. Because some of them are dumb enough to choose No Fun, even when they hate marshmallows. Save us from ourselves.

Posted on February 15, 2019, in Philosophy and tagged , , , , , . Bookmark the permalink. 10 Comments.

  1. I can’t help but be reminded of the initial implementation of Island Expeditions in the new WoW expansion. You got worse loot if you went around killing mobs and racking up the score that was displayed at the top of your screen. You were supposed to drag out the round to allow more mobs to spawn and only kill those late arriving ones. Killing the big flashy mobs marked on the minimap would hurt you. Terrible design.


  2. I don’t think Prey is an example of this. The consequence of taking the Typhon mods is pretty minimal. The turrets do fire at you but they are easy to kill and if it really bothers you you can just hack them. I think taking no typhon mods does slightly alter the ending, but it isn’t on the level of an entirely new ending. A lot of things will slightly alter the ending in this manner and I don’t know if it’s really worth worrying about.

    The main reason to not take the typhon mods is that they are crazy OP and make the game much easier. In particular they turn decently tough enemies like the Technopath into pushovers.


  3. I live my entire life based on deferred gratification – I totally love it. Anticipated fun is hugely more satisfying than experienced fun in almost all cases. Looking forward to things is infinitely enjoyable. When the anticipated thing happens the envelope collapses.

    Potentiality is empowering and thrilling in a way that actuality can never be.


    • Ehh… I don’t buy it. Even from you.

      Why play videogames at all, instead of simply tapping the reservoir of fun that is “looking forward to” playing videogames? Substitute games with books, movies, or anything. I am assuming you are not simply dictating comments while in Lotus Pose, staring at the wall, imagining infinite joy from things not experienced.

      In the real-world Marshmallow experiment, the original conclusion was wrong – researchers had thought a child’s willpower explained the better life result correlation, but it turns out to be simple economics. The children who were well-off and experienced little food instability were the most likely to delay gratification.

      Similarly, it’s pretty easy to say things like “the journey is always better than the destination” when you are already enjoying the journey. If the journey sucks and you are miserable the whole time, the only saving grace is time and cognitive dissonance to redeem the experience.


  4. If it didn’t change the ending, would it be okay?

    It’s seems like a basic “upgrade with drawbacks” to me. You get better powers, but they have a drawback you have to work around. That seems like normal game play to me.

    Or are you suggesting that every upgrade must be an strictly better upgrade over the baseline? That the only reasonable choice is mutually exclusive upgrades, but both of them make you better than the baseline?

    Liked by 1 person

    • It’s pretty much entirely the unknown factor. Blind Choices are a peeve of mind.

      Drawbacks are okay, although I consider them a bit lazy. What happens in practice is that we can choose the ones that aren’t really drawbacks at all. In Sundered, one of the Perks gives you +10 Armor but reduces Gun damage by 25%. I never use the Gun, I don’t even like the Gun, and so that Perk just reads “+10 Armor.” Meanwhile, other Perks say things like “Enemies drop 40% more shards (currency)” with no drawbacks, because you have a limited amount of Perk slots to begin with. Drawbacks are a bit more interesting when they completely change the way you play. Sundered has a Perk that gives you 40% of your total HP as Shields, but then reduces your HP to 1. That one is an interesting choice.

      The drawback to choices is having to make a choice at all, IMO. Not being able to acquire all the upgrades, or only being able to equip X number of upgrades, is choice enough.


  5. I’m ok with delayed gratification, if it comes from my own decision to put off something to get better returns later. That’s autonomy.

    Less ok with delayed gratification of something that exists to be fun, mostly because that is stupid and illogical design. Why take the effort to make and design a system and reward players for doing it with a valid pathway if the goal is for them to do LESS of it? It’s like designing a bounty-criminal-jail system for sandboxes and then being surprised that a good number of players will be encouraged to try and test it out simply because the system exists and gives them both tacit permission and awareness of the possibility to murder and grief other people.

    And don’t even get me started on blind choices. I loathe them. Especially those that exist to pull the rug under you after, instead of trying to accept in good faith what a player thinks might logically happen from such a decision. I could not get through the first few decision points of the first Witcher game because it set up a chain of “oh you want to be ethical and do the good thing here? Haha, you’ll regret it three chapters later.” Even after consulting a walkthrough I ended up with the “I don’t think there is any choice that I would be content with” and quit the game option.

    I struggled through the Walking Dead only after all the episodes were out and everything was mapped, branching storyline decision tree in hand. Life is Strange was even worse, all five episodes are apparently a giant rug-pulling joke on the player. “Haha, think you can change fate and causality? Think again!” I lost the will to even go through the first major decision point, after having an ikling of what would transpire.


  6. dachengsgravatar

    The marshmallow test is very poorly designed. The experimenters do not at all consider the possibility that some of the subjects may simply not believe the experimenters when they tell their subjects that there will be more marshmallows if they don’t eat this one right here right now. Some of the subjects out-think the rigid minds of the experimenters and conclude that they might be lying. Or circumstances may change. A bird in the hand is worth two in the bush.


    • Yep, the original experiment results were not reproducible.

      And there’s a nice parallel with what you said and games. Do we trust the designers that the better ending will be worth the less fun we have now? There is the possibility that the two marshmallows they offer later will be stale, as compared to the fresh marshmallow in front of us.


  1. Pingback: Prey in Conclusion | In An Age

%d bloggers like this: