Amazing article! I'm a Roblox developer myself, so I can relate to this in a few ways.
I'd love to know about any instances of consistently popular / trending games where you didn't encounter this kind of gambling BS, and maybe some investigation on what those games succeed at to bring in attention without this exploitation. I think it could be helpful to developers who want to create a successful game but don't want to venture down the wrong path. As a Roblox developer myself, this is information that would be really valuable to me, since I absolutely DO NOT want to implement any of these techniques into my games.
I also want to point out that Roblox has actually enforced this to some degree recently, mainly through age regulations. Nowadays, playing games with these kinds of mechanics requires age verification, either through FaceID, official documents, or something else. This wasn't the case in the past, but there is a little mistrust over whether the verification technology is being used to steal data from people or not. That's not really relevant to this issue, but I felt like mentioning it here.
Overall though, very insightful paper; I might even show it to other Roblox developers in case it teaches them something!
What your taxonomy highlights really well is the difference between manufactured uncertainty and real-world uncertainty.
In most of the mechanics you document, the platform itself creates the randomness. The probability distribution is hidden, adjustable, and controlled by the developer. That turns chance into a tool of extraction.
Prediction markets often get lumped into the same category, but structurally they’re quite different. The uncertainty there doesn’t come from an RNG inside the system. It comes from reality itself: elections, sports, policy decisions, weather, etc. The platform can’t change the probability distribution of the event.
That distinction may end up mattering a lot for regulation. One system is essentially synthetic randomness optimized for monetization, while the other is collective price discovery around real events.
The interesting policy question might not be “does chance exist in the system?” but rather who controls the source of uncertainty.
If the platform controls it, the incentive is extraction.
If the world resolves it, the incentive shifts toward information aggregation.
Curious how you think about that distinction given the parallels you draw between loot boxes, casinos, and prediction markets.
I’d like to see surveys of kids before/after playing games. What characteristics of games are predictive of positive wellbeing?
Yes that would be great. There are some cool experiential studies of this in the gaming literature, but it’s hard to get a large sample size.
Amazing article! I'm a Roblox developer myself, so I can relate to this in a few ways.
I'd love to know about any instances of consistently popular / trending games where you didn't encounter this kind of gambling BS, and maybe some investigation on what those games succeed at to bring in attention without this exploitation. I think it could be helpful to developers who want to create a successful game but don't want to venture down the wrong path. As a Roblox developer myself, this is information that would be really valuable to me, since I absolutely DO NOT want to implement any of these techniques into my games.
I also want to point out that Roblox has actually enforced this to some degree recently, mainly through age regulations. Nowadays, playing games with these kinds of mechanics requires age verification, either through FaceID, official documents, or something else. This wasn't the case in the past, but there is a little mistrust over whether the verification technology is being used to steal data from people or not. That's not really relevant to this issue, but I felt like mentioning it here.
Overall though, very insightful paper; I might even show it to other Roblox developers in case it teaches them something!
What your taxonomy highlights really well is the difference between manufactured uncertainty and real-world uncertainty.
In most of the mechanics you document, the platform itself creates the randomness. The probability distribution is hidden, adjustable, and controlled by the developer. That turns chance into a tool of extraction.
Prediction markets often get lumped into the same category, but structurally they’re quite different. The uncertainty there doesn’t come from an RNG inside the system. It comes from reality itself: elections, sports, policy decisions, weather, etc. The platform can’t change the probability distribution of the event.
That distinction may end up mattering a lot for regulation. One system is essentially synthetic randomness optimized for monetization, while the other is collective price discovery around real events.
The interesting policy question might not be “does chance exist in the system?” but rather who controls the source of uncertainty.
If the platform controls it, the incentive is extraction.
If the world resolves it, the incentive shifts toward information aggregation.
Curious how you think about that distinction given the parallels you draw between loot boxes, casinos, and prediction markets.
Amazing research. Hopefully the courts use it!