Moneyball, but for Kids: Games as Vehicles for Predatory Prediction Markets

Key insights

7 takeaways

01
The paper frames the proposal to bring prediction markets into children's games on Roblox as a continuation of 'gamblification,' the broader expansion of gambling mechanics into digital entertainment, and treats it as a policy problem exposing the limits of current platform governance.
02
Prediction platforms such as Polymarket and Kalshi are placed in a longer history of gambling's growth, with more than $3.3 billion wagered on the 2024 US presidential election outcome on Polymarket alone.
03
More than 35 US states have legalized some form of sports betting since the 2018 repeal of the federal sports betting prohibition, and European casino and sports betting apps have grown with young adults driving usage.
04
The authors treat loot boxes as sharing nearly all the attributes of gambling, yet they remain largely unregulated as gambling in most markets, with Belgium cited as a notable exception.
05
Current loot box regulation depends heavily on whether virtual items can be extracted or traded for real money, a condition a child-directed prediction market would likely circumvent.
06
Roblox's terms of use require developers to disclose odds for random virtual items, and the authors argue new clauses would be needed to govern real-world-linked rewards, given children's status as a vulnerable audience.
07
The work is a short theoretical position paper presented at DiGRA 2026 by Florence Chee and Daniel James Joseph rather than a systematic study, so its claims rest on documented events and secondary literature, not original data collection.

Summary

This paper, presented at DiGRA 2026 and authored by Florence Chee and Daniel James Joseph, argues that the prospect of prediction markets entering children's digital games, specifically Roblox, should be understood as a continuation of the broader "gamblification" of digital entertainment. The authors contextualize prediction platforms such as Polymarket and Kalshi within a longer history of gambling's expansion into sports, media, and games, and they frame the Roblox proposal as a policy puzzle that exposes the limits of current platform governance. Their stated aim is primarily theoretical and policy-oriented rather than empirical.

The paper highlights several data points to support its concerns. It notes that more than $3.3 billion was wagered on the 2024 US presidential election outcome on Polymarket, that more than 35 US states have legalized some form of sports betting since the 2018 repeal of the federal sports betting prohibition, and that European casino and sports betting apps have grown with young adults driving usage. The authors observe that loot boxes, which they treat as sharing nearly all the attributes of gambling, remain largely unregulated as gambling in most markets, with Belgium cited as a notable exception.

A central conclusion is that current regulatory approaches to loot boxes depend heavily on whether virtual items can be extracted or traded for real money, a condition a child-directed prediction market would likely circumvent. The authors argue that Roblox's terms of use, which require developers to disclose odds for random virtual items, would need new clauses to govern real-world-linked rewards. The paper concludes that such developments raise serious ethical and regulatory challenges, particularly given children's status as a vulnerable audience, and it calls for greater industry leadership and policy attention. The work is a short position paper rather than a systematic study, so its claims rest largely on documented events and secondary literature rather than original data collection.