Loot boxes, streaks, and compulsion loops: how modern game design keeps players hooked

In November 2017, Hawaii state representative Chris Lee stood at a press conference and described Star Wars Battlefront II's loot box system in specific terms:

"These kinds of loot boxes and microtransactions are explicitly designed to prey upon and exploit human psychology in the same way casino games are so designed." Chris Lee, Hawaii House of Representatives, November 2017

He was talking about a game rated appropriate for players aged 13 and over, sold by one of the world's largest publishers. Electronic Arts removed the paid loot boxes from the game within days. The underlying mechanic remained standard practice across the industry.

What Lee described had already been documented in academic literature, and has been studied more extensively in the years since. That certain game design features apply gambling mechanics to a context where players are often unaware that is what they are experiencing has been the subject of peer-reviewed papers, parliamentary inquiries, and regulatory action across multiple countries.

The compulsion loop

Most modern games are built around a compulsion loop: complete an action, receive a reward, receive a new action. The structure works because it aligns with how humans are motivated by visible progress. Completing a task feels satisfying. A reward confirms that effort produced something. A new objective gives forward momentum.

The design becomes harder to defend when the reward in the loop is unpredictable. Variable ratio reinforcement, identified by B.F. Skinner in the 1930s, describes what happens when rewards arrive on an uncertain schedule rather than a fixed one: behaviour becomes more persistent and more resistant to stopping than when rewards are predictable. Skinner's research documented animals pressing levers hundreds of times after rewards had ceased entirely. The same mechanism underlies slot machine design, and it has been deliberately embedded into game reward systems at scale.

Game features that function as variable ratio schedules:

• Random item drops from defeated enemies
• Loot chest and crate openings with randomised contents
• Gacha pulls for characters or equipment
• Critical hit animations with celebratory visual and audio effects
• Daily reward calendars with escalating surprise elements

Flayelle et al. (2023) documents these as contributors to loss of control in the gaming context.

Loot boxes: what the research shows

Loot boxes are purchasable in-game containers yielding randomised rewards: skins, weapons, currency, cosmetic upgrades. Research has established that they share structural characteristics with slot machines.

Shared features between loot boxes and slot machines (Drummond and Sauer, 2018):

• Random distribution of prizes
• Variable prize value
• Near-miss features designed to increase play
• Visual and audio cues calibrated to heighten excitement during the opening

What the data shows:

• A meta-analysis of 15 studies found a moderate positive correlation between loot box spending and problem gambling symptom severity
• The top 5% of loot box spenders generate approximately half of all loot box revenue
• Loot boxes appear in 49% of games rated suitable for children aged 7 and above
• Loot boxes appear in 93% of games rated suitable for children aged 12 and above

Researchers King and Delfabbro described the revenue concentration around heavy spenders as a structural feature of predatory monetisation in a 2018 paper in Computers in Human Behavior, noting that higher spending correlates with higher problem gambling risk scores. The business model is disproportionately dependent on the most vulnerable players.

How different countries have responded:

• Belgium (2018): classified certain loot boxes as gambling; required their removal
• Netherlands (2018): reached a similar classification
• China: requires disclosure of the probability of obtaining specific items
• US and UK: have not classified loot boxes as gambling; classification remains contested

Streaks and the cost of not playing

Daily login streaks reward consecutive days of engagement and impose a cost, visible or symbolic, when the streak breaks.

Duolingo's flame streak is probably the most widely recognised implementation. The counter is prominent in the interface. Notifications fire when a streak is at risk. Losing a streak after several weeks of consecutive days produces genuine disappointment for most users, a response the design is engineered to create.

The ADDICT study (2026) categorises streak mechanics as high-risk, defining them as systems that penalise users for insufficient engagement. Some games attach streak mechanics to digital companions or virtual pets that visibly deteriorate or die if the player does not check in. The emotional consequence of absence is built directly into the character design.

Structurally, what streak mechanics do is convert the decision not to engage into a felt loss. Loss aversion, the tendency to weight losses more heavily than equivalent gains, is one of the most consistently replicated findings in behavioural economics research. Designing for it is a deliberate choice.

Free-to-Play and the gradual commitment

Many of the highest-grossing games globally cost nothing to download. The early hours are often genuinely enjoyable and free of financial pressure. Players build characters, form social connections within the game, and accumulate progress over weeks or months.

The transition toward spending tends to happen gradually:

• Progression slows noticeably at a specific level or point in the game
• Competitive features require resources that take much longer to earn for free
• Social dynamics within the game create visible gaps between players who have spent and those who have not
• Limited-time events create urgency around items that may not return

By the time significant spending begins, the investment in time and identity is substantial. The escalation of commitment, a documented tendency to continue investing because of what has already been spent rather than what future value is available, is not something most players recognise in themselves while experiencing it.

King and Delfabbro's (2019) analysis of internal game patent documents from major studios found explicit descriptions of systems designed to identify individual spending capacity, target players at moments of emotional engagement, and create social dynamics that increased the felt cost of leaving.

At FairPatterns, we believe that fundamental human rights like freedom, dignité and privacy, should not « dissolve in the digital world ». Addictive design is a predatory practice, preying on the people these platforms pretend to serve. That’s why we did 3 years of R&D to create the concept of “fair patterns” (interfaces that empower users to make their own, free and informed choices” and built a multimodal AI that scans sites, apps and social media to find and fix dark patterns and addictive design.
We’re building the Human Safety Tech architecture that’s now indispensable to protect humans online and when interacting with AI.

Sources: Flayelle et al., Nature Reviews Psychology (2023); King and Delfabbro, Computers in Human Behavior (2019); meta-analysis of loot boxes and problem gambling, 15 studies; ADDICT Study, IHS Vienna (2026); Hawaii state legislative proceedings (November 2017); Drummond and Sauer (2018)

‍