Loot boxes have long raised concerns among public health specialists. Their mechanics — based on random rewards, repeated purchases, and broad accessibility — closely resemble those of gambling. A particularly prominent role in this landscape is played by public broadcasts on Twitch and YouTube, where streamers open loot boxes before thousands of viewers, turning the process into a spectacle. An international team of researchers set out to study the actual behavior of streamers and uncovered an unexpected yet logical finding: the vast majority of openings occur not in isolation, but in long streaks triggered by monetization elements within the in-game interface.

Scale of observations and platform coverage

The total observation base amounted to approximately 65 hours and 28 minutes of streaming material. The primary analytical phase covered 13 public broadcasts on Twitch and YouTube with a combined duration of roughly 27 hours and 16 minutes. During this period, 490 individual loot box openings were coded. Eleven streams were conducted in Spanish and two in English. The games included in the sample were diverse: EA Sports FC 26, Brawl Stars, Clash Royale, Counter-Strike, and Pokémon TCG Pocket. The channels varied considerably in audience size: follower counts ranged from approximately 30,000 to 1.3 million, while average concurrent viewers ranged from 150 to 6,000.

93.1% of openings occurred as part of streaks rather than isolated attempts

The study’s primary quantitative finding speaks for itself: 93.1% of all coded openings were part of streaks — that is, chains of two or more consecutive openings within a single episode. Isolated openings proved to be the exception rather than the rule. This means that the real unit of streamer behavior when engaging with loot boxes is not the individual opening, but the chain itself, in which each successive action is fueled by the momentum of the previous one.

Episode, streak, opening: the anatomy of observation

The analytical structure follows a nested “matryoshka” principle. Each entry in the dataset corresponds to a single loot box opening. Openings are grouped into episodes tied to a specific context: one in-game menu, one reward screen sequence, one purchase sequence. An episode containing a single opening is classified as isolated, while an episode with two or more openings is classified as a streak. Within each streak, it was recorded for every opening whether another followed (continuation) or whether it was the final one (termination).

Time-limited offers and in-game currency as triggers

The formation of streaks was found to be significantly more strongly associated with visible monetization elements in the interface than with broader contextual triggers such as in-game moments or chat interaction. Key triggering cues included:

• time-limited offers displayed on screen
• a visible in-game currency balance
• an open in-game shop or storefront

In other words, the game interface itself “nudges” the streamer toward not stopping at a single opening.

Similar attention-capturing mechanisms are found in online casinos, where time-limited promotions, bonuses, and other offers become embedded in the gaming interface. A distinct category is represented by no-deposit bonuses, which according to nodepositau.com have become considerably more varied in recent times. Such promotions allow players to begin without an initial deposit, though their use is typically subject to wagering requirements and restrictions. The presence of a bonus, however, says nothing in itself about the likelihood of winning or the ability to withdraw funds.

A streak sustains itself, as if gathering momentum

One of the most notable observations was that once a streak had begun, continuation remained the dominant pattern. This continuation could not be convincingly explained by the rarity of the reward received, the streamer’s visible emotional reaction, or a near-miss effect. The authors describe this as a process resembling inertia: a sequence of openings, once initiated, develops its own momentum. The analogy to a flywheel is apt — once set in motion, it continues to spin without any additional external push.

A disruption in rhythm ends a streak more often than a specific outcome does

In contrast to continuation, streak termination was more frequently associated with a break in the current flow of actions: noticeable pauses, shifts in attention, or behaviorally marked moments. Certain categories of opening outcomes were also associated with streak endings, but no consistent pattern of “received a rare item and stopped” was identified.

Streamers’ commentary tends to sustain the “show” rather than slow it down

Qualitative analysis of streamers’ remarks showed that verbal commentary most commonly served functions of evaluation, interpretation, and performance for the audience. Streamers commented on the value of items, reflected on luck, expressed disappointment, and justified their spending. However, these statements almost never functioned as an effective mechanism of self-interruption. The public nature of speech makes it part of the spectacle rather than a brake on purchasing impulse.

Four questions the researchers set out to answer

The overall aim of the study was to understand how streaks of loot box openings are organized in public broadcasts. The specific research questions were formulated as follows:

• to what extent openings are structured as streaks rather than isolated episodes
• which interface and contextual cues are associated with the initiation of a streak
• what is associated with streak continuation after each successive opening
• what is associated with streak termination
• how the streamer’s verbal responses accompany continuation and stopping

The authors anticipated from the outset that streak behavior would predominate, that streak initiation would be tied to monetization cues, that continuation would not be explained by simple rarity or emotional reaction, and that streamer commentary would tend to frame the process rather than interrupt it. All four expectations were confirmed by the data collected.

From exploration to analysis: how the observation was structured

The study followed an “exploration, then analysis” design. In the first phase (October through December 2025), seven observers reviewed and preliminarily coded approximately 38 hours of streams, refining the coding scheme. In the second phase (February through April 2026), the revised coding scheme was applied. The material consisted of publicly available VODs (video-on-demand recordings), which allowed episodes to be paused and reviewed.

Inclusion criteria required clear visibility of openings, sufficient audio and video quality, the ability to reconstruct the sequence without gaps, a single game within the segment, and a language intelligible to the coders. Excluded were streams with fragmented or low-quality material, game switches, duplication, and cases where openings appeared commercially prompted through promo codes or direct purchase incentives. Two streams were excluded on precisely these grounds.

For each opening, the following were recorded: monetization interface features (shop, time-limited offers, currency balance), outcome characteristics (rarity, item type), the presence of a near-miss, the streamer’s visible emotional reaction, verbal remarks, and basic chat activity. A streak was defined as two or more openings within a single episode, and a stricter threshold of three or more openings was also applied to test the robustness of the results.

A public ritual with implications for a young audience

Serial loot box openings on streams transform loot boxes into a public, repetitive ritual accessible to millions of viewers, a significant proportion of whom are adolescents and young people. As the study demonstrates, the visibility of monetization elements in the interface triggers a chain of openings more effectively than the prospect of a rare reward. Meanwhile, streamers’ commentary — which viewers often treat as a reference point — more frequently normalizes and frames what is happening than it cautions against spending money. These findings are directly relevant to ongoing discussions about monetization transparency in games, the need for platform-level warnings, and the protection of underage audiences from gambling-like mechanics.