Among the games often called “GOAT,” there are those where the narrative story takes center stage.
When you think about it, these games are structured so that the player enters a fixed story and moves along that predetermined flow. Even though we are directly controlling the game with a mouse and keyboard, we cannot escape the set narrative, and the ending is also fixed.
In that case, the player seems to occupy a passive position, yet strangely, the more fully a game fits this mold, the deeper the immersion becomes, and a uniquely strong shiver and lingering resonance remain at the ending.
Why is it that, before a conclusion whose outcome is already decided, the player feels so overwhelmed?
「Cybertext: Perspectives on Ergodic Literature」(1997)
In Espen Aarseth’s Cybertext: Perspectives on Ergodic Literature (1997), applying the concept of “non-trivial effort” to the gaming experience reveals that the difference between a movie and a story-driven game is not simply whether the outcome changes, but whether the player invests non-trivial effort to reach that outcome.
In a story-driven game, there is only one ending, but the ‘trajectory of the process’ to reach it belongs entirely to the player. How many times you died, what strategy you used to advance—this does not mean the narrative changes, but that the performance is unique, and this is the point where it essentially differs from film.
However, Aarseth’s concept merely explains the structural condition of ‘having invested effort’; it does not answer why that effort feels like ‘something I did’. At this point, a psychological explanation is needed.
「The Illusion of Conscious Will」(2002)
According to the “apparent mental causation” theory in Daniel Wegner’s The Illusion of Conscious Will (2002), the sense of agency itself is not directly perceived from actual causal relationships in the first place, but inferred after the fact. The three principles this theory proposes are as follows:
- Priority: My intention preceded the action
- Consistency: The time gap between the action and the result is appropriate
- Exclusivity: There are no alternative explanations
The more these three conditions are met, the stronger the sense of agency the player feels, regardless of actual causal contribution.
This is precisely why, in games, immediate combat feedback provides strong stimulation even without narrative branching. Narrative branching and the sense of agency operate on different mechanisms to begin with.
Ultimately, the thrill a player feels at the ending is not the fixed conclusion itself, but their own accumulated performance up to that point condensing into a single moment and returning to them.
An interesting implication follows from this. A common trap in designing LLM-based interactions is the misconception that multiplying “choice branches” will increase agency, but viewed through the lens of the sense of agency, choice branches may not even be a necessary condition.
Rather, the temporal coupling between input and output, the visibility of feedback, and the manner in which user intervention is reflected in the outcome can be just as important as the diversity of content. Even if the final destination is fixed, if the system is designed so that the user’s intervention during the process is perceived as ‘non-trivial effort,’ agency can be strengthened without actual narrative branching. Conversely, even if given multiple choices, if their impact on the result is minimal, agency crumbles. The key, in the end, is aligning ‘actual influence’ with ‘perceived influence’.
The possibility shown by story-driven games is important here. Good interaction does not necessarily need to leave every ending open. What matters is a structure where the user invests non-trivial effort in the process of reaching that outcome, and that effort returns as a meaningful value within the system’s responses. That is, agency does not arise solely from infinite freedom, but also when one feels that their own performance accumulates and is reflected even within a bounded structure.