Heuristic Analysis: How do principles of level design help players to effectively navigate a 3D space? December 2021
Introduction
When crafting levels, it is important that designers keep the player’s subjective experience in mind. This essay will explore how principles of level design are used in Cycle, with a focus on how they help to effectively guide players toward their goals. Cycle is a first-person exploration and adventure game centred around the traversal of hand-crafted 3D spaces and light puzzle solving, with a minimal HUD. Because of this, effective use of level design becomes crucial to the player experience, particularly when they are moving towards an objective. Heuristic analysis will be used to evaluate how effectively the game is able to achieve this. To inform this playtesting, research will be conducted into the fields of visual attention, lighting psychology, and theories of architecture. The Legend of Zelda: Breath of the Wild (2017) and Gone Home (2013) will also be used as case studies, being examined to assess how they utilize and arrange their spaces in order to lead players.
Contextual Research
According to the Encyclopedia of Neuroscience, ‘visual attention’ refers to “a set of cognitive operations that mediate the selection of relevant and the filtering out of irrelevant information from cluttered visual scenes,” (2009). Visual attention theories are an area of neuroscience that have had a profound impact on how we understand the way humans perceive their surroundings. Research into the area is beneficial in understanding how players relate to and interpret the play space they engage with. Among researchers in the field, it is commonly held that humans perceive their environment using either a ‘top-down’, or ‘bottom-up’ process. Richard Gregory proposed that perception is built on ‘top-down processing’, with stimulus being interpreted through stored knowledge and inference, as perception is constructed through hypothesis (Gregory 1970). In games, this could be considered as the goal-driven visual processing that occurs when players are actively searching for their objective, using assumptions and prior knowledge to filter out irrelevant information. Conversely, bottom-up processing proposes that perception is data-driven, with sensory information being interpreted in real-time, and beginning with environmental stimuli, before extending to wider surroundings. Gibson argues that our ‘perceptual systems’ are “ways of seeking and extracting information about the environment from the flowing array of ambient energy” (1966: 5). This directly relates to level design, as skilful use of lighting and form can draw players’ attention to particular areas of a level, thus guiding them on a desired path. With these theories in mind, it is therefore important for designers to make sure that bottom-up visual attention is used in support of the player’s already established top-down visual processing to ensure that players are aware of how to advance towards their goals. This bottom-up attention can be drawn in a number of ways.
A study conducted in 2008 found that the salience of lighting in comparison to that of its surrounding or previous conditions often determines where our visual attention is directed (Vincent et al. 2008). Studies such as this indicate that throughout our evolution as humans, we have grown highly sensitive to variations and contrasts in lighting conditions. This knowledge can be levied by designers in order to better direct and signpost players towards specific elements of a level. In painting, ‘chiaroscuro’ is a technique that refers to “clear tonal contrasts which are often used to suggest the volume and modelling of the subjects depicted” (The National Gallery n.d.) Fundamentals of chiaroscuro can also be applied to the lighting used in 3D game spaces where through deliberate alternations of lightness and darkness, a designer can guide a player to particular areas of an environment. Importantly, this application of lighting extends past general areas to include notable objects, or even boundaries of a level. Kaplan and Kaplan’s influential 1988 study of environmental cognition further supports this idea, in which they found that “a change in texture or brightness in the visual array is associated with something important […] something that draws one’s attention within the scene should turn out to be an important object or boundary…” (cited in Davis 2013). By illuminating salient areas and highlighting key objects or passages using strong, concentrated lighting, particularly that which stands out from the global lighting currently being used, a player’s attention can be drawn to effectively guide them through a level. Gone Home (2013), a similar title to Cycle, is an atmospheric ‘walking simulator’. Because of its minimal presentation, the level designers relied heavily on the use of chiaroscuro and lighting to aid their players in moving through the house. As can be seen in the game’s basement section (Figure 1), a key light, emanating from behind the door, draws attention in through its drastic alternation in salience, signposting the way forward and enticing the player to enter. From the literature discussed, it is clear that a well-designed 3D space uses lighting cues to lead the player’s attention toward their goals.
Figure 1: Screenshot from Gone Home showing the use of lighting in the basement section.
Evidenced by the emergence of texts such as An Architectural Approach to Level Design (Totten 2014), overlap between level design and architectural theories has been well documented in the discourse of game development. Like designers, architects employ stalwart theories and principles to guide observers through their authored spaces. In Image of the City, Kevin Lynch identified that the legibility of a city typically relies on the following elements: paths, landmarks, nodes, districts, and boundaries (Lynch 1960: 46). On our relationship with paths, Lynch writes “people observe the city while moving through it, and along these paths the other environmental elements are arranged and related” (Lynch 1960: 47). This application of ‘paths’ as a spatial guide can be found within real-life contexts in both hospitals and theme parks, places where easy navigation and orientation is imperative. Hospital architects have found that “aligning the patient journey with key architecture and interior elements alleviates the need for excessive signage,” (Staczek 2019). Within the context of level design, this can be understood to mean that play spaces and the objects within them should be arranged and structured in a way that allows for clear, diegetic player guidance, without the need for explicit, immersion-breaking signposting.
Drawing on Lynch’s concept of landmarks (1960), Totten writes “landmarks not only call attention to themselves, but also allow players to orient themselves by observing their relationship to the landmark in space.” (Totten 2014: 138) Therefore we can understand that if a player becomes lost within a play space, they should be able to quickly re-orient themselves using their surroundings and their character’s position in relation to certain landmarks. In fact, it could be said that the bottom-up attention of a landmark serves to support and guide the player in their top-down search for their goals. To be effective in this, landmarks should be visually distinct from their surrounding environment, and look different when viewed at various angles, allowing the player to better position themselves in the overall play space. The Legend of Zelda: Breath of the Wild (2017) features a sprawling map that the player could easily get lost in with little guidance. As can be seen in Figure 2, there are a variety of structures that players can use as reference points while exploring. Hyrule Castle, the centre of the map can be seen in the distance, backed by the imposing form of Death Mountain, the north-eastern point of the map. Large landmarks such as these are visible most of the time and aid the player with positioning on a global scale, something especially crucial when designing games with expansive maps. Amongst these, less prominent landmarks can be seen in the form of towers, which serve to help with positioning on a more local scale. These can be particularly useful when observed in relation to their larger counterparts, as the player can use this visual information to gain a deeper insight into their exact location without needing to consult the in-game map. As this case study shows, multiple landmarks of varying scales and visibility can work together to help form a more complete mental map in the player’s mind.
Figure 2: The Guardian. 2017. Screenshot overlooking Hyrule in Breath of the Wild.
Playtesting methods
Playtesting of Cycle was conducted using heuristic analysis. As outlined in Games User Research, this is a method used to help “research evaluators produce insights and design recommendations about a variety of aspects of player experience” (Anders Drachen et al. 2018: 254). Prior to testing, researchers outline a set of heuristics (core principles that uphold a particular player experience) against which the game is to be assessed. For a number of reasons, this method of playtesting can be a highly useful tool to critically assess a game. Firstly, the fact that researchers are testing against a set of previously established criteria makes it a highly technical and detail-oriented form of playtesting. Furthermore, the act of taking time to define these heuristics requires that the tester fundamentally understands the product in question, alongside its desirable and undesirable qualities. On the other hand, due to its highly technical nature, to get the most informative data it is traditionally conducted by experts and data scientists, making it largely inaccessible to smaller teams with less funding. Additionally, the data gathered is largely qualitative, and can be at risk of bias resulting from researchers’ own backgrounds and preferences, especially when evaluating player experience rather than usability heuristics. Not all playtesting is conducted via heuristic analysis, and development teams often choose to employ a variety of alternative methods of evaluating their games’ usability.
Think-Aloud Protocol is a popular method used to playtest level design and player navigation. It involves users being required to verbalise their thoughts as they complete tasks in games, providing the accompanying researcher with “an insight into the cognitive processes that go into using a product” (Anders Drachen et al. 2018: 189). This is especially effective for testing level design where, as previously discussed, cognition and visual processing is of high pertinence. In theory, this gives a more accurate representation of the player’s moment-to-moment experience than a post-game questionnaire or interview in which the user must recall their experience. However, as they are required to be constantly vocalizing their thought processes and experience, this form of playtesting can be mentally taxing on players. Social desirability bias can occur, making it crucial that researchers are external to the development team with no personal stakes in its success. It also involves slower data collection than other forms of playtesting, due to most sessions being one-to-one.
I will be internally presenting the results of my heuristic analysis using a GUR (Games User Research) report (see appendix). I will be detailing any examples of heuristic criteria being met, and more importantly, any issues. These issues will be labelled using a three-point scale from low to high, to prioritize issues in order of importance. The cause of any usability issues will be detailed, as well as the potential impact on the player experience, before going on to suggest possible steps that could be taken to mitigate the issue. This method allows for usability issues to be presented in a clear, concise, and comprehensive manner, often detailed in an easily readable table.
Playtesting results
The heuristics that Cycle was assessed against were largely informed by the contextual research undertaken for this essay. They have each been given a shorthand code (ranging from H1 to H7). See the appendix for the full heuristics table.
To start positively, H4 is a heuristic that was consistently met throughout playtesting, as architectural forms are typically aligned to the player’s journey. Rising structures such as spiral staircases are employed often for the player to ascend, contributing to the vertical progression that they are tasked with. The frequent use of archways, both natural and built, is also a strong motif which draws the user in to pass through them, making it an effective application of pathing. These elements of architecture were found to be particularly effective at leading the player when combined together (see Figure 3).
Figure 3: Screenshot from Cycle showing the combination of architectural forms to create a feeling of progression.
There are also examples of H6, the heuristic stating that “key objects and paths are highlighted and distinct from their surroundings”, being met. Notably, the first book that you encounter is highlighted by a ray of bright light (see Figure 4). Books are an important interactable in Cycle, being used in multiple puzzles, so this initial clarity in presentation does a good job of drawing the players visual attention towards them and underscoring their importance. Not all books are equally presented, however, with the subsequent one that the player encounters being sat in shadow. This could result in it being easily lost within the environment of the scene.
Figure 4: Screenshot from Cycle showing the use of light to highlight a book.
When playtesting, one key heuristic came up as being of high priority, with multiple instances of it not being met. H1 is a key heuristic, stating that “navigation is clear and logical, with the player knowing where to go”. The initial staircase encountered by the player features three directions which they can choose to take (see Figure 5), however only one leads them to the next section of the level. If the player is to pursue either of the other routes, after a while they will be met with a dead-end. Providing the player with multiple, equally viable routes that lead nowhere could potentially frustrate them and break immersion. Furthermore, this room is a prefabricated asset, being used a total of four times throughout gameplay, making it an especially prevalent impediment to the player experience. An obvious way to mitigate this issue would be to block off these pathways, so that only the stairs are available. Alternatively, in line with H7’s criteria, elements of chiaroscuro could be employed, with the ‘correct’ path being brighter and more visually distinct than the others. This would serve to make navigation clearer and more logical.
Figure 5: Screenshot from Cycle showing the routes the staircase asset.
Throughout playtesting, the experience fell short of meeting H7’s criteria that “lighting is used to highlight salient areas and the path forward”. On the whole, there were very few instances of lighting being used to draw players through the level. As previously discussed, spotlights are occasionally used to highlight important objects, however general lighting and chiaroscuro are largely ignored, with lighting appearing consistent regardless of an area’s salience to the player’s goal. This is likely to lead to events in which players are unsure of exactly where to go. With what we understand about how lighting is able to capture visual attention, it would be beneficial to utilise this tool more within the level design of Cycle.
Regarding H4, the large tower that the player is tasked with scaling initially serves as a key landmark for orientation. The first gameplay section features a wide play space, with the player being required to travel between various points within it. Because of this, having a large global landmark to use as a reference point proved to be helpful in orientation and navigation. However, issues arise in later sections of the game, as players progress to be underneath and eventually inside of the tower, in the process fully obscuring it and rendering it unusable as a spatial reference point (see Figure 6). With little help from distinct landmarks, the player could become lost in the play space. To remedy this, I would recommend the employment of more immediate local landmarks that the player can use to orient themselves in each section of the game. These would draw bottom-up attention that service as key aids to the player’s top-down processing.
Figure 6: Screenshot from Cycle showing how main landmark of the tower becomes obscured.
Conclusion
The heuristic analysis conducted found that there are examples of cognitive processes being effectively drawn in Cycle, with bottom-up signposting supporting the top-down visual processing undertaken by players in pursuit of their goals. However, there is still room for improvement within the game’s level design. Key techniques such as lighting and landmarks are underutilized, leading to an increased chance for player disorientation and frustration. A firm understanding of the research discussed in this essay could help level designers to better conceptualise design techniques that support the player’s experience. Application of these techniques through the crafting of levels will help to make engaging 3D spaces that players can navigate with ease.
Bibliography
Cycle. 2021. Wrong World Studios.
DAVIS, Robert. 2013. ‘Lighting Psychology: Cognitive and Emotional Responses to Lighting.’ LEDinside [online]. Available at: https://www.ledinside.com/knowledge/2013/12/lighting_psychology_cognitive_and_emotional_responses_to_lighting [accessed 13 November 2021].
GIBSON, James J. 1966. The Senses Considered as Perceptual Systems. Westport: Greenwood.
GREGORY, R L. 1970. The Intelligent Eye. London: Weidenfeld And Nicolson.
Gone Home. 2013. Fullbright: Fullbright.
DRACHEN, Anders, Pejman MIRZA-BABAEI and Lennart E NACKE. 2018. Games User Research. Oxford: Oxford University Press.
LYNCH, Kevin. 1960. The Image of the City. Massachusetts: M.I.T. Press.
MCMAINS, Stephanie A. and Sabine KASTNER. 2019. ‘Visual Attention.’ Encyclopaedia of Neuroscience [online]. Available at: https://link.springer.com/referenceworkentry/10.1007%2F978-3-540-29678-2_6344 [accessed 10 October 2021].
STACZEK, David. 2019. ‘10 Elements of the Perfect Hospital Design.’ Architizer [online]. Available at: https://architizer.com/blog/practice/details/perfect-hospital-design/ [accessed 27 October 2021].
The Legend of Zelda: Breath of the Wild. 2017. Nintendo Entertainment Planning & Development: Nintendo
THE NATIONAL GALLERY. No date. ‘Chiaroscuro.’ The National Gallery [online]. Available at: https://www.nationalgallery.org.uk/paintings/glossary/chiaroscuro [accessed 14 November 2021].
TOTTEN, Christopher W. 2019. An Architectural Approach to Level Design. Boca Raton: CRC Press.
VINCENT, Benjamin T. et al. 2009. ‘Do We Look at Lights? Using Mixture Modelling to Distinguish between Low- and High-Level Factors in Natural Image Viewing.’ Visual Cognition 17(6-7), 856 79. Available at: https://www.tandfonline.com/doi/abs/10.1080/13506280902916691 [accessed 14 November 2021].
List of figures:
Figure 1: In-game screenshot of Gone Home.
Figure 2: Kate Gray. 2017. The Guardian [online]. Available at: https://www.theguardian.com/technology/2017/may/30/the-legend-of-zelda-breath-of-the-wild-nintendo-game-design-open-world-player-explore [accessed 29 November 2021].
Figure 3: In-game screenshot of Cycle.
Figure 4: In-game screenshot of Cycle.
Figure 5: In-game screenshot of Cycle.
Figure 6: In-game screenshot of Cycle.