Wednesday, May 17, 2017

A Study of Spaces in Yooka-Laylee



Steam tells me that at the time of writing this post, I have ~30 hours put into Yooka-Laylee.  I thoroughly enjoyed the game, and I still look to complete it by collecting all of the collectibles.  Playtonic did a fantastic job bringing their original formula into the modern day, with updated graphics and animations and large, sprawling worlds for the player to explore.

However, there seems to be a similar complaint among those who are faithful to the genre - after the first world, there seems to be less of a desire to hunt down each and every item or challenge.  Many seem to feel that the levels become "disconnected", and progression becomes very drawn out.  I too have felt that this game has had a few flaws, and I want to explore exactly what those are and what could have been done to fix them.  The question this puts in my head is this - How do you create a game in this genre, with the aim to create the same feel as it's predecessors, while at the same time meeting the trend of growing the space in which your game takes place?

Why write this?

Yooka-Laylee, and other games that fall under the Free-Range Platforming genre, appeal to gamers who like to explore, discover, and interact within an interesting 3D space.  While the game mechanics are important to the impression the player has over their experience, creating a space that places the player in a state of flow is almost more important.  Everything down to the tactile interaction the player has with the objects placed around them assists in creating a memorable and enjoyable experience.  It is in the expansion of these virtual spaces that we often see a change between generations of gaming, as an increase in memory and processing power allows for bigger and more interactive worlds to explore.  To create a world that in non-linear, but also naturally guides and encourages the player to take specific actions is a challenge indeed, and I hope to get a firm grasp on how 3D level design should be done for my own work in the future.

**Please Read This:  Despite how critical I may sound in the following paragraphs, I have thoroughly enjoyed this game and cannot wait to see what Playtonic does next.  I am by no means an expert on game design, but I'd like to take a step closer to understanding why a game leaves a particular impression on me.  This blog post is simply a practice of taking what I have learned from others and applying it to what I have experienced.  Please feel free to hit me up for some constructive criticism or if you have any counter-arguments to the ideas that I present here.

Overworld and Interest Curves

In his book, The Art of Game Design, Jesse Schell describes the methods in which developers organize the space in which their games take place.  The pattern in which these spaces are organized dictates the flow and direction that the player will travel when progressing through the game.  Free-Range platformers of the 90's took the approach of arranging their levels and stages into a web, like so:



 Each of the yellow circles in the web above would represent a level of the hub world (the floors of Grunty's Castle), while the branching purple circles represent the smaller stages.  The lines separating each of these areas represent some kind of level transition or another (i.e. doors or book portals).  This is a quick sketch of my understanding of the map, so if there are any shortcuts I missed, I'll edit the chart and add them in.

This segmented approach to level design marks an interesting transition between the linear level progression that platformers have seen up to this point, and the interconnected, open-world design that is often favored among many adventure games today.  This branching hub layout, however, aids in creating an alternating interest curve for the player.  After the exhilaration of exploring a brand new world, returning to the hub grant the player a bit of down time before striking out on their next big adventure.

Yooka-Laylee's hub world is Hivory Towers, the corporate lair of Capital B.  The size of the hub world matches that of one stage, each area carrying it's own unique challenges and characteristics.  As opposed to several stages that branch from this hub, however, Playtonic opted to create 5 expandable worlds, giving the player the freedom of choice to either try to progress further into the overworld, or stick around and complete the level they are currently playing.  When this option is presented, it presents a striking peak to the interest curve:  either way, you are going to have something brand new to experience, it's just a matter of deciding where to go and what you spend your pagies on.

*Attempted  representation of my personal first play-through experience.


Hivory Towers, on it's own, serves as the valleys to the interest peaks.  Even though it is considered as the hub and overworld of the game, it is usually treated as more of a sub-space in comparison to the other stages.  There is often less to collect, challenges for pagies are few and far-between, and it is really just a pathway between levels.  Having this familiarity in space in-between worlds that might be largely unexplored, however, offers an important break from the excitement.  When Schell discusses his experience as a juggler in the past, he describes these breaks as a chance for the audience "to relax and get to know you".  Giving the player time to become absorbed into this world could be considered an investment, as they will experience much more excitement for the finale.

This kind of interest curve can even be applied on a smaller scale, within each playable world.  Part of what made the collect-athon so popular among it's followers was the addictive quality of consistently being rewarded for gathering items.  This creates a particular kind of arc in player interest - as they explore a new world, they are picking up these items along the way, until they have come full circle and realize that they missed a few notes somewhere along the way.  Below is a reflection of my own interest curve when I played through Tribalstack Tropics:

*Attempted  representation of my personal first play-through experience.


Context

It may be fate that around the time I started doing my research into 3D Platformers, I finally got around to opening up Swink's Game Feel.  Glossing over the table of contents made me jump ahead to his masterful analysis on Super Mario 64, which was a mistake, as I completely passed over his descriptions of the metrics he uses to compare the "feel" of games.  His discussion on the metric of "Context" is one that felt the most relevant to the previously mentioned criticisms of Y-L.

The word context itself is defined as "the circumstances that form a setting for an event, statement, or idea, and in terms of which it can be fully understood and assessed" (via google).  In this regard, Swink observes context in games as the interaction between physical objects and their placement within a space.  This includes most attributes that are applied to the avatar, including size, speed, and behavior.

Swink uses 3 levels of context to measure the impression that the space leaves on the player:  "High-Level Context", which refers to the impressions of space, speed/motion, and size, "Medium-Level Context", which observes immediate space and object avoidance, and "Low-Level Context", which observes the tactile interactions between object within personal space.  Each of these impressions can make or break the design of a level, especially within a 3D space.

High-Level Context


The impression of space within any level in this genre is one that evolves over time, as you begin to get a better understanding of your position by memorizing particular landmarks.  This is made even easier by the circular shape of each map, that typically has a structure or large geographical feature at it's center - a strong, centered landmark.  This sort of design allows the player to map the area as they explore without spelling out directions for them.  Anyone who has attended a good-sized carnival can relate to this usage of landmarks - "Where was the cotton candy stand?  Over by that Ferris Wheel, right?".

Tribalstack Tropics, fully expanded.

The first world that opens up to the player is Tribalstack Tropics, and it's layout demonstrates how the designer can guide this evolution quite well.  The screenshot below is one of the earliest views that the player has after they spawn into this new and colorful environment.  



After following a short path from the spawn point, you are immediately met with a fork in the road and a highly characteristic horizon.  Directly ahead of you is a temple that marks the center of the map, to your left is a plain old dirt path, and to your right is a wooden bridge.  If your eye is a little more keen, you might be taking a particular interest in the giant, neon pink arcade cabinet that also has a spot in your line of sight...  This early on, you would already be taking mental note of this vantage point, giving you some notion of where to travel next if you happen to find yourself at these crossroads again.  There is no mini-map or objective marker provided in the free-range platformer; the player has to rely on their understanding of the world's layout based upon the placement of these landmarks.


Landmarks are what players remember and what they talk about, for they are what make a space memorable.
- Jesse Schell

This is not a new idea, and it doesn't need to be.  Swink's breakdown of Mario discusses Nintendo's use of context in their level design in great detail.  His analysis on the High-Level Context application in Super Mario 64 compares each stage's layout to "the zones of a theme park" where the "landmark-focused approach" provides a "delightful sense of vastness and exploration".  As I previously mentioned, Banjo-Kazooie also took this approach...




Now, while the most memorable levels from this genre typically followed this design, there were exceptions.  SM64 featured levels like "Big Boo's Haunt", which took place inside a haunted mansion, or "Hazy Maze Cave", which lives up to it's name.  Banjo-Kazooie had Gobi's Valley and Rusty Bucket Bay, both of which had many smaller stages/puzzles that took place indoors, separated from the overall perspective of the stage.  While these may have been considered to be refreshing changes of pace from the typical stage layout, they are often lauded even by the most devoted of fans, as it provides a disconnect from the spatial awareness that the player draws in their mind as they explore the landscape.

This is also the case with Glitterglaze Glacier, as much of the level takes place within caves or the Isymetric Castle.  In this manner, the majority of the world's landmarks and challenges are connected by a series of doors and bridges, rather than being a part of an interconnected loop.  In many ways, this instills the feeling that Glitterglaze Glacier was built on the fly, with a little less deliberation in design than Tribalstack Tropics.



Moodymaze Marsh takes this a step further, removing a central landmark from the world altogether, and limiting the draw distance to challenge the player's navigational skills.  This is, of course, deliberate, as the name of the stage itself suggests a disorienting perspective.  The designer balances this a bit through the use of pipe-tunnels that can be used as a sort of fast-travel to areas you have managed to find in the past.  However, this provides a similar disconnectedness that is present in Glitterglaze Glacier - a sort of incomplete logic in the map layout.  Ultimately, I found myself wanting to spend as little time here as possible, as I became more and more frustrated when trying to keep track of where I've been.



Capital Cashino steps back to the "circular theme park" layout, with a few rooms branching off that feature their own puzzles and challenges.  At the center of this map is a theme-appropriate water fountain featuring a lot of... character.


Returning to this style of map design, paired with the highly characteristic landmarks (by way of artwork, props, architecture, and puzzles/challenges), gave the impression of an arcade, where the playing and completing these games and challenges are more emphasized than the act of exploring.



The final stage before battling Capital B, Galleon Galaxy returns to the bridge-building design that we encountered in Glitterglaze Glacier, but in a manner that suits the overall theme of the stage.  Points of interest on the map are separated by a cosmic ocean, which can be traversed by flight or by teleportation.  As each landmark is large and just kind of floating out in space, it is rather difficult to get lost, but you are still met with an impression of openness and majesty on your first visit.

Similar also to Glitterglaze Glacier, much of the play space in Galleon Galaxy is actually separated from the outside space in which you begin.  Many challenges and puzzles take place inside of one of the many landmarks that populate the map, but they are self-contained.  Having one entrance and one exit for a room with a single challenge allows the player to keep better track of where they have been and what they have accomplished. What is jarring about these rooms, however, is how large and displaced they may seem - a small meteor holds a massive golf course puzzle, a massive moon holds a rather average sized sliding challenge.



This, of course, was not necessarily an issue in previous adventures.  The anthill in Mumbo's mountain was much larger on the inside than outside, as was The Salty Hippo in Treasure Trove Cove.  Even reaching the top of Grunty's Castle, when you think about it, didn't make a whole lot of sense.  However, in one way or another, you made your way through these rooms with some logical sense.  You start at the base of the anthill, and travel upwards, while the interior gets narrower and narrower; the Salty Hippo was partially flooded after being beached, so the player would have to jump atop floating boxes or swim to the floor to gather Blubber's gold.  In Yooka-Laylee, it felt like most of the rooms that branched away from the rest of the level didn't provide any natural guidance.  The Isymetric Castle, while an interesting concept, seemed like it was intended to be it's own world altogether, continuing on and on and making it easy for the player to lose track of where they've been.  The rooms of Galleon Galaxy felted simply like detached rooms that really only kept the theme of the stage, rather than utilizing the spherical shape of the structures in which they were placed or playing with gravity.

It is worth noting that before entering Galleon Galaxy, Yooka and Laylee are taught "Flappy Flight"  which can be used at any time and lasts as long as your energy bar does.  Granting this ability at the end of the game simplifies a lot of the issues I had encountered in the levels previously, and makes it a lot easier to search high and low for those last few missing quills.  The levels, for the most part, were built with this in mind, as they permit a lot of vertical travel space to the player.  It's how I managed to get a lot of the screenshots for this blog post.

 Mid-Level Context


The concept of Mid-Level Context observes the player's immediate surroundings, referring to the objects that can be interacted with.  Swink suggests that when measuring this level of context, the designer needs to examine 5 object traits:
  • The Number of Objects
  • The Size of Objects
  • The Nature of Objects
  • The Layout of Objects
  • The Distance Between Objects
The use of the term "object" can be rather broad when talking about the free-range platformer.  The first things that may come to mind are the collectibles that populate your path, or the number of enemies that will get in your way.  It can, however, even refer to simple static obstacles, like platforms or walls, that require a little added maneuverability to get around.  In a genre that prioritizes movement and navigation above all else, it would be fair to say that the feel of this level of context is extremely important to the impression a world leaves on the player.



Swink relates this level of context to his experience playing World of Warcraft; his traveling experience in WoW is compared to a long drive between cities, where a lack of scenery or change in direction result in "highway hypnosis".  He states, "Across multiple games, we can compare how far apart objects are spaced relative to the speed and motion of the avatar and can examine how this changes game feel". Swink later points out one of the biggest factors that benefited the feel of Super Mario 64 was the manner in which object placement was tuned to Mario's jump heights and trajectories, and how these relationships are presented to the player throughout the game "regardless of what else is going on in the level".  To put it a little more simply, you can get an impression of this level of context based upon the 5 traits in relation to the abilities of the avatar.

With Yooka-Laylee providing such an arsenal of abilities the player can use, utilizing each one in a meaningful way when the player is exploring a level can get very tricky.  A prime example of what this could look like, however, can bee seen in the very simplistic formula presented by Prospect's "Unbox".

From Unbox's press kit.

I've come to look at this game as the Goat Simulator of 3D Platformers - it is simple, with a lot to do, and by no means a perfect game, but I've found a lot of joy in it regardless.  What Unbox did so brilliantly is strip platforming mechanics back down to their basics and exaggerating them.  You have the clumsy action of rolling a box around on the ground, and the ability to "unbox" yourself up to 6 times, which essentially means that the player has 6 extra jumps that you can perform in mid air.  Pair these abilities with the appropriate balance of physics code, and you have a really satisfying avatar that is difficult to master.


What is great about this exaggeration of basic abilities, however, is that it allowed the designers to create massive, sprawling worlds for each level.  At least, they only appear massive in relation to your size as a cardboard box.  The ability to fly through the air using fans or your hexa-jump(?) allows you to cross these worlds in a matter

of seconds.  These added jumps allow the player to make-up for misjudged jumps or falling off the path, accounting for the unfamiliar sensation of trying to roll a box along a path.

Although Swink has stated multiple times in his book that context can only be measured as a soft metric, this put the thought in my head that I could measure object-in-level density by simply traveling across a level in a straight line, using the fastest method of transportation Y-L would currently have, measuring how much change in controller input is required.

This concept is by no means scientific, so take the following with a grain of salt.  However, the results that I saw were of interest to me.  When traveling in a straight line on Tribalstack Tropics, I never found myself holding a steady course for more than 10 seconds when using the roll, a move that I would find myself using often to make my travel time as short as possible (given to the player before entering Level 1).  Utilizing the fastest method available to the player for this measurement is important because, as previously quoted, the spacing of objects and it's effect of game feel is dependent largely on the speed at which the avatar can move.

 
Compare that time of 10 seconds in Tribalstack Tropics to the length of time it took me to traverse the lake in the image above (in Glitterglaze Glacier), which took approximately 45 seconds (accomplished by frequently double jumping out of the water to cover ground a bit faster than by swimming).  There are a few scattered jellyfish lurking beneath the surface, but if you are simply trying to cross the map, they won't give you much trouble.  There is, of course, an icy path of floating platforms that you could use to cross.  But providing only one pathway across a vast void like this one is too linear of a solution.

What should be used to fill these voids?  Typically, the types of objects that populate Yooka-Laylee's game space are collectibles, enemies, or platforms (not really counting the NPCs and their challenges).  The combination of these items strike an interesting balance of "pleasure seeking" and "pain avoidance"; collecting items provides the player with a small reward, usually in the form of fanfare or a point increase, while platforms or enemies can break rhythm and flow, encouraging the player to further master their control and timing of actions the avatar can take.

Tribalstack Tropics offered a near-perfect balance between these two sensations in a space that felt open with a lot of room for maneuverability.   The important thing to note, however, is how the placement of these objects created a space that was neither claustrophobic, nor barren.  Object placement in this manner allows the player to think about the space as a whole and reflect on where they need to go next, while simultaneously focused on the obstacle that lies directly ahead of them.  Going back through the level and trying to find those last few notes doesn't necessarily become much of a chore, as the level of attentiveness required of the player to simply traverse the 3D space is enough to keep them from getting bored, but not too much to be a burden.


Playtonic did break some of the conventions that we saw in the free-range platformer of the 1990's, however.  In the above image, you can see that the placement of quills, the more common collectible item, isn't reserved specifically to platforms or paths that players are expected to take.  They are often found obstacles that might usually be treated as walls rather than platforms; something to simply avoid rather than climb.  This, in a way, breaks the conventional boundary and gives the player freedom to explore every inch of the world no matter the terrain.  We'll look at the effect this has on boundaries with a bit more detail later on.

By using quills to suggest this level of freedom, however, these "common" collectibles become hard to find rather quickly.  With a set number of quills established for every stage (excluding Hivory Towers), they can feel pretty scarce when you're sticking to the path.

Low-Level Context


Low-level context is concerned with the avatar's personal space; the way in which the player can interact with object and collision code.  In his analysis of Super Mario 64, Swink mostly addresses how Mario interacts with the ground under his feet.  Designers used steep inclines to make Mario lose his footing and slide, which is applied both as a way to mark boundaries and as a new way to control his movements.  Once more, the emphasis on this kind of interaction makes sense in a game where movement is key, and the challenges in Yooka-Laylee are designed with this in mind.

Momentum and friction are both simulated in a logical manner (at least, with regards to interactions with common terrain and platforms).  Slopes and inclines get a little weird, however, as what is defined as "slippery" is not based purely on steepness.  Some slopes of similar degree can be walked on, some require rolling, and some require rolling plus a power-up, like honey.  There really isn't much of a way to determine which of these are necessary until you simply attempt the climb.  Slippery slopes in SM64 were based purely on the degree of steepness, which the player could judge after becoming acclimated to the game's physical rules.


The "Reptile Role" is a move considered to be the successor to the "Talon Trot" in Banjo-Kazooie, with a bit more speed and the ability to knock back enemies that come into contact.  Both of these moves are acquired very early in the game, and are often relied on by the average player to traverse through a course quickly.  While Talon Trot was quick, it was also precise in control; the player had the ability to make sharp turns and stop on a dime, while climbing nearly any surface.  Reptile Roll sought to balance this by countering the aforementioned speed boost and knock-back capability with controls that felt a little more slippery and hard to control with precision.  By creating a mode of transport in this manner, Playtonic has ensured versatility in game play, so that no single moveset is favored.

Boundary Play


Swink sees these extreme angles in SM64 as a soft boundary, a "gentle negative reinforcement with a clear, physical relationship".  In other words, it's a not-so-punishing way of saying "turn around".  Boundaries in Y-L are treated quite differently.  In Tribalstack Tropics, they almost don't exist whatsoever.  As I mentioned in the section regarding Low-Level Context, climbing objects that aren't platforms is encouraged (boulders in this case), and the edges of the world are marked with those objects.

While this instills a feeling of openness to the stage, playing with these boundaries can easily result in death-by-void, and even if you are skilled enough to traverse the outermost edges of the map, you'll discover that there isn't really much waiting for you.

Moodymaze Marsh and Galleon Galaxy, on the other hand, employ the old "floor is lava" technique, where rather than falling to your death, the player is given a chance to turn tail and get back to the intended play space.  While this is less harsh of a punishment, the Marsh's lowered draw distance would consistently entice me into thinking that there might be more ahead of me, if I could find a way across this hazardous waste.

The above is, of course, only addressing boundaries on the horizontal axis.  Boundaries below the player either end at your feet, or somewhere below the course, watching for you to fall into the void.  A boundary above you is usually just a ceiling that stops any further ascension.  The ceiling of each stage in Y-L is set pretty high to accommodate the grand scale of structures and landmarks (see previous screenshots), and with the end-game ability of Flappy Flight, these generous barriers offer some interesting perches (again, see previous screenshots).  

The one and only place where vertical travel feels more limited is in Glitterglaze Glacier; the ceiling is noticeably lower, and forbids the player to reach vantage points of interest.  Not only that, but rather than being a solid wall of collision, the ceiling will actually push the avatar back down, a kind of jarring way of saying "No Trespassing".  I couldn't help but get the feeling that this particular ceiling, combined with the variance in low-level context between the avatar and other objects, demonstrated a lack on uniformity between rule sets between stages.


Conclusion


Upon reviewing everything I discussed in the above paragraphs, it feels more like I just spent a lot of time picking apart the things I didn't like about Yooka-Laylee rather than answer the question I posed in the introduction.  Truthfully, many of these details are just that - small details of a bigger picture.  Playtonic sought to revive the 3D Platforming genre by adapting the original formula into modern-day gaming, and they accomplished this by creating captivating worlds that were massive and packed with things to do.  They weren't perfect, as you will no doubt see in reviews on Metacritic, but they were a good starting point.  Tribalstack Tropics specifically will go down as one of the most entertaining virtual spaces I have explored in a long while.

This research was a really fascinating crash course on 3D level design, especially for games where exploration is key.  I aim to use what I learned from Schell and Swink in my future analysis of games and in my own designs as well.  For now, it's time to spend a little less studying design and actually get my hands dirty.

No comments:

Post a Comment