Introduction: Angle Jungle is an award winning puzzle game built by a team of students at Carnegie Mellon University’s Entertainment Technology Center in 15 weeks for Pennsylvania’s Intermediate Unit 1. Angle Jungle has value to first graders and above, its primary purpose though is as a supplement for 4th to 6th graders learning basic geometry.
Platform: iOS | Time: 15 weeks | Role: Game Designer | Team Size: 4
Design Goal: The goal of the project was to achieve the following transformations in our target demographic:
Primary Transformation: Build familiarity with the angle by having players solve puzzles that use a mechanic that encodes the numeric and spatial representations of angles
Introduce positive and negative angles
Introduce clockwise and anticlockwise rotation
Introduce angles greater than 180 degrees
Build familiarity with the protractor tool
Design Challenges: We faced a number of design challenges during this project:
Protractor tool introduction
Finding an mechanic which made angles essential
Crafting fun and engaging puzzles
Crafting additional sources of motivation
My Contributions: As the game designer on the project I took the lead on directing our creative efforts. My efforts helped create a well received, fun, and engaging experience which made a good attempt to achieve our transformational goals. Other areas I made significant contributions in were:
An ideation process that created the main mechanic of the game
In this article I will chronicle my design process in creating Angle Jungle an award winning transformational puzzle. Then how I went creating the puzzles within the experience, and finally lessons learned.
Angle Jungle is an award winning educational puzzle game for fourth to sixth graders studying geometry. At the start of development our requirements were up in the air. Following discussions with our client we settled on the following objectives:
From our paper prototypes, we choose to refine two based on feedback.
In parallel we began the process of creating digital prototypes based off these paper prototypes.
Our breakthrough moment came when Jesse Schell, a faculty member at the ETC, posed to us that though these games used angles, both could be played without thinking about angles. We needed to make an angles essential experience. This priceless notion lead us to create Angle Jungle’s progenitor which we called Treasure Hunter.
Treasure Hunters mechanic encoded the relationship between the numeric and spatial representation of angles. This was achieved by having players use numeric representations to create spatial representations in-order to solve a puzzle. We believed this embodied a system where angles were essential. We then began refining Treasure Hunter.
After positive feedback from playtesting we next created a digital prototype.
In the above video players slot numeric values into a beam maker which creates a spatial value. A certain spatial value is required to hit an objective to solve a puzzle and receive treasure. This digital prototype then went through many more iterations.
At this point in development we had the foundations for an experience. What was needed next was to design that experience.
How does one go about creating an experience? There are infinite ways, but we began with considering the difficulty curve within our experience.
The above graph is an abstract difficulty curve which displays a sequence of tense and release cycles of increasing difficulty. This curve would form the underlying foundation of our experience.
With an idea of what we wanted the experience to look like, next we conceptualized the elements within the greater experience. The inspiration for this process came from a number of sources including the learning materials of our target demographic.
Our aim was essentially to gamify our target demographics learning material. We would achieve this through gameplay elements which attempted to capture aspects of the kind of problems they faced in the classroom. These gameplay elements would form the core components of the experience.
Whilst conceptualizing our gameplay elements we also considered the possibility that the puzzle may not be intrinsically motivating enough for players. Therefore we created two additional supporting motivational factors.
A gender-neutral character that needed assistance (inspired by Jesse Schell’s Lens of Help). Given the use of supporting characters in educational experiences is common, and there exists research on the potential beneficial effects for players. We hoped this would augment learning within our experience.
In addition we created The Cabin. The Cabin would contain rewards in the form of treasures and trophies. The Cabin would act as motivational element by creating Golden Expectations (expectation of rewards) through the aesthetic use of empty shelves as well as serve as a measure of game progress.
We also recognized the need to space out our rewards for better impact. We therefore arranged rewards into evenly spaced intervals.
Together these pieces could further flesh out the difficulty curve of our experience. The peaks of our difficulty curve would now commonly correspond to the introduction of gameplay elements, and the dips would be periods of rest at The Cabin.
The experience needed more though, it cried out for substance in the form of puzzle content.
Transformational Puzzle Complexity
With a high-level view, and the fundamental elements of the experience in mind we went about crafting a set of transformational puzzles.
This process resulted in a jumbled pile of puzzles. This was a good first step, but it did not fit the experience structure we wanted. We therefore turned to a mighty tool. The spreadsheet.
The spreadsheet consisted of columns of each gameplay element which we incrementally increased to raise puzzle complexity. This tool complemented the design process as we created more puzzles based on these new complexity constraints.
Two additional considerations came to mind during this process:
Include drops in puzzle complexity when introducing new gameplay elements to allow for more effective tutorials.
Have the majority of learning occur early when complexity is low.
The result of this work was a structure of thirty levels which we then playtested.
Although initial playtests were largely positive they revealed two design issues:
Lack of Angle Diversity – High occurrence totals of fewer number of angle values in the total experience meant a lesser exposure to different angle values.
One Gem Solutions – Solutions which required only one angle gem on more complex levels meant less interaction with different angle values.
Both issues were detrimental to our goal of building familiarity with the angle system. Therefore, two methods of analysis were used to solve these issues:
Angle Distribution Analysis – Counts of each angle value used.
Angle Solution Analysis – A comparison of solution angles against angle values used.
These methods revealed a number of such ‘issue’ levels.
The result of iteratively applying this analysis was that both the complexity and angle diversity was maintained and improved. This ultimately meant a better attempt at achieving our transformational goal.
At the end of the project we ended up with a concrete primary transformational objective, and several secondary transformational objectives.
Build familiarity with the angle system by having players practice solving puzzles using a mechanic that has an encoded relationship between the numeric and spatial representations of angles.
Sharon Carver – ‘The actual angle choices at the various levels and the angle meter seemed to work well and COULD promote learning of the concepts and spatial relations of angles, as long as students don’t game the system’.
In addition to our primary transformational objective we took the opportunity to introduce a number of secondary transformational objectives in manners that were natural extensions of the core experience (providing the experience with more puzzle content).
Protractor Tool Usage
To solve a puzzle, players had to work out the angle that was required to be made. This was difficult for some playtesters and therefore provided a natural opportunity to introduce a protractor scaffolding tool.
By making this tool available we built in the protractor in a manner that was of a natural clear benefit to our players. We hoped by doing so to build familiarity and appreciation of the tool by creating a puzzle environment where it was undoubtedly helpful. Playtesting showed that this strategy ‘seemed’ to work.
Sharon Carver – ‘I especially like the meter that shows the full 360 degrees while the player is working on selecting angles. It would definitely be worth testing the impact’
Introduce both anticlockwise and clockwise rotation, and angle addition and subtraction.
Angles Above 180
Expose students to angles greater than 180 degrees.
Whilst exposing students to our core mechanic (an encoding between the numeric and spatial representation of angles), initial levels would allow brute force approaches to be rewarded in order to draw in the player with easy rewards.
Allowing for such ‘brute force’ (choices made without solid reasoning) approaches, resulted in the following criticism being raised:
What if players are not doing the thinking you want?
In the defense of brute force, we responded with the following counter points:
Absolute mindless play is rare, so since the use of numeric angle values are essential even with a brute force approach, players are likely to at least reason about this aspect of the game.
Supporting brute force approaches makes the experience more accessible (we had first graders reach level 22 with help!).
Brute force approaches are only reasonably satisfying in low complexity puzzles (playtesters who solely practiced a brute force approach experienced frustration on more complex puzzles).
Most importantly though, we admitted that when complexity was low players would not have to think ‘much’. This was intentional. The experience allowed it for a deeper purpose.
We intended to combine that brute force motivation together with puzzle complexity as a transformative tool to incentivize a ‘logical’ approach. As puzzle complexity slowly increased the experience would naturally create skill appropriate ‘teachable moments’ for teachers to capitalize on.
The results of this process created an experience that contained:
Suitable learning and puzzle complexity curves
An appropriate pattern of tense and release
Appropriately interspersed rewards
An exposure to a wide variety of angle values
A mechanic where angles were essential (encoded the relationship between spatial and numeric representations of angles)
As part of the educational game project my team was working on we were required to build a reward system. This system took the form of a trophy room which would display trophies that players had earned. After playtesting though we found we had created an expectation for treasure which we were not fulfilling. The following is a gameplay video where our players would collect treasure chests at the end of each level.
So in order to fulfill this expectation we created additional art assets which we would use to fill up our empty room. We faced a dilemma in this regard. We did not want to force players to see treasure added to the room at the end of every level. This would be far too disruptive to the game experience. So how does one fulfill the expectation of reward without forcibly having the player see the reward appear?
Well one thing helped us in this regard. We already designed fixed reward intervals through the trophy system which forced players to go to the trophy room and observe the new trophy being added to the trophy room.
In our experience we had periods of fixed visitation where the player would be guaranteed to be seeing the Trophy Room. Looking at the experience more methodically we were giving trophy’s at the following intervals (we had thirty levels).
One and thirty were absolutely necessary since they began and ended the experience. The others were decided based on difficulty curve which was designed in previous weeks. Again we asked ourselves the question. How does one fulfill the expectation of reward without forcibly having the player see the reward appear?
During Spring break we had the chance to playtest a digital prototype of our game. The game consisted of five puzzles, and the intention of the playtest was to see if our target demographic and client (Colonial School) liked the game, and their thoughts. Feedback from both the teacher, and our target demographic was as follows:
Kids like the game
Thought it was easy, wanted more challenge
Understood the mechanic immediately
Completed the game within 5 minutes
When asked about characters they wanted they mentioned all kinds of animals they saw in the jungle
Again asked for a wrestler
Had no major complaints about art or mechanic or story
One kid wanted dragons
One kid recognized it was a maths game but kept playing
Asked for more levels!
Teacher liked the game
Said reverse angle gems (move in opposite direction) would be fine but only on advanced levels
Wanted some source of competition so star rating system should have a total for students to compete against each other
Teacher said using games to teach angle of shapes would be fine
Teacher said students are not taught physics at their level (leaving physics out is a good idea)
This week was spent working on UX changes as well as polish to the game.
A number of UX changes were made .
One Gem Solutions
One gem solutions are The changes this work consisted of solving a number of one gem solutions that appeared during playtesting.
Changed protractor tool tutorial to an earlier level, then introduced it again in a later level to hopefully increase the probability that players will use it.
During our playtest it was revealed that slotting and removing a gem constantly could be used as a cheat to beat a level. We solved this issue technically by having a check for slotting, and not allowing a win to occur if a slot had occur within sometime.
We reconsidered the flow of the first time play experience. Initially the first time players played the game they start directly at level one. The intention behind this was done in attempt to get players attention by showing them the most interesting thing first. This was changed to start with the map first because:
It was our actual homepage.
Many other games followed a standard of showing the map first rather than introducing the gameplay.
At the start of week twelve polishing the game was on the forefront of our minds. In this regard, design wise we continued to struggle with small, but vitally important decisions namely considering the visual representation of angles during gameplay and the introduction our scaffolding tool (the protractor from week eleven).
We met with Jessica Hammer on Thursday to get a perspective on what we had done and the issues facing us. She told us the following:
Clarify our learning goals and sort it out into a table
Make red and blue gems beam movement uniform, so red always goes anticlockwise, and blue always goes clockwise
Reconsider the visual representation of clockwise movements
Interest in protractor tool introduction and suggested we put it on level three where we introduce no new things and so cognitive load is not high
Jesse to the Rescue!
Following this we met with Jesse Schell on the evening of the same day. Being the masterful designer he is, Jesse gave us a suggestion of displaying the spatial representation of the angle.
Jesse’s suggestion was when the beam rotated clockwise, the beam maker would make the full 360 degree representation pop out, and be subtracted from when the beam moved past 0. In the case of the beam rotating anticlockwise the sector would grow as the beam moved anticlockwise.
We implemented this feature, then spent the rest of the week playtesting the levels we had, and weeding out one gem solution angles.
Starting Week 11 we finished creating digital versions of our remaining puzzles. In addition we began working on the various aspects of the game that we presented to our playtesters at the end of Week 10.
We added a map to replace the original level select screen. The new map would serve two functions.
It would display the progression of the game to the player
Create a more visually appealing method of level section
We also implemented a reward system in the form of trophy’s added to ones treasure room after completing a ‘boss level’. We hoped such an addition would add a motivational factor for completing the game.
Later in the week Jesse Schell played the game, and suggested a new way to show treasure room. Instead of having trophys placed on the desk, have shelves arranged in a geometric way with numbers on them to reinforce the central theme of angles. In addition to this we considered including random treasures which we hoped would add a surprise factor.
During Week Ten we prepared designs for the final levels of the game. These levels were in line with the complexity metrics we established during Week 9.
During this process we also documented our puzzles, and their solutions. This document would not only help recreate these puzzles during development, but could be handed off to teachers as a supporting document.
Meanwhile we began preparation for The Entertainment Technology Centers playtest day. This would involve members of our target demographic visiting our project rooms to playtest our game. For this day we came up with a number of questions to ask our playtesters as well as prepared video and screen recording equipment to capture gameplay footage.
On Playtest day we had five groups of playtesters. Each group played the game for approximately fifteen minutes. We then conducted a short interview with them, and found several good insights such as:
They really enjoyed the game, we never had a case of a bored playtester
Even when playtesters got stuck they cried out for help, and we had cases of playtesters working together to solve puzzles
The protractor tool was useful, but since there was no clear tutorial playtesters found it by mistake
Playtesters liked the art, music as well as the treasures we would reward them with
Playtesters didn’t object to the main character, but found certain animations weird
Recently we have been working to create an educational game on angles. Part of that requires designing puzzles that try to provide educational value. The following blog post is a continuation of a look at our process.
The most important part when analyzing our puzzles was first to recognize our puzzle metrics. Initially these metrics were as follows:
Number of slots
Number of gems
We began our first pass using these metrics to craft the thirty puzzles that would form the core structure of our game. The process essentially boiled down to a table of each of these metrics listed in columns. We incrementally increased metrics until key climax moments which we referred to as ‘boss levels’. Following a boss level we dropped the metrics to allow for the introduction of a new system in a simpler environment.
Our first pass at developing the puzzles allowed us to create the initial structure of the experience. On further examination, points three and four actually had more depth to them. We broke these points into each and every gem value. This additional depth warranted further analysis.
We then went about constructing a meaningful method of presenting what we called ‘angle distribution’. Using this we mapped out each and every gem per level. This method of analysis revealed several levels that were problematic for different reasons such as:
High angle overlap
Had no garbage
Levels that were similarly structured
These key points conflicted with our main educational objective of improving familiarity with both numeric and visual representations of angles. As for one having a large degree of similar angles meant that the exposure to different angle values in the 360 angle system was lower. So for our second pass we went about redesigning certain levels adding in garbage, and choosing angle gems carefully to avoid overlap.
On making a third pass at the we again found a problem. Our third pass took the form of playing the levels. What we found was some gems were included that were direct solutions to problems in hard puzzles.
We needed to weed out as though it is good that players are able to discern such a solution, we felt that doing so would mean engaging less with the angle gems in the level as several other gems were left out entirely in the solution. Thus we weeded such scenarios out during our third pass.
Essentially the process boiled down to a number of steps:
Carefully study the components within our structure
Extrapolate areas for further fine grained analysis
Develop a tool for analysis
Apply the tool
Identify and address problem areas
Replay the experience
Using this process we iteratively analyzed our puzzles redesigning when necessary to ensure levels had particular solutions to problems with minimal overlap. Now with a clear design process, all thats left to do is playtest and hope the design worked!
At the beginning of the week 9 we had our halves presentation. Following this we met Jesse Schell on Tuesday, and presented our thoughts on how we would go about designing our puzzles. His suggestion was simple.
JUST MAKE PUZZLES. Worry about the details later.
So that is what we did.
The inspiration for our puzzles came from a combination of two sources:
The teaching material that our client used
A map of element complexity against time
The process of considering elemental complexity began with a consideration for the interest curve of the experience. Essentially we wanted an initial large peak then a period of rest, followed by ascending peaks with rests until a climax at the end.
When designing puzzles Level Design for Games by Phil Cosuggested listing the elements of a game, and systematically designing puzzles with incrementally harder arrangements of elements.
In our case we intended to use the elements to increase complexity, but explore fundamentally the same (problems related to the 360 angle system). The elements of our game were:
Receivers & Obstacles
With these elements we create a table of level against elements, and incrementally increased the number of elements. When a new element was introduced we would drop other elements to lower the difficulty experience for players to more clearly grasp the new element.