This site showcases the thesis capstone projects for the Full Sail Mobile Gaming Master of Science program. Students completing the program post their end of program project self evaluation here examining what went right and what went wrong during production.

The site provides examples of all completed projects, without regard to the quality of work. Final faculty evaluation of your project is separate from your postmortem. It is a place to share student work and start dialogue with faculty about completed and upcoming projects.

If you are adding a postmortem for a completed project to this blog, please do your best to provide a meaningful meta-level evaluation of your project. This helps students currently in the program have a better understanding of the critical points related to independent production, game development and design and project management. The template for the blog content and instructions can be found in the first post from July 2014.

Thank You,
MGMS Faculty

Sunday, April 26, 2015

Capstone Game Post Mortem: Treasure Craze

Capstone Game Post Mortem: Treasure Craze

Game Summary:
Game App Icon

Treasure Craze

First Person Multiplayer Melee Combat and Item Collection.

All Microsoft Windows 8.1 and later devices. (Phones, Tablets and Desktops)

Revenue model
Treasure Craze (Norri, 2015) is completely free in the Capstone edition, however in-app purchases were planned for the final retail version. Purchases would include upgraded flashlights & shovels.

Development tools/Language
Treasure Craze was built completely from scratch in C++ using a custom 3D engine built specifically for the project. All software was authored and tested using Visual Studio 2013. Game artwork was created and exported from Adobe Photoshop and Autodesk Maya 3D Modeling & Animation software. The only APIs used were low level hardware interfaces such as D3D11, XAudio2 and DatagramSocket.  

Game audience
Teens and Young Adults 13 – 35 who like competitive games. Treasure Craze is designed to appeal to players who fall under both “Achievers” & “Killers” under Bartle's player taxonomy (Bartle, 1996). The game also targets players who enjoy PC & Console multiplayer combat games but want a more streamlined experience for mobile devices.

Lari Norri: Lead 3D/Network Engineer, Game Designer, Gameplay Programmer.
Joshua Fox: Lead 3D Character and Level Modeler, 2D Texture Artist, 3D Character Animations.  

Treasure Craze is copyright 7thGate Software, LLC. (Norri, 2015)

Sound Bite
Get ready for a Rootin' Tootin', Doubloon Lootin' good time!

Executive Summary
Treasure Craze is a fast paced first person treasure hunting game where you battle other prospectors for loot & glory! Use the in-game compass to navigate a scary maze and find shiny doubloons. But be on the lookout for other prospectors who want the treasure all to themselves! 

The final design of Treasure Craze was influenced by many different games and experiences from my youth. I enjoyed playing many competitive first-person shooter games like Doom and Unreal Tournament. I also loved to play role-playing games with a particular fondness for dungeon crawlers. As a kid I used to love to explore the woods in North Carolina. I chose a “gold rush” era art style to try and capture that sense of adventure.  

Capstone Scope
While the full retail game called for internet play and in-app purchases, the actual capstone portion focused on just getting the fundamental technology and gameplay operational. This included: Random level generation, Player movement, Environment collision, Player matchmaking, Networking, 3D Rendering, Sound Effects and Music. Besides the critical features, I was actually able to get in two “B” level features on my list. These included “shovel combat” and “pre & post match screens”. Last minute additions that were never planned but I was able to cram in are: Active game player drop-ins, Player rankings and a compass indicator for the Gem and Ladder. These were added primarily based on faculty feedback.   

In a perfect world I would have had the time needed to add some additional features and polish the graphics of the game to a much higher level. Visually I am happy with the results, but given a little more time I could have added normal & specular mapping to the game which would have bumped the graphics to the next level. Another feature which was on my “B” list was doubloon digging. Since you have a shovel it would have been nice to use it for something other than just hitting enemy players. I would have implemented this feature as a sparkling crevice in a wall that you could smack to yield even more treasure. The last item that I wish could have made it in, was a tutorial level that taught player the ropes of the game. Even though the game is fairly simple, additional help could have made things go smoother for new players.

The Critique: What went right…

Design & Aesthetics
Nearly three months into the program (after the prototyping class) I already had a very strong idea of what game I wanted to make. This meant that very early I was already documenting features, mocking up user interfaces and creating pre-production sketches. This proved immensely helpful moving forward as I could focus directly on how I would be putting the game together and worry less on design details as they had already been addressed. Access to the Weststar Music and Sound FX libraries (Sound Ideas, 2014) turned out to be invaluable. It was easy to find music and effects that complemented my game almost as if they had been made specifically for it. Last but not least are the overall visuals of the game. Writing my own engine allowed me to extract significant performance from mobile devices and to fine tune the game’s atmosphere directly. Another huge win came in terms of art & animation content but more on that later…

Project Management
Using Underdog was simple but effective. I was actually able to integrate GIT directly into Visual Studio 2013 which made commits and pull-downs a total breeze. The task and time management systems on Underdog were fairly straightforward if a bit barebones. I didn't have any problems using them though I do prefer more full-featured project management solutions such as “FogBugz” (Fog Creek Software, 2015). Even though 99% of the time source control really didn't do much for me, it’s always that 1% of the time where it is indispensable. For me it was when I had introduced a severe networking bug that caused constant disconnects in my custom network protocol written on top of UDP. Being able to “diff” with previous versions of my code ultimately led to finding where the problem had been introduced. Events like this are why source control should be considered “non-negotiable” on projects with any sort of reasonable scope.    

Towards the end of development writing my own 3D engine really started to pay off. It became very easy to tweak systems or add new features since there was absolutely no confusion about how something worked. I was able to keep the engine/system code and the code driving the game fairly isolated from one another. This will come in handy if I want to develop future games using the same underlying technology. As you will see in the “what went wrong” section, creating a 3D engine is not a decision to make lightly. However, it does allow for focusing on just the essential technology that you need for your game. The other major benefit is that you end up with a game that has a very unique feel due to the custom tech driving it. Last but not least, there is no “learning curve” to things you make for yourself.

Play sessions and user-analysis ended up being the two most beneficial contributors to the Treasure Craze project. Multiple core gameplay revisions and feature additions are the results of direct player feedback. Even though a pre-multiplayer build was used during user-analysis it still provided significantly valuable information. Due to this feedback, I changed fundamental player controls and scheduled a compass feature for the game. Towards the end of the project, multi-player game sessions served to reveal game breaking bugs in the matchmaking and networking sub-systems. These sessions also provided helpful user suggestions that could further enhance game-play such as ranking order and in-game notifications. My Advice: Testing is seriously useful, do it early do it often. 

Business Model
Though the capstone version of Treasure Craze is a free game with no monetization, this is a good spot to discuss what the full game would have. In the Windows Store the game would be a free download but contain a way to purchase extra doubloons at an in-game bank. These doubloons can then be used to purchase upgraded shovels & flashlights as well as single-use items used to help win a match. Of course these doubloons can also be earned just by playing, but this offers a faster way for impatient players.

I really have to hand it to my lead artist Joshua Fox! When I started this project I was expecting to simply cobble together whatever artwork I could scrounge from the internet. Josh is a good friend and colleague of mine who stepped up to the plate. He not only offered to create & animate my main character (my biggest concern), but he also provided beautiful level textures for my maze! I am beyond thrilled with what he came up with. He was also nice enough to put up with me constantly asking for tweaks and changes. I owe him a great debt of gratitude! (And some beers!)

The Critique: What went wrong…

Design & Aesthetics
Networking in general caused a lot of headaches throughout the project, but even something technical like this can trace some problems back to a few early design issues. Treasure Craze was designed with an emphasis on the mobile user, therefore the matchmaking system was designed to be one-click and go. This idea is great if you are playing on the internet and just want anybody to play with. Unfortunately the scope of the capstone led me to only support play over a local network. This ended up clashing heavily during play testing when we had trouble getting all the players into one game. The traditional lobby system would have made a great deal more sense for a LAN only game. At the end of the project Nick Penny suggested adding a drop-in capability to the game. After this was added, it significantly improved the ability for players to connect to a single game.

Project Management
Although most of my milestones went off without a hitch, I did end up falling behind briefly on one of my network milestones and one of my animation milestones. Though I had budgeted a week for each of these tasks, I ended up running into some issues beyond my control. The animation format I chose (*.CMO) ended up having some fairly significant bugs & limitations that had to be discovered through trial and error. On the networking side I chose to use the brand new Microsoft “DatagramSocket” class to do my low-level networking. It too turned out to have a serious glitch that I thankfully found a workaround for online. In both cases I had to hustle to make up time lost due to these setbacks.

On the scheduling side, “Sprints” seemed to work well enough for most of the project. However, I found them somewhat lacking in certain areas. For example: It would have been nice to have a pre-production milestone as well as an “Alpha” & “Beta” submission that were put through more rigorous faculty testing. I also ended up with a giant pile of minor tweaks and fixes that got pushed back because there was no time scheduled to address them. I ended up having so many that I struggled the last month to address as many of them as I could. Having scheduled points throughout development where “polishing is a priority” might have helped alleviate this.  

What went right: Making your own 3D engine.
What went wrong: Making your own 3D engine.

Although the end product turned out well, initial development was at times seriously overwhelming. I knew writing the 3D graphics code would not be an issue (I already knew how). However, I had never before used a networking system let-alone built a new one from scratch! I ended finding some really great resources and articles online that taught me the fundamentals of UDP networking, but it was still a very real struggle to create something usable & reliable enough to build a game around.

By choosing to forego a pre-built engine, I started my project with massive technology debt. Before I could even start thinking about making an actual game I had to write code that would load my assets and draw them efficiently among many other things. I enjoy this type of work, but I certainly wouldn't recommend it for anyone who simply wants to get right to making a game. If you’re wondering why anyone would be crazy enough to do this when things such as Unity & Unreal exist, here are a few reasons:

1. It can be a great learning experience. (I learned all about networking)
2. I didn't have any problems understanding how to use my own technology. (Obviously)
3. Its free! Both now and later if you become successful. (No licensing fees or royalties)
Of course there are significant downsides as well, but if you can make your own tech it’s not necessarily a bad idea to do so as long as you know what you’re getting into.

As noted earlier testing overall was far more of a plus than a minus. If there was any downside to it, it is that I didn’t get started early enough! Because I wrote my own tech, networked gameplay was not available until late in the project. This meant I could not gather large scale gameplay data until things were coming to a close. Had I been able to do this earlier I may have been able to nip some of the early design issues in the bud before they became more serious problems. Although Unit Testing and A/B Testing were interesting ideas, neither ended up benefitting this project. In retrospect I would have preferred to use that time to do another user-analysis study.  

Networking is tough, do not add it lightly to your project. Even if you have an engine that does most of the heavy lifting, it can still be difficult to detect and debug errors that are caused by the asynchronous nature of game logic based on network events. I wish I had added a logging system to my engine, it would have helped me track and record problems more effectively. I did have output logging in my compiler, but that was of no use once I was running large scale tests with multiple players. An error log is also useful once your game is out in the field being tested on a variety of hardware configurations!


Overall I am pretty thrilled with what I was able to produce in roughly three to four months. There are still quite a few rough edges, but that is to be expected for what is supposed to be an early preview of a full-game. I had to make quite few trade-offs during the development to focus on what I felt really mattered. I really wanted to have the game be a fun experience, something that you could play with other people and have a good time. I think I got pretty close to that target. To do so, I did end up trading time I had wanted to spend polishing the graphics or adding an in-game tutorial among other things.

In the end though, I am glad I made those difficult trades. I can always bring those things back into focus in the future. I do want to release the game on the Windows Store, it would be great to do so soon after the Windows 10 roll out. That being said, there is still a great deal of work to be done. I will need to re-evaluate my final design and tone down its scope if I want to make release a reality. I have always wanted to really hunker down and build a game from the ground up. I am very grateful to have had the opportunity to do so.


Bartle, R. (April, 1996) HEARTS, CLUBS, DIAMONDS, SPADES: PLAYERS WHO SUIT MUDS Retrieved From http://mud.co.uk/richard/hcds.htm#1

Fog Creek Software. (2015). FogBugz [Web Application]. New York: Fog Creek Software.

Norri, L. (2015). Treasure Craze ( [Windows video game]. Florida: 7th Gate Software LLC.

Sound Ideas (2014) Sub-site of Westar Music Catalog. Retrieved from http://soundideas.sourceaudio.com/