|1The Lost Spires Post-Mortem
In fall 2006 I decided to make a mod showcasing my design/directorial vision to prospective employers. There was also a need to prove, to myself, that I could make games as well as I thought I could. Browsing around the Bethesda Oblivion mods forum I spotted a thread asking for more quest mods. I immediately understood that this was the way to go for making a great mod. I didn’t merely want a new horse armor expansion, but a self-contained adventure that could actually be played-through and enjoyed. I wanted custom dungeons, a satisfying story, and cool loot. Being a Morrowind fan, I though Cyrodiil lacked Morrowind’s mysticism and romanticism. Morrowind featured obscure environments that made me feel as though I were in another world. Oblivion was lacking in this department; it’s white caves, and bland exteriors made me feel as though I was exploring the real world rather than an enchanted, time-forgotten realm.
The theme of this new mod was decisive: ancient magic & lore. The goal was to give Oblivion as taste of Morrowind. Cohesion was also a necessary component— I didn’t want to just build a couple of loose dungeons. I don’t exactly remember how, but the Archeology Guild sprang forth to marry both these aims. It was perfect. The Archeology Guild epitomized both. Every high-level aim was satisfied: the romanticism of obscure artifacts and legends, a hub from which the player could embark on new expeditions, and a group of comrades providing moral support. And it made perfect sense, for in a realm as lore filled, ancient, and mysterious as Tamriel, there must be individuals who study its secrets. Now, if only there was a way for the player to tap in.
Industry vets often say that great ideas are easy to come by, it’s the implementation that’s hard. Maybe so, and while implementation is important, never underestimate the power of a well-though out idea. A truly great idea is rare, and being a good designer requires one to quickly weed out crap ideas. Be selective about which ideas to pursue, precisely because they take so much time and effort to implement. Anything short of a terrific idea — a one in a million idea — doesn’t deserve to get made.
A decision was made, from the get-go, to work solo. And, although I resorted to outside help in the end, this was an excellent decision. Working solo affords great efficiency. No time is lost to debate or idle chatter. There are no communication delays, nor is time wasted on differences of creative vision, nor on quality control. Working solo eliminates the need for a producer.
Originally the entire story set out to serve one purpose, to give the player a unique weapon — the Horn of Zyyr. The Horn had been modeled in Maya several years back as a lightsaber variation for a sci-fi RPG weapon design. In the sci-fi version, the weapon was meant to be wielded by Jedi-type characters who used magical powers to siphon energy through the Horn. The Horn functioned as a heat-pump, sucking energy out of the air at one end (producing a freezing blue blade) while expelling it at the other (producing a red hot hilt). Fond of the model, I decided to port it into 3ds Max and give it a back-story.
The first crude draft of the story was much different than it actually turned out. It all started with a single meteorite. Falling from the sky, this shard crashes somewhere near the player. Witnessing the ka-boom, the player was to come and investigate the area to find a brilliant celestial shard. Not knowing what to make of it, he takes the shard to a gem specialist. This specialist secretly tricks the player into performing a series of tasks that eventually lead to the unleashing of a great evil. This was the basic plot.
Work on the narrative started in December 2006. I wanted a deep plot, with a shocking Starwars-esque revelation. The revelation was arrived at early on, but there were serious doubts as to how the player would react to having been misled throughout the entire story. Would the player feel betrayed by the evil mentor, or by me, the designer? One month of writing resulted in a solid plot -- the backbone of the design document. The key pivot points were set, with just the details left to fill. The result was pleasing — it was dramatic, lore-filled, and engaging. The only doubt I had was not being able to provide a sufficient motive for the mentor to unleash the evil Avatar knowingly. I didn’t want the mentor to truly be “evil”, just to appear so. In the end, no justification was ever created. There’s something philosophical I still don’t understand about “evil” people and how to depict them. For me, this area of writing requires further research.
Having started several game projects in the past, I knew that the greatest hurdle was actually following through. My greatest fear was quitting mid-production due to waning interest. I had recently quit an ambitious space-RPG project due to technical difficulties and failure to appreciate project scope. This would not happen again. Firstly, I regularized workflow. No more late nights at the computer. No more working weekends. Now there was a strict regimen: wake up at 8:30 am every weekday, work, break for lunch at 11:00 am, come back at 11:30, and work until 5:00 pm. No work was allowed past 5:00 pm or on weekends. This was done to prevent burnout. After 5:00 pm, and on weekends, I was to relax as much as possible so as to be maximally productive during work hours the following day. Moreover, 1 hour was set aside each workday to research game development skills and workflow efficiency. Between 4:00 to 5:00 pm, daily, all work gave way to research. Research involved reading books on 3D modeling, texturing, writing, Roman history, graphic design, and even science fiction novels. Sometimes research involved watching 3ds Max and Zbrush tutorials, or listening to GDC design lectures. Without this strict work regimen, Lost Spires could not have been completed. Many times I hated work and thought of quitting. But, tenacity kicked in and kept work continued. I forced myself to put in a full day’s work regardless of how I felt, or how much I didn’t want to work that day. Research time developed vital skills and helped take my mind off the tedium of work. Besides the regimen, research time was the smartest decision made throughout the project. Anyone interested in mastering a subject-matter ought to dedicate time for increasing workflow efficiency.
Production began with basic level layout. The approach was to paint with broad strokes and fill in the fine details iteratively. Base interiors were laid out first. The Archeology Guild was constructed from a vanilla tileset. After this, I realized that the cave retextures would have to be done sooner rather than later. The Primeval Hollow retexture required each tile (over 200) to be edited in 3ds Max, re-exported, and heavily Nifskoped. This was incredibly boring and painful work, however, the initial results looked fantastic, so I was motivated to move on. Designing the Warlock ruins was the most frightening of all tasks. These caves had to actually look like ancient ruins and they had to keep the player’s interest over several levels. I needed a sinister retexture and various pieces of Warlock ruin architecture. Several dark retextures were tried, but didn’t work. It would take several months to finally get something satisfactory. One of the first ideas for the Warlock ruin architecture was the Spire. Originally, spires were small, decorative, column-like elements found inside Warlock ruins, but their potential to serve as landmarks for each Warlock ruin entrance was soon realized. This idea came after reading a book on level design (during my research hour) which praised the use of landmark geometry in a multiplayer map in Halo 2. See, research pays off!
Before even starting to design Warlock architecture, I had decided to create a set of custom Warlock hieroglyphics. From my work on a previous project, I discovered the power of eerie hieroglyphics, and intended to imbue Oblivion with that mystical ambience it lacked. This struck at the heart of the project’s mission statement. An entire day was set aside to design 31 custom glyph shapes in Photoshop. The results were immediately pleasing, and I knew that the right decision was made. Generally speaking, I’ve found it very useful and even fun, to delve head-long into such seemingly small details. Work actually becomes less monotonous when you slow down, forget about the big picture, and just focus on doing one little aspect exceptionally well.
Much of production was a bore. After laying out the interior tiles for all the caves, I started cluttering them with geometry. Slowly throughout this process, custom props like mushrooms for the Primeval Hollow and ruined Warlock columns and crystals were modeled. One day, while working on a large Ayleid Ruin, an idea came to make a true boss-type encounter: the enormous Minotaur. It needed to be a total shock to the player’s system. An oversized Minotaur, a loud gong flurry, tense battle music, and a deep voiceover work well to this end. A day of scripting led to a satisfying boss encounter, which was iterated upon several times later in the project.
Many of the best features in Lost Spire were the result of quick, on-the-spot thinking. The Warlock celestial planes, for example, were though up in the final month of development as a way of spicing up the otherwise uneventful Warlock ruins. They took only a few days to implement but are easily the best part of the game. Likewise, the Archeology Guild artifact display cases came in the final month of development after a suggestion on the LS work-in-progress thread. I had planned to release Lost Spires in early August, but the idea of cluttering the Archeology Guild with mythical artifacts so intriguing that two solid weeks were devoted to modeling and texturing several dozen objects. This idea directly perpetuated a key goal of Lost Spires and seemed warranted. The player needed to walk into the Archeology Guild and feel as though he’d entered a magical space — a place he’d want to call his own. There was also a need to reward the player after each mission and provide incentive to start the next.
I am very grateful to have teamed up with XMS on creature creation. The Avatar was an essential part of the plot and could not have been done without he. After receiving the awesomely rigged Avatar, I started playing around with the final battle. This was late in development, and I knew that rest of the mod, although large in scope, could not truly be epic without a grand final confrontation. A promise was made early on in the LS feature list of a “climatic ending”, and now was the moment to deliver. Slowly, the final battle evolved into something I never foresaw. This is the magic of diligent, moment-to-moment implementation. You forget about the end-goal, you focus on a small sub-component of the entire project as though it were the entire project itself, and you just tinker around until something great results — something to be proud of in-and-of-itself, absent the rest of the project.
Although Lost Spires development was in full-steam by January 2007, I had purposefully waited until July 2007 announce its development on the Bethesda forums. This was an excellent decision. Firstly, it allowed me to concentrate on the work itself, rather than discussing flights of fancy regarding its release. Secondly, I didn’t want to lose steam by discussing LS development. I also didn’t want to reveal screenshots or story so as not to spoil the surprise. Thirdly, I wanted to present LS as an epic mod, with proper teaser material. Doing so meant dedicating lots of time for to website development, which was another great decision.
It was also a good decision to start the PR effort in July, a month prior to release, rather than wait longer. From my Pi release, I knew that publishing a game with no hype or prior advertisement is very dangerous. Without a primed audience, the game will fail to gain traction and slowly settle to the bottom, even if it’s good.
In general, PR relations were good, albeit they did start to eat up time and eat away at productivity. This was worth it, however, because I received moral support, a few good suggestions, and buzz that would be capitalized on upon release. Setting a tentative release date (August 2007) also made me work harder. At this point I was well accustomed to my work regimen and decided to stray from it by working overtime. I stopped the 1 hour research sessions and felt comfortable working weekends and nights. Without the crunch, LS would surely have been delayed for at least another month or two. Generally, I regret having announced the August 2007 release window as well as later having announced the August 27th release date. To make the August 30th release required working a 27 hour day, with no sleep, which negatively impacted quality.
Release was slightly premature, up to 1 week. I had completely underestimated the length of the mod, the number of bugs, the number of typos, the amount of awkward sounding dialogue, and the number of miscellaneous loose ends. On a positive note, the previous month’s PR effort really paid off, generating good buzz and lots of downloads. I underestimated the number of severe bugs, the post-release time needed to fix them, and the need to attend to players’ questions, comments, and compatibility problems. It took approximately one week to incrementally patch LS up to a decently polished state. In my defense, many reported issues would have been near impossible to identify personally, without having released the game. I regret not having organized a closed-beta, which would have smoothed out the release, albeit a belated one.
1What Went Right:
1The original idea to create a deep, lore-filled Archeology Guild
1Establishing a strict work regimen
1Devoting daily work time to research
1Spending a month developing a solid narrative outline
1Understanding the scope of the project
1Designing custom hieroglyphics
1Scripted boss battles
1Devoting 2 weeks to modeling Archeology Guild artifacts
1Devoting several weekends to develop an appealing website
1Properly timed forum announcement
1Working with XMS and The_Shadow
Hosting all files myself, ensuring high download speeds, and monitoring site traffic
1What Went Wrong:
1Failure to play other popular mods and identify possible compatibility issues
1Failure to sufficiently play vanilla Oblivion
1Failure to organize a closed beta
1Discounting voiceovers in pro-production as unfeasible
Pre-mature 1announcement of release window and specific release date
1Not having kept a post-mortem list from day-one
1Underestimation of the time required for debugging and polishing
Number of Principle Developers: 1
Number of Supporting Developers: 10 (mainly for VO work)
Production Span: 8 months
Estimated Development Time: 1,500 hours
Estimated Cost of Project: $50,000
Average Download Rate: 120/ day
Tools Used: Photoshop CS3, 3dsMax 9, 3dsMax 6, Dreamweaver, Flash, Premiere
All-in-all, far more things went right than wrong. The final version ended up with numerous cool features that couldn’t have even been anticipated in pre-production and community reaction has equally surpassed all expectations.