4. Naïve Multiplayer Expectations
Sure, we knew that mean people would tear other players' kingdoms down, but we treated it like an abstract theoretical problem that didn't need a lot of attention. Hosts can kick other players, games can be saved and reloaded, anything broken can be rebuilt, and, after all, people will mostly be playing in private games with their friends anyway, right?
In practice, kicking other players is slow and not obvious. Expecting an ongoing game to be stopped so the host can reload it is quite unreasonable.
Broken banner towers cannot be rebuilt, and other repairs are a pain when components are broken all the way down to their resources.
It also turns out there are a ton of people playing in public games, especially when they're looking for their banner tower-related achievements.
A simple system of controlling who has permission to break things would have gone a long way toward helping with this very painful problem. Of all the things we wish we had done differently, this probably tops the list.
5. Small Developer Blues
Our pockets are not deep. We had only the money we had saved up and additional cash borrowed from friends to make this game. Budget pressure meant cutting a few corners here and there, not fixing a lot of things we would loved to have fixed at the end, and cutting features we would loved to have added, like local multiplayer.
On top of this, we need to keep everyone at the company gainfully employed, so as team members came off other projects we had them help out with this one even when the timing was wrong.
Designers were working on particle effects before the game-specific technology really let them see what the results would look like. Artists were doing production work when the leads weren't available to work with them on direction. All of this meant less efficiency and more pressure on the pocketbook.
We went over budget and way over schedule, risking damage to the company. As a consequence of playtesting and an evolving design, we felt there were certain features that had to be included, and felt it was critical to scrape up the resources necessary and delay release to get them in. We don't regret this decision, but we certainly will consider these lessons when planning future budgets.
That's about it. I believe we can directly apply every one of these "what went wrong" lessons to our next project, and build on the things that seem to be going right.
The experience of making and releasing this game, even with the bad moments mixed in with the good, ranks among the most interesting and rewarding things I've done with my career so far. Life is good. This concludes our therapy session.
Release Date: Nov 19, 2008
Platform: Xbox Live Arcade
Number of Full-Time Developers: 3 -- 12 depending on stage
Number of Contractors: about a dozen, including audio, QA, and localization
Length of Development: 18 months
Development Software: Visual Studio, Photoshop, 3DS Max, Excel
Technology: NinjaBee's Wraith Engine