     |
 |
 |
 | | | | |
 |
|||
 |
||||||
          |
 |
 |
||||
|
Postmortem: Naked Sky Entertainment's RoboBlitz  Underestimating the Trouble with New Technology
Building a game based entirely on physics is difficult. Building a fun game based entirely on physics is OMGWTFBBQ difficult! The problem we encountered was that reality just wasn't very entertaining, so we were constantly balancing between "real" physics and "fun" physics. One example is the way Blitz drives – it took weeks to balance the programming and tweak the parameters so that he'd drive as smoothly as he does now. He still doesn't move around like Mario, but in a game where everything is physical, that's just not possible. Blitz's grabby arms were another sore spot. We re-wrote the code 14 times until we came up with something that was user-friendly and enjoyable. The key there came down to building several A.I. heuristics into Blitz's grabby routine. He'll actually auto-adjust his alignment to pick things up better, and if something is too low for him, he'll bend down just the right amount to pick it up. We decided to implement this feature because it’s just not fun for the player to have to do everything manually. We believe using A.I. to drive physical animation is going to be the future of procedurally animated games, but we didn't realize it until late in the development process and it cost us quite some time.
On top of all that, building a physical simulation using an engine still in development made things a whole lot more challenging. We began development using version 2.3 of AGEIA's PhysX tool, and we found that a lot of their features just weren't working the way we needed them to – usually because no one else was using those features in their games yet. AGEIA was responsive to our needs and tried hard to get things functioning the way we needed. However, every time a new build of the engine came out, we had to re-adjust and tweak all of our old settings, and work around any regression bugs in the betas. All the unexpected iteration work hurt us when trying to stay on schedule. Ran Out of MoneyCash flow, cash flow, cash gone! This is not a recommended scenario for any game developer, or any business for that matter. Unfortunately, this is what happened to us. The game was delayed by multiple months due to all the production and technical hiccups mentioned above. Because we weren’t willing to put out an inferior product, we just kept going until we were satisfied with the quality. This, of course, led to an empty bank account and an over-stressed management. Fortunately, we were able to borrow some money from our family members and close friends. Even so, calling up your ex-girlfriends for $20K is not something you’d want to go through, unless your skin is as thick as ours.
To avoid or at least minimize such a scenario, the rule is to always raise more money than you think you need. Trust us, you’ll need it, and more. We raised several hundred thousand dollars to fund RoboBlitz, which even included a significant buffer (or so we thought), but we still ended up needing more to take the game to completion. |
|||||||||||||||
Features
Previous
|
|
|