Definitely our biggest challenge was a tight timeline and a small team. This meant we didn't have any room for mistakes. Any significant re-design would likely have caused us to miss our critical deadline. The way we overcame this challenge was through good planning and process. For example, the first thing we did was identify risky elements and quickly find solutions or take another approach.

Another big challenge was designing really large environments, something the Unreal engine wasn't necessarily designed for. We are pretty sure that Desert Thunder has the largest environments of any Unreal game. It came down to optimization. We had to balance the size of the environment with the number of enemies, and still have decent performance.

The next big challenge was vehicles physics. At the time, the out of the box vehicle physics solution was very computationally expensive. We could only get a few vehicles on the screen at a time on our test machine without the frame rate bottoming out. This was unacceptable, so we had to scrap the system and create our own. In the end, we were able to pull it off.