How to build a Splosion Man in six months

Who Was There: Splosion Man lead programmer Mike Henry and lead designer Sean Riley dissected what went right and wrong during the six-month development cycle of their latest Xbox Live Arcade hit.

What They Talked About: How does one perform a postmortem on a corpse that has been blown up thousands of times?

Riley began the half-hour-long session that would answer that question by going over what went right. With a development schedule that tight, the focus for the 12-developer team was to have as little wasted work as possible. Riley said clinging to a dying game mechanic wastes a lot of time on most projects, so they had to take a battlefield medic approach to the game, amputating appendages and gameplay features that looked suspicious in order to save the whole.

Henry said the team needed the right tools to cut down on the amount of interdependence. For example, the level design tool had developers mock up a 2D outline for the level, and it would automatically fill in all the geometry of the 3D level, letting the designers iterate without needing to call in other help.

Prototyping was another major point, and Riley said the keywords with early versions were "ugly and quick." The team wanted to make the game fun before they made it pretty and were up and running quickly, swapping in finished art for placeholder models relatively late in the relatively short development cycle.

The team also kept its focus well, Riley said. After laying out the team's vision points, the specific areas they wanted to get perfect, Riley said they had more points than they did months to work on the project. They cut that list to just one unspecified vision point, which helped them keep the project on track and reach internal milestones regularly. Henry said there were plenty of cuts both in terms of levels (originally planned at 75 and scaled back to 50 in the final version) and features, like parts of the game specific to the multiplayer mode.

Henry said experience paid off for Twisted Pixel with the project. Most of the team had about seven years' experience in the industry, and they had "sage wisdom" to contribute. For Riley, that was scheduling broader tasks rather than micromanaging time. Another lesson learned from experience was to have the team working in relatively close quarters so that communication between departments was virtually instantaneous, Henry said.

Also on the list of things they did right was a vertical slice of the game. The pair showed a video clip of their one-level demo of the game, which proved core gameplay was fun and ensured the tools and engine were up to the demands of the game.

As for the list of things that went wrong, time management topped the list. Henry said the team made up for the lack of time on the schedule by using their own personal time. Toward the end of the schedule, 90-hour work weeks were not unusual. Riley said the team was also allowed to work from home, with hours between 1:30 and 5 p.m. required at the office to ensure some communication took place. He also stressed that it was important to take time and "let your family know you exist," even during the harshest of crunch periods.

Networking was another problem. They brought in an expert near the end of the project to incorporate online multiplayer after the fact, but Riley said it should have been incorporated from the beginning. He also added that a patch for the game's online multiplayer is being worked on.

Riley said the team dropped the ball when it came to testing the game as well. There has never enough time for testing, Riley said, even though it's crucial. There were features and optimizations the team had to leave out because they started testing late, and by that time there wasn't enough time to fix the issues brought up as a result.

While the pair said the focus was a win for the game, the things they had to cut were especially painful. Henry said the team wanted to add a proper tutorial and more diversity of environments, Rube Goldberg-style contraptions for the scientists (that would inevitably turn on their creators), and more. Riley pointed out that while the cuts are painful to the development team, they're much less so to consumers who didn't know the ideas were considered in the first place. Cutting a feature is much better than putting one in that isn't especially polished.

As for why the developers tied themselves to the six-month window in the first place, Henry said they were desperate to make it into Microsoft's Xbox Live Arcade Summer of Arcade. Partly because Microsoft puts the most marketing behind XBLA at that time, games released in the promotion sell significantly better than those from the rest of the year, Henry said. Riley added that the Summer of Arcade's time frame is ideal for downloadable games, coming during seasonal doldrums and before the rush of holiday retail titles.

Quote: "The real takeaway from this session is, 'Just don't try to make a game in six months.'"--Henry, at the beginning of the postmortem.

Takeaway: See above.