Croal opened the session by asking about Hartsman's title, Chief Creative Officer, and what it meant. Hartsman explained that he's not the origin of all creative ideas in the game, but rather, he's a creative problem solver. MMOs, he said, are a source of new and exciting problems that no one can see ahead of time, and as a result, companies have to adapt.
When asked what he did after RIFT launched, Hartsman replied, "Took a nap." But he went on to explain that lots of first-timers don't know that launch actually gets harder. Lots of MMOs either never make it to the finish line or just barely limp across, and the difficulties become evident in the following months. Trion intentionally started working on post-launch features before launch so that it could get a good pace set and be used to the post-launch live pace. This led to five game updates in six months.
Croal then asked, "Does this make your beta your launch?" Absolutely, Hartsman replied. Internally, Trion had a couple of alpha shards running 24/7 as the team worked on improvements and scaling up. By the time of the beta events, the devs were still making changes, and the feedback was valuable, but it was all about mass-scale at that point. They were focused on going from one server of hardware to dozens of servers across multiple continents and all the potential issues that crop up from that.
How does one go about pacing the team? Hartsman explained that he joined Trion two years before launch, and from the moment he joined, the team was on the cusp of being ready to essentially be a fake live team. That team emphasized being ready to ship something to someone every single day and taking care of the things devs worry about so they're ready to tackle the unknown issues that are bound to arise.
Croal next asked about the differences between beta events, with the downtime in between vs. the launch of the live servers when you're running 24/7. Hartsman said that it's more than just putting out fires -- it's looking at the ways players react to the game. You might think a certain feature is awesome, but the players might either ignore it, or worse, hate it. It might even cause them to leave the game completely. So you have to throw your core game assumptions aside and take risks to get the game on track.
This was the case with the invasions and dynamic content, which was still evolving during the beta events. In fact, the biggest iteration, he said, was between the beta event downtimes. The episodic betas allowed for things to quiet down and give the team a chance to focus on improvements.
The metrics helped measure the mood as well. The team could tell which quests weren't completed or were abandoned by players. The devs saw the success and failure rates on zone events and used them to adjust accordingly. If the zone events were stickier for players when they succeeded 80% of the time vs. 10%, then the team would adjust them because that's what makes the game fun.
"Are MMOs a service?" Croal wondered.
"Think of the game as a dialtone," Hartsman answered. "It needs to be up and dependable." Over the last eight or nine years, he said, companies have begun to realize this, that no matter what the problem, there has to be a place for answers. Holding up his phone, he said that he considers himself on call all the time because the game a service.
Croal then asked whether all of the technical stuff and the metrics become a barrier to creativity when it comes to game design. Hartsman suggested that "anything that detracts from making a game more fun for audiences is a distraction." Initially, as the team was building the tech behind the platform, things were in flux, but now that it's built, it should be much easier for Trion to get right into the art of making fun games and good experiences.
Hartsman then addressed the question of how you distinguish between a good mistake and a bad one. He used the example of the dynamic content layer in RIFT. The team assumed more people would be interested in it, that natural leaders would emerge to rally the troops and play through the goals of the world map. There would be the soldiers, happy to participate, and the leaders getting their leader on, and everyone would have a good time. What happened, though, was that an event would start and there would be 200 soldiers and no leaders. No one wanted to be that guy in front of strangers because "strangers are scary."
A member of the development team pointed it out right away and said that the devs needed to get a public grouping system in to let people organically blend into groups and raids and merge and unmerge without needing that first person to step up and take the lead.
As for a bad mistake, he used the example of designing a database that's physically impossible to back up (he added that Trion did not do this). The earlier you make any given mistake, though, the less chance of making a bad one. It's all about how much time you need to solve it.
Lastly, Croal asked about the post-release progress. As Hartsman said, the team didn't take a big vacation after launch. Instead, it got right to working on the first update after the game shipped. The biggest shift overall for the team was figuring out how to stay focused on what's important for the live schedule of live updates and larger scale things for the future.
Trion is starting to work with other partners in other territories now, like Russia and Korea, so those are sub-projects for the main team. Figuring out how to fragment without compromising the state of the live game, he said, is everything.