Ghost Recon Online (released in summer last year) was the first game in Ubisoft's Ghost Recon series to be built for free-to-play from the ground up. This business model is a unique challenge for any core-oriented game, as more of the industry leans in that direction. We spoke to Ubisoft Singapore's Ghost Recon Online producer Adrian Blunt to learn five important tentpoles for the creation of a new free-to-play third-person shooter.

A starting point for skillsets

In the case of GRO, they began with basic rock-paper-scissors principles for their skills, layering and nesting them as time went on. "Essentially, we're building the game out of core principles," Blunt told us. "I think the first thing that we did was work on having a fluid cover system. Then there was a case of looking at purely systems' design -- so you've got three classes, they've each got two abilities, and so how are those abilities going to work in such a way that they counteract each other? You want to have a system that is always balanced, and so you effectively design the small pieces, make sure that they're balanced, and then layer on top of that," he said. "So, once we've got that, then we've got the weapons system on top of that, which again, from the stats perspective, is balanced in such a way that you never have something that feels like it's overpowered against each other," he adds. At the end of the day, it becomes a huge, complex system of nested powers and dependencies. "We have designers just completely focused on the weapons," he added. "Our lead weapon designer knows more weapons than I would ever care to imagine, but basically her life is in numbers, and making sure that those key elements of the weapons do not make something that's overpowered. The reality is that this is something that's taken time to tune. This game has been in development for three years."

Map creation

Map design starts with the level designers creating something they think might be fun. "They'll create a map that they think is fun and they'll build it in SketchUp," says Blunt. "They'll run through it, and then they'll put it into the game. We run daily play tests on our levels. So, the level designer will say, 'I've got this map that's ready to go.' And so we'll get a bunch of developers into it, they'll play through it, and we'll do a couple of things." First, they look at heat maps, of course, seeing where people are spending time in the maps. But they also discuss the user stories people have had on each map. "We talk about experiences," he says. "We talk about the firefights we had, whether it was fun, whether it actually felt engaging. And then the level designers will take that feedback. I think that's really where the dark magic of level design comes from, taking that feedback and turning it through. "At the early conception of a map, these maps will go through radical changes. Points will move, levels will change, and it can take a while before we start to hone in on what actually feels great. Because it's not just about 'is there good gameplay there?,' but is it actually Ghost Recon Online gameplay," he says.

Iterate, iterate, iterate

Not long after launch, the team found that they needed to tweak a map, because it wasn't performing the way they expected. This is just part of running a live service, Blunt says. "One of the big things that we did on that map -- it's a three-point map, and it's actually the final point, point 'C,' we found that it was too easy to attack," he specified. "We had felt that something was maybe wrong with it. But also looking at the heat maps, seeing how many times is this map captured... Are teams able to roll other teams if you get into a certain position? But also where do the firefights happen? Where is the actual gameplay happening? So we took all of that information, plus the feedback we were getting from the community, that they felt there was something about that that was unfair," he said. They fed this information to their level designers, who then tweaked it a bit, moving that point a little further out. "They built a little more barricade, turned it more into a tunnel, and that just solved the problem," he said. "So, that's the sort of iteration that we'll go through. And we do that on a lot of our maps. Some of these maps are huge. They're massive. Especially the five-point capture maps. They're massive, massive maps. So, there's always going to be something to tweak. And so we will keep going through while at the same time building new maps."

Identify what data is "good" data

Sometimes players will tell you something is unbalanced, but the numbers indicate otherwise. How do you know when to listen to your players, and when to listen to your data? "I think there's a couple of tactics that we can use to hone in on that," Blunt says. "One thing is if the data is telling us that there isn't an issue, but we have a majority of players saying there is, who's right? It may be that we're reading the data wrong, or it may be that that majority is actually just a really vocal minority. So, we will always take it seriously, when we get that sort of feedback," he adds. "Whenever we get that, we will go and investigate," he says. "We will go through the data. As part of the feedback loop, we will then re-engage with the community and say, 'Actually, we've looked at this. We don't think that's the case.' And then we'll get into more of it through our community teams and really hone in on what that issue is. Because I think if a lot of people are experiencing something or feel something, then something has to change. "The challenge is in finding what it is. If the data says that that particular thing is fine, is it something else? Is there something that is related to that that we can make a tweak to?" Blunt suggests that you look at your numbers for a wider perspective, if you want to find what's really at fault when players feel like something is unfair.

Integration of live and new content teams

Essentially, for the GRO team, everyone is a member of the "live team," says Blunt. "The reason for that is because everything we do has an impact on live," he adds. "We don't want to have that separation of people making stuff and not thinking about what's happening in live, because that feedback loop always needs to keep coming back to us." But the company does differentiate via strike teams, working on specific things at certain times. "We will have a strike team building a feature set, and we'll also have a strike team that's really there to make sure that the live service, any issues that are coming up, are being addressed," he says. "And that's a dedicated team that can rotate through, but at any given time, there's a dedicated bunch of guys who will be looking at bugs, investigating them, maybe fixing them, and really being the front barrier for the dev team so that they can concentrate on building features. "One of the great things that strike teams allows us to do is to have a lot of communication among the teams," he says. "We have a lot of standoffs where everybody talks about what they're doing, what the challenges are across the teams. The producers who are project managing all of these strike teams are also talking to each other so they're aware of what's going on. But the other thing that this approach allows is that we can have a steady rotation of people coming through. So, if the idea is that we may need strike teams for other features to remain slightly more static, the live team is actually on a quicker cycle so that basically anything that's coming through, they have awareness of stuff that's being built. But it's basically through communication. It's talking to people." Editor's note: Ubisoft provided travel accommodations in order to facilitate this interview.