How to Build a Base in the 26th Century

Published Monday, May 18, 2009 1:10 PM by Aloysius

Halo Wars had its fair share of challenges. We were trying to bring the strategy genre successfully to the consoles for the first time. We were working with someone else’s IP. We had never done a console game. The list could easily be bigger than that, but you get the idea.

Now that Halo Wars has shipped and everyone’s had a chance to see, play, and evaluate it, we have the opportunity to discuss more about the process of making Halo Wars. For example, everyone knows how base building works in Halo Wars. Each base has a fixed set of slots (though you have to upgrade to get access to all of them). You click on a building site, make a deft move with the left stick, and tap A. Presto, here comes that Barracks you ordered. Nifty. It’s fast. It’s quick.

But is that a good thing? Most strategy games are noted for their deep city building and economic elements. WTF Halo Wars??? Why don’t you have an Age of Empires-like infrastructure game where the truly skilled players can “out econ” the newbies? The fairly un-sexy answer is that we just wanted a faster game with a larger focus on combat. Okay, fair enough. Let’s look at the iterations that base building went through to achieve those goals.

When we started Halo Wars, we had the relatively simple idea of bringing the “Age Experience” to consoles. The longer we worked on Halo Wars, the more we realized that we were doing something else: bringing a new strategy experience to consoles. Age is a PC game with PC sensibilities. Halo Wars needed to be something else given its platform. The games needed to be shorter. The action needed to be faster.

So how does that affect bases? Well, given our background with Age, the first base building system we implemented on Halo Wars involved placing buildings wherever the player wanted. Just like Age. For us, this was a logical starting point. Over the years, we’d fought hard to hold onto the idea that nothing invests a player in his base more than building it himself. And, most designers will tell you that base investment is a good thing for a strategy game.

After we got farther, though, we ran into problems. Players were spending too long building out their bases. At one point, we were even letting players rotate individual buildings. That made the problems twice as bad, not to mention exposing some brittleness in our new engine’s placement code. Players would agonize over the decision of where to optimally place a barracks instead of spending time watching their troops. On the one hand, that’s good. We want folks to care about that non-combat stuff. On the other hand, every second spent in your base was a second not spent in combat.

So, we tried to simplify building placement to make it simpler and faster. We made the buildings all the same size, so it was easier to visualize/control the layout. Yuck. The buildings all ended up being on the “big” side of the spectrum to accommodate our need to drive tanks out of them. That was a lot of wasted space. So, we made them all smaller. Too hard to tell them apart. So, we went back to varied building sizes. But, we took out buildings so you had less to build. Blech. That was definitely too far. It was clear that players didn’t have enough to do in their bases when we did that. All the time we were struggling with this issue, players weren’t having fun. Building placement was frustrating and it was killing the game.

Halo Wars has a really cool terrain system, perhaps the best we’ve ever built. It needed it, frankly, to even compete on the console. The game had to look hot. As we moved away from our random map heritage, though, things got more complicated. The terrain was more varied, so simply plopping a building down anywhere was harder because the ground was never really very flat. The code to dynamically raise and lower the terrain hadn’t turned out as well as we wanted, so we had problems just letting you build anywhere. As we looked at our options for fixing building placement, we decided to live with those limits and build the design around having players only build in “mostly flat” spots.

As you might expect, our next iterations were played on exceedingly boring, flat maps. Not good. After a while, we couldn’t take it anymore and the maps went back to having height variation. That was really a blessing in disguise. That issue pushed us over the edge on finally accepting that we needed to keep our bases “collected” in small areas. This turned out to be great for helping people parse and learn the maps quickly, which suited our faster play goal. But, that also sucked royally with our placement system. Things were too cramped.

As the playtest feedback inevitably started to re-suggest making buildings smaller again to work around the cramped towns, we stepped back. Something big had to change. After what turned out to be a couple of months of hemming and hawing, we left behind our “place anywhere” traditions and adopted the so-called socketed building model.

In hindsight, I think this was one of our best decisions. It instantly made the gameplay in bases quick like we wanted. The bases also looked a helluva lot better, too. The random, ramshackle hodgepodge was gone. There was some order, which sorta suited a heavy military game anyway. Design-wise, it was also a clear system that visibly presented you with your tradeoffs. You knew how many sockets each base had. You didn’t have to play Building Tetris only to find out that there wasn’t any space for the building you wanted to place.

It was somewhere during this 5th redo of the buildings that we realized the next problem. The bases were still too big. They looked awesome, sure, but they were too big. We couldn’t have enough bases on our 3v3 maps with the base size as it was. I think the only time I’ve ever truly feared for my life was when I had to tell the artists that we had to redo the buildings again. If only that was the last time.

The next (and last) big problem on base building and building size turned out to be the need for the units to exit their buildings. We really wanted to showcase that. It had been received very well in our E3 2007 demo. We had gutted that base scheme during the various un-fun iterations of the “place anywhere” systems, so most of the E3 2007 stuff was gone. But, we were determined to keep the units exiting their buildings. But those tanks are so f’ing big. Each building site had to be big enough to accommodate the Vehicle Depot, which was killing our attempts to reduce the base size.

I honestly don’t recall who suggested that we have each unit exit from some building. I do recall that it was met with the burning hatred of 1000 suns when I pitched it to the team. It wasn’t a popular idea. It was another piece of the vaunted “Age gameplay” that we were taking away. But, taking it away was the right thing. We could make the central building bigger and the outlying buildings smaller, which fixed our base size issue. Plus, design-wise, it made gather points much easier to manage since players only had one. Another “win” for ease of play on the console.

Okay, I lied. We actually had to redo the buildings one more time after that because the Design Department kinda goofed on some calculations with the turrets. But, let’s not go into that one.

Dave Pottinger, Lead Designer