You’d think having too many players are launch would be a good thing. But what at first is a dream, could quickly turn into a nightmare.
It’s the dream scenario. After years of slogging through late nights and building a multiplayer game, you release to critical exclaim. The reviews are great, people are hyped and everybody decides to stay up until midnight to be the first to play your game. A million eager players all log in at the same time and log into their first match.
Not many studios prepare for that kind of success. We’re taught scepticism, not optimism. But if you don’t prepare, ecstasy can quickly turn to misery. Servers can crash. Players abandon the game. Refunds start rolling in.
We at Gameye recently helped Tripwire and Torn Banner release Chivalry 2, a multiplayer first-person slasher. While we expected a good launch, it got twice as many players as we’d predicted. But we’d learnt that the path to success is paved with tests.
So here’s how to prepare for the unpredictable.
Launching a game is like intricate clockwork. You likely have a dozen third-party tools all plugged into each other. So it’s really important to make sure they all work together properly. Test them and make sure that your providers can work together effectively. Go through systematically and make sure there are no gaps.
While working on Chivalry 2, we found that communication was key here. You need to be honest with one another, without getting frustrated by changes. The more you can be open and honest in your conversations, the easier it’ll be.
“The relationship has to be a two-way street. Even if we didn’t agree on a specific issue, you were always willing to see it through. There was always room for a conversation,” said Mike Stone, senior producer at Tripwire.
You never really know how many players you’ll get. Over the course of the testing for Chivalry 2, we slowly ramped up the numbers of players on each alpha test. This made sure that the servers could automatically handle the number of players coming in.
But it also meant we could experiment a little with locations. We could test what would happen if other countries suddenly had a huge number of players. Could we spin up sessions in other countries in an instant? What if the game is suddenly popular in Asia? This ended up being vitally important, as it turned out that Korea and Japan got far more players than we expected. Those tests meant we knew we could automatically respond as the demand started rolling in.
Your players are one of your most valuable assets. During our alpha tests of Chivalry 2, we’d all built up a strong relationship with our testing community. They were active and involved, not just pawns. If we hadn’t been so vocal, we might have missed a bug that was hiding in the servers. But with the help from the community, we were able to spot it, identify it, and quash it in two weeks.
Expect changes to your predictions. Models can only tell you so much, and should only be used as guidelines. The real world is much messier and much more unpredictable than a model can estimate. Towards the end of testing, we suddenly had to update our predictions based on new information. It drastically altered how many players we thought we’d get (which was still under the real figure).
Thankfully, we’d already tested for this. We’d looked at what would happen if the predictions were wrong. And so we were able to make all the necessary changes immediately. It took just two days from start to finish for everything to be rolled out.
At Gameye, we know how important multiplayer servers are to a game’s success. So we’ve developed a platform that can automatically identify and spin up new sessions where and when they’re needed. And without breaking the bank: it costs significantly less than a typical solution.
If you’d like to learn more about how we worked with Chivalry 2, and how we work, read our case study.