Game companies are increasingly faced with a certain dilemma related to the release of every little thing in a game: test servers. In the beginning, test servers were used strictly as development areas that wouldn’t crash the regular servers should something go wrong. Infinite loops, those bastid children of deeply nested call functions, could kill an environment over and over again until someone was able to pin it to a particular action in the debug logs. Did we really want paying customers to go on this painful journey with us?
Over time, the rules for pushing new assets – code in particular – changed. The test server became a more stable environment and could host select players who enjoyed the sort of things done on test servers. Chiefly, they loved seeing the features before anyone else and they loved breaking things.
The best feedback can be had by a solid population on a test server. So why are they under-performing?
My personal belief is that the importance of loot changed the appeal of test servers. If you were to brag about owning that superuberleet Battle Axe of Doom you wouldn’t be taken seriously if you won it on a test server. This may be due in part to the interval wipe system employed by some companies; sure you won it this week but next week will be a different matter. Maybe you won it via some strange bug you helped squash or maybe you didn’t win it at all and are just saying you did. Maybe?
So while games still have a devoted audience of trailblazers who may want to be the first to try anything (and the first to break it), they’re not the robust gang of players they were in the cold, dark past. To be a tester back then meant you knew enough about the game to bust things up…you were sometimes invited because of that expertise. But the real game now is on paid servers, where that same player and his buddies are trumping everyone else with the superuberleet stuff they find during regular play. The bragging rights are beyond reproach. The screenshots flood the internet. The challenge to (and from) other servers is keen. The lore for the game will always include “so-and-so was the first player to ever win the superuberleet Battle Axe of Doom.”
But s/he wasn’t, not really. Someone on the test server won it first. See what I mean about them not getting any respect?
How, then, does one attract more people to the test environment? We’re talking about people who are more likely to be test server players than regular server players. How does one appeal to the entrepreneurial spirit of these individuals and keep them where they’re needed most? I have some ideas.
- Games should invite knowledgeable players to the test server for scheduled discussions and play. Sure you can run a Stratics chat, but why not a test server chat? Why not sit in the middle of the environment you helped create and talk with your players and have them walk you through the things they find fabulous or fubar’d to hell? This doesn’t have to be a daily thing. Once a week, twice a month, once a month, every two months. Something!
- Devs should play in the test environment. This doesn’t mean exclusively, but as part of the work day or night. Interacting with players is vital to getting their thoughts on how things feel. Watch the players going through the paces – do they look like they’re having fun? Are they struggling too much on a part of the quest? Are they constantly wiped by the first few mobs? How are they playing their skill sets to achieve the goal? And what are they saying to each other when it’s said and done that perhaps they aren’t posting because it isn’t so polite?
- Anyone who designs a system should join a group on test and actually use the system with them. If you designed something with a certain usage in mind, see if it plays out the way you intended. Chances are very good that player A finds a better way to do it, and instead of running back and forth in forum posts you can discuss it right there and put your fixes into your stand-alone client because you just saw it with your own eyes and yes, it needs to be tweaked.
- Things need to stay in the test environment longer for polish and iteration. Two or three weeks just isn’t enough time in many cases. A month and a half should do the trick, though your mileage may vary from product to product. If there’s an exploitable anything you’ll find out only after a player has had time to figure it out. Sometimes the exploit isn’t apparent until months later – you’re not going to catch everything. But a hefty new system or a change to an important existing one should develop on test as much as possible before release. If you’re actively pursuing perfection, chances are very good the players will stick with you.
These are just some of my thoughts and not necessarily the same ones that other people have. I’m speaking solely from my time on test servers both as a GM and as a player. It’s a great place to be when it’s a good working environment; it’s a drag when it’s a barren wasteland. Players are extremely talented people who should be tapped for ideas and suggestions whenever possible. That one awesome idea may show up in the midst of a sea of not-so-awesome ones but like a hit album, you only need one cut to make the whole thing worth purchasing.
What would you do to revitalize a test server?
