Guild Wars 2 Dev Tracker

For test automation, we built endpoints into the game client so we can interact with different aspects of the game. We have internal commands we can send to the game server to put ourselves in specific server states, and a verbose set of observable and control endpoints for different client states.

The back-end for the game is basically a cloud of clouds of services. The server teams write unit/component/integration tests for common regressions and high risk areas, but QAE focuses on testing the whole system rather than diving deep into component testing for individual APIs.

I think the music team is doing a good job with what they have now:). Maclaine, Lena and Stan all work together and of course with all the live players they use are able to keep up with what we need.

QA Engineering (QAE) started up after launch, so our strategy was focused around covering the breadth of the game's systems. Now that we have thousands of tests, we've been focusing on automating high risk areas, keeping automation results reliable, catching up with new content changes, proactively automating new features, etc.

We use Tableau and JIRA for analytics and bug tracking. The rest of our tools (test execution, harnesses, instrumentation, reporting) were built in-house. If you're looking for more info, I gave a talk about our test automation tools at GDC last year.

QAE has been "agile" since we started. We made some early mistakes trying partial implementations of SCRUM, but we switched to Kanban a long while back and it just worked for us. I'd say we are better off for having made those mistakes, and learning from them, because the team is still invested in our process.

We have a lot of environments internally. Most dedicated content teams have t...

Right now, for the purposes of matchmaking, MMR is currently used when searching for players to be on your team (unless you're in a pre-made group, of course). We do not plan on making any changes to matchmaking until after Season 2 has ended.

How players move between maps in PvP is different than in the rest of the game and because of that it there are technical limitations on our ability to have players queue from outside the Mists. If we can find a solution to the technical challenges around this we would love to be able to do this but it is a very difficult problem given how the code between these systems work.

QA at launch was a very large group (something like 60ish people) in-house, but since then we moved to a leaner model with more testing being done as the content is developed. QAE started up after launch and we are a little more than a hand full of developers now.

Before launch we used Fogbugz, we switched to DevSuite after launch, and recently transitioned to JIRA after the HoT release.

Hank the Honey Badger of DOOM!

Or did you mean from the devs?

She is indeed!

From John Corpening earlier in the thread:

It is definitely our intent to build new maps for PvP for both Conquest and Stronghold. Before we dig into that, we are taking a pass on making some improvements to our existing maps. We've had some feedback from some of our top players on how some existing maps could be adjusted to be more competitive. Expect to see some adjustments along those lines first.

I'll look at setting up something with the whole Narrative team! It would be fun to chat story with folks.

We're aware that there are still existing pain points on the Battle of Champion's Dusk map and we are actively working to address them.

Expansion

Chauncey agrees with you.

We're aware that there are frustrations with the initial starting period for Season 2 and we are currently planning to improve overall match quality for Season 3, particularly during the start of the season.

The team has been evaluating what to do in the off season. We don't have a firm plan yet but it is something that is being discussed.

Regarding ranked play, it is our highly competitive mode. While everyone can participate in it, you do need to think of it more like a tournament. That said, we are looking at ways to get higher tier players to where they need to be faster. In S2 we added win streaks which our data has shown has helped a lot to achieve this. Again, no solid plans yet but this is something that the team is keeping a close eye on.

We believe that having distinct meta periods is healthy for the state of PvP, but keeping the meta constantly in flux was frustrating for most players, because things that were effective one day were ineffective the next week.

As we have shown, we're willing to make changes to improve the health of the meta mid-season, but we also don't wish to constantly change the meta during PvP League Seasons in general.

The story will definitely pick up in Season 3 where we left off in HoT. I totally understand your concern. As Mike mentioned in his letter, above, we had to take a lot of people off of Living World to focus on getting HoT out the door. This meant that we were left with a larger gap between that content and resuming Living World content. We had originally intended one to more quickly lead into the other. But the next chapter will be coming!

And yes, Canach does rock! He's one of my favorites.

We definitely have a long-term plan! I feel that is really important and something that will help us be sure to emphasize key beats and wrap up threads moving forward, rather than create loose ends.

That being said, there are many ways for us to get to the big milestones along the way, and with every new release/season/xpac, we always do a lot of brainstorming on the best ways to get us there.

And fun fact, we Devs have taken to calling it Steve in meetings. It isn't official, but it makes us laugh!

We currently don't have plans for match sizes larger than 5v5 in PvP. When it comes to PvP, we've seen in the past that once we get above 5v5, match quality and individual skill tends to matter less the more players we add, so we won't be going higher than 5v5 at the present.

We are doing both. :) The upcoming blog posts will talk about the features Tyler listed in more details.