What Makes an Alpha?

In all projects, there has to be a beginning. One day in the past there was no game, and no commitment to produce a game. The next day there was a commitment. At that point the project had begun. That was the start of pre-Alpha. For us, that was in April 2016, and it took a lot of work to get from there to Alpha.

Ship of Heroes was developed using a project plan in which our first major milestone was to develop a working prototype of the game; this included a fully functioning character creator, the first version on the playable environment, a set of complete, high-quality signature heroes, and basic animations, and to integrate those core pieces into a working prototype. That took us about eight months, starting with a team of two and expanding to a team of about eight, most of whom were volunteers.

Of course, we also had to create a website, 2D concept art, forums, a newsletter process and format, and our first videos. And then we had to add powers, animations, flight, SFX, enemies, etc., and integrate all of those pieces. And code to handle teaming, kicking, instanced missions, XP, Dust, and drops. And we needed to upgrade our engine version, fix bugs, improve the look of Apotheosis City, and create costumes, weapons, and hair models. We developed our Design Studio to show off costumes, and the Powers Testing Center to show off powersets.​

In addition, we have prioritized optimization since the first day of development. Ship of Heroes is an MMORPG, and this imposes tougher constraints than single-player games typically face. Our code, our character models and costumes, our animations, and every asset in Apotheosis City has been rebuilt from the ground up at least once during the last two years to keep our MMO from unexpectedly becoming a near-single-player game. In our view, MMO optimization is not something a team can do at the end; it needs to be constantly evaluated and improved. As an aside, our first published stress test was in March 2017, with just 14 characters logged in simultaneously with no problems. We called it Remember the Torch. But we don’t consider that to be a real Alpha test, because the only participants were members of our dev team. It wasn’t a real Alpha, but it was fun.

We consider our first Alpha activity to have been our character creator test, which we conducted in September and October of 2017. It took us just under eighteen months from the start of the project to get to our first Alpha. What made this event an Alpha, and not a pre-Alpha or a Beta activity? These are our conditions:

The CCT Alpha involved non-developers in a test of a major part of the game – in this case, the character creator.
The testers logged into SoH and used part of the game, and then made their results, good and bad, publicly viewable.
The feedback from the testing has been used to upgrade the CCT in significant ways. This aspect is not strictly required to meet the definition, but it would be a poor Alpha if the devs didn’t listen to their players.

After the first Alpha, we did a groundbreaking stress test in November 2017. We called that one the Night of the Clones. We tested how many cloned heroes we could get into a mission map that we built ourselves, and still maintain our target of 30 FPS. This showed the effectiveness of our game optimization. But again, this was not an Alpha, because only devs participated.

Our second Alpha activity was the combat testing we did in February and March of 2018. The same criteria apply; we even managed to stream some of the combat live, and to release at least a few hilarious screengrabs of combat gone wrong from the early dev testing. But the key point is that several supporters and members of the press were able to log into Ship of Heroes, grab a character, and fight enemies in Apotheosis City.

Our third Alpha activity will be the upcoming login test we’re planning for the next month or so. We’re going to conduct a stress test to make sure that our code actually allows at least 50 players to log into the same spot in Apotheosis City and move around a bit – sorry, no combat in this one. But this stress test will definitely be an Alpha activity, because members of the community will get a chance to log in alongside the dev team. Of course, we’ve already successfully tested our multiplayer framerate many times with automated systems, but automated systems don’t deliver the same challenge as real-world testing.

Assuming all goes well in our third Alpha, we plan to follow that test with a fourth Alpha, which will be a raid test. The raid test will be similar to the login test, except that we will be bringing in a large group of Nagdellian enemies for the players to fight in an extended invasion scenario.

So, what is a Beta?

For our team, a Beta is when players get to test the game prototype, or at least one aspect or feature, in a form that the dev team thinks is pretty close to what they’re going to launch with. Of course, a Beta can blow up or fail so that significant re-work is required before launch. A Beta that only includes one or two elements, like a character creator or an environment, is not a Beta of the full game, though even a full Beta does not have to be 100% feature-complete; if it were, the game would be launched. In our case, for example, our first full Beta will probably not have a working auction house.

Open vs closed

An Alpha is open if any person can gain access to the game or feature. This includes paid Alphas; the important part is that people not selected by the devs can pay the fee and play. A closed Alpha is one where some non-devs can play, but the opportunity to participate is not open to everyone. The same definitions apply to open and closed Betas. Open Alphas and Betas are more significant than closed ones, but they are also harder to do.

We have had two closed Alphas so far, and we currently plan to skip the open Alpha and closed Beta phases and go straight to open Betas after the raid test. We’re now planning to do more than one Beta for Ship of Heroes prior to launch, both because we think we’ll receive a lot of good feedback from the community, and because we expect to make changes and improvements as a result of the feedback.