Wednesday, August 08, 2012

The Dynamic Event Process or Getting from A to B pretty fast.

I was thinking about the many reviews of ANet’s dynamic event system. They range from amazing and heralding in the next generation of questing systems to it’s a standard WoWesque quest system wrapped in a more streamline package. One amazing thing ANet’s Dynamic Event System (can I call it DES?) One thing the DES can do is produce quality content rapidly. The BWE3 Hunger Royale"end event" was reported to have been programed in about 12 hours.  Holy Data Crunching Batman that is pretty amazing. Let me try to put this into perspective in hopes I can explain why it’s impressive. I’ll give my best guess at all the work that could go into creating a Dynamic Event.
NOTE: I’m not a programmer, I’ve never designed a game but I have some experience with creating and development projects and I’ve field tested software systems where my main mission was to stress test it; nevertheless, I’m sure I’ve missed something. Please feel free to chime in.

Dynamic Event Design

The design phase would be the first step in creating a new event. All the possible ideas are brought in and trimmed down, refined and set in order. The timeline for this stage depends on the scope of the event(s). For example, an new event wherein you teach a Quaggan barber shop quartet to sing a song will have a shorter design period in comparison to a world scale event like Tyria vs Smaug the Trademarked. The following lists some areas ANet might address. Again this is not all-inclusive.
· Event Concept:
o Event Description: A very broad description of the event.
o Genre: What kind of event is it? Holiday, end-game?
· Game Mechanics:
o Core Gameplay:
§ What actions will be available to the player consistently and how will those actions influence the world?
§ Mode of play: How many different modes of play will be available in the event?
§ Event Flow: How will the player progress through the event? How will it be organized? Is it broken into mini-events? What triggers its beginning & end? How will in-game assessments be integrated into the flow of the event?
o Types of Characters:
§ How many different types of characters are there?
§ How do they behave differently?
o Gameplay Elements:
§ What environmental elements exist in the game that adds to the event?
§ Will there be items that act as a power-up, etc?
§ Are there different types of weapons?
o User Interface Functionality:
§ What are the user interface items and what functionality is needed for each of the items to include screens and menus?
o Narrative:
§ What is the back-story for the event and characters?
§ What scripting is required?
· Art:
o User Interface:
§ Will screen elements or menus alter?
§ What do they need to look like? Color scheme, design, resolution, fonts, etc.
o Gameplay Elements:
§ What do the event elements (characters and the setting ) look like?
o Sound and Music:
§ What music will be needed?
§ Will new sound effects and voice acting be required?

Dynamic Event Implementation

        After the event is designed ANet would have to implement the event in the game engine aka the game’s program. It doesn’t mean it goes into the online and running game, just into a local server for testing. Now if ANet wants to showcase how fast they can produce content and push it out, the data to prove it would most likely start at this point. The design process is more abstract; as a result, it is difficult to assign concrete data points since it’s a creative process and each design session is different. The implementation phase however can be measured since the game engine is constant, it can be compared to existing standards (the live game) and ANet can record the date it went in and a date it went live. 

Again, I’m making an educated guess but the new event will have to integrate correctly with probably hundreds, maybe thousands, of game process such as class mechanics, race mechanics, skills, timelines, dialogue, instructions, rendering, AI, physics, code, music, data tracking, correctly receiving and distributing data packets, in game analysis, collecting data for later analysis for future improvement and more things I’m probably missing.    

 After the event is in the local system, testing and maybe limit outside testing will occur to ensure that everything works together correctly. Glitches and bugs will be found and the team will address, repair and then test again. Depending on the severity of the glitch they may have to redesign the event. Additional testing will occur as needed and after all the kinks are worked out ANet will still have one more “scary” step - Deploying the event into the current, live and running game without disrupting hundreds of thousands of players’ online game experience. I say scary because you can test and test and test but you never know 100% what might come up when you “deploy something new” as they say in my business. ANet will know real fast if things go bad. Think Diablo III error code 37, yeah.

     So, after reading all that, and I probably missed a number of things, for ANet to successfully push a Hunger Games type event in ~12 hours for tens of thousands of players (I’m guessing the BWE3 number here) that worked, and from what I can tell without flaw, is pretty amazing. Also, in between each beta event/stress test, GW2 made improvements in many areas. In some cases adding new design elements to classes and races or altering the entire talent system.  Again, I’m guessing but I imagine the glitch list was very long, add to that changes in class mechanics, visuals, etc and all done in roughly an month’s time shows ANets ability to proactively address issues.  All of this tells me they built a game engine and a process that can create, implement, test and field working content efficiently and effectively. From just a preliminary review the ability to think-up new events seems to be the limiting factor, not the software. Future content drops will either reinforce or challenge this assessment.

     For my part, I’m confident ANet will be able to produce new events and content in a manner gamers are not use to seeing.  For those who have had concerns about “end-game” hopefully this will help soothe some of those worries.  It may not be end game raiding as many are acustom to but hey, since ANet is already into breaking paradigms why not include end game content?  Can you imagine BWE "end-events" popping up for all the endgamers to play in?  It will be fun to see.


  1. Orr sounds like it will have a raid-like vibe according to recent interviews, but open to anyone in the persistent world, and a lot more complex. Although WvW has always been closest to what I envision as raiding anyway. As I've said elsewhere, the rare city invasion was far more raid-like to me than actual raids in games like WoW. As for end-game, that's such a weird notion in MMOs anyway.

    I have no doubt in ANet's ability to push out content quickly but there's so much already. I suspect new content will come out before all but the most ambitious players have even scratched the surface of existing content.

    1. I agree that the quantity of existing material is immense. In addition, with down-leveling, any toon can explore starter areas from other races and still have a relevant playing expereince, something you could not do in say WoW.

      I hope folks do not burn through the game to 80. GW2 strikes me as a game where focusing on the journey will reap a more rewarding experience than focusing on the destination.

  2. First, Smaug. The other thing is a 'normal' atmospheric condition in LA, Beijing, and other cities with too many vehicles and too much industry.

    I'd say that you are missing the major point on a lot of game or game-like projects - availability of resources and existing systems. As a project team lead I would be far more concerned with a 'teach the barbershop quartet to sing' event than a 'major dragon threatens everyone' event. We would likely have art and system resources for the dragon, the combat effects, the burning houses, the screaming villagers, and so on. Do we have the right models for a barbershop quartet or do I have to get with the designers and wait on the art team? What about the sounds, that has to be important. Hang on, can they only learn one song or do I have to provide options, maybe scaling options? And did anyone check to see if we have the right type of user interface elements for teaching someone to sing or do we have to get with the systems and UI teams?

    It sounds like they may have designed an engine that permits them to quickly link existing elements to create a new event. That's great, it adds a good layer of CM and makes it easier to utilize existing resources. But it will only be fast if they are using existing UI, systems, and art. Anything new will have to be designed, developed, tested, and implemented before it could be released to the scenario design and development teams.

    And this is all flat speculation. I've never seen their architecture but based on supposition and past experience this seems to be correct.

    1. Darn auto correct. Smog is what comes out of Smaug's mouth. :) Thanks for the finding the error.

      You are very correct regarding resources, especially human resources. You need the people to actualy do the inital work for anything new and for continuous maintenance.