Wednesday, March 30, 2011

First Public Version of A Valley Without Wind To Be Beta, Not Alpha

It pains me to say that we're delaying the first public version, but Keith and I spent a good while talking about it today, and we agree it's for the best.  We've decided to scrap the idea of a public alpha, and instead are going to start with a public beta. 

In practical terms, what does that mean?  It means that instead of having a version for preorder/demo in 3-4 weeks, our timetable is shifted to "we don't know."  It might be May or June, or it might be even further along.  We don't have a crystal ball, but things are proceeding well so far, and we'll know when we see it.  In the meantime we'll keep you as informed as possible about both our progress and plans.

The Obvious Question: Is The Project Behind Schedule?
Not in so many words, no.  It's simply lopsided.  We have an awesome engine at this point, but not a lot of actual game so far.  We've been focusing on the technical aspects, and the construction of this infinite world itself, and multiplayer, and the basic systems for gameplay such as crafting and item use and combat.  These were the bits we were most unsure about, and the bits that underlie everything else.

In many respects, we're a lot further along than I'd expected to be at this point.  It's also high quality work: the worldbuilding code that is there is practically final, and we just need a lot more layers on top to really make a truly varied world; the parts of combat implemented so far are also really solid, and work in multiplayer, and most importantly are fun.

But that's not an entire game yet.  To put it bluntly, what we have at the moment is this fairly repetitive world that is nevertheless infinite, and which has very repetitive gameplay due to simple lack of content.  Skelebots chase you, and you fire spells at them, and that's about it.  This tiny scrap of gameplay is fun, but most people would wear it out after 20 minutes or so

Why Didn't We Start With The Gameplay?
A few people have asked why we didn't start with the gameplay first, as they've heard game designers should.  The answer is that we did -- in terms of what we designed.  In terms of what actually gets implemented first, that differs from project to project, even in AAA games.  In order for us to prove out interesting gameplay for this specific game, we had to prove we were capable of making an interesting, dynamic world to house it.  So that's what we've done.

We also had to prove out that we -- I -- could do the art in a way that looked good, and that was reasonable in the amount of time it would take.  I've done the art for some space games so far -- Light of the Spire and AI War 5.0, mostly -- but doing all the art for a game of this sort was a new challenge that took a lot of figuring out.  Until we had that sorted, we couldn't be sure we could really make this game, given our very tight budget and staff constraints.  And I think that's been successfully demonstrated at this point, as well.

We also had to implement a whole new networking model -- this one fell entirely to Keith, to my great fortune -- and we weren't sure we could do that in any reasonable amount of time.  Or what sort of refactoring it would take if we waited until later.  So we hit that up early, too, and it's working quite well if not fully optimized (aka, no smoothing or prediction).

When starting out with a brand-new project, especially a large one, you want to start with the bits that are worrying you.  The bits you aren't sure you can pull off.  If you can reassure yourself that those bits are feasible, then you can proceed with confidence on the entire rest of the project.  If you leave those uncertain bits until the last, you might be in for a nasty surprise, and the whole project might fall apart.  As of a few days ago, we've now hit all the big points of uncertainty -- lighting was the last of the brand-new things that we weren't sure about.


All Projects Start Small, Then Accelerate
The description of the gameplay two sections back might sound lame, but at one point AI War was just a game with fighters, light starships, and engineers.  I guarantee you every other game started out the same way.  At some point Chrono Trigger must have been just some empty-ish fields and a couple of repetitive  battles with one or two playable characters.  Forget all that story stuff, time travel, variety, and interesting locales. 

I remember screenshots from what turned into Ocarina of Time when it was for the Nintendo "Ultra 64," and all it involved was Link fighting a single Stalfos.  They were soooo excited about the revolutionary new "Z targeting" system they'd invented, though they didn't call it that yet. 

And they were right!  Boy was that revolutionary.  They made some amazing engine strides, and were able to do something that no game had done before.  The entire game rest of the game was designed and built around what that one mechanic -- everything from camera angles to boss fights.

But it wasn't a game yet, or at least not a very fun one.  This is more or less where AVWW is at these days.  We have a working prototype in hand, it's all sorts of unprecedented, and we're super excited about it.  But this is the game-lifecycle stage of videos and screenshots for a reason, when it comes to larger companies.

Why Were We Promising A Public Alpha, Then?
This was my error.  See, we've been doing AI War expansions and engine upgrades for the last 9 months.  The last full game we did was Tidalis, and puzzle games are necessarily smaller in scope than most other kinds of games.  The last time I was this early in to a project the size of AVWW was about two years ago, with AI War's alpha versions.  Consistently, what I had at that stage was an overgrown prototype.  I should have remembered this!

But we've been doing all these expansions lately, and that was fresher in my mind.  When you do an expansion, a funny thing happens: you can do one day of work, make a new ship or something, and then put that out as a public beta on preorder/demo.  This is no problem, because the fans already have this huge game to play with, and they are happy to get that 1 extra ship added in and see what it does.  Then every day as we do more work, we release more betas quickly, and everyone stays happy.  Before you know it, a few months have passed, you're doing balance testing and polish, and you release.  Success!

I think that this led me down the wrong path, paired with thinking about how Minecraft spent so long in alpha (and I think I read it went into alpha after only one week of development, if I'm not mistaken).  Minecraft is another one of those special cases: it has an enormous creative component, so players could have fun in the sandbox of the game even when all you were doing is digging up dirt and rearranging it into houses or whatever other patterns.  Minecraft could start out extremely primitive in its first versions because of its very nature, but you can't build anything out of skelebot corpses; there has to be more of a game here right from the start.  It's not better or worse, it's just a different kind of game.

Why Not Release A Public Alpha For Those Who Really Want It?
We were criticized for releasing our first footage and screens of AVWW too early, in too raw a state, and I'm still on the fence about whether that criticism was entirely correct.  Some of our fans were really salivating for anything we could show them, and in the end it turned out we got some really useful feedback that helped us to improve our graphics in a way I might not have thought to on my own.

Regardless, I'm increasingly coming to feel that it would be a colossal, perhaps unrecoverable mistake to release any playable build of the game "too early," whatever that subjectively means.  The obvious reason is that the press, and people with a passing interest in the game, might download it and not like it in an early, unpolished state.  But even when it comes to the fans who are most excited about the game, there's this problem: an alpha could never live up to their expectations. 

The press would actually probably be more forgiving, because they at least are used to seeing early builds of game software -- they know how the hotdog is made, so to speak.  But when it comes to the hardcore fans, they want what we've been promising -- but it isn't done!  They want settlements and to see a big interesting world that you can explore around in, and they want NPCs that they care about, and intriguing bits of story, and goals that actually matter.  They want to be able to affect the world in ways that are lasting.

So... should we hand them a game with a few enemies, basic sketches of interiors and exteriors, no real overarching goals, and no memory-of-deeds system?  That's insanity.  Those fans that have been around us a long time, and have seen us develop games before, know the process and would probably give us the benefit of the doubt that we'd make a good game.  In their view, we've done it before.  But the excitement would be gone.  It wouldn't have done them or us any service.

So Is The Project On Schedule Or Not?
Oh, yes, it's very much on schedule.  Our goal is 1.0 in October, and that seems imminently hittable.  The engine is vastly further along than I thought it would be at this time, and we're finally getting into the vertical development phase of the game in several areas.  We have all the basics of the explore/craft/combat trio down, and those are fun and simple.  We even have the very basics of NPCs.

What we need before we start showing this around in a playable format, though:

1. A lot more content of all sorts.  We've already been planning that for the next 3-4 weeks, and those plans haven't changed.

2. For the game to set dynamic goals that the player actually cares about, in terms of them being able to help out other NPCs, or find the lairs of bad guys and kill them, or whatever.  We don't presently have this at all, and there's no way we cold do this in 3-4 weeks on top of #1.

3. For the game to remember past events in a meaningful way, so that the whole benefit of the perma-death system becomes clear.  This in turn feeds back into #2, and there's a cycle of complexity here that we need to sort through.  The engine is ready and waiting for this, as is most of the design, but we have to actually put this in place.

4. Ways that the players can strategically affect their world.  As of today we have the basics of the wind shelter system, which is a great stride forward, but that isn't enough (and it isn't even fully finished yet).  What we really need are settlements, as well as the more complex interactions with NPCs that are related to #2 and #3.  And that also requires more crafting and other forms of content from #1, to really make it so that the player has enough choices for there to be any strategy to it.  See how this all interrelates?

5. More ways for the game to set up obstacles for you, that you can then overcome.  Even simple things like locked doors provide a surprising amount of  interest, because those represent something you can't do until you solve whatever puzzle is associated with the door (find the key, or the lever, or whatever).  That's a superficial example, but a lot of this also comes back to having notable arch-foes and the hopes of characters that you can work toward solving... again, speaking to #2  and somewhat #4.

When all of those things are in place, what we will have is a "broadly feature complete" version of the game.  In other words, all the major points of horizontal development will be done.  We'll still have loads of work to do in terms of adding yet more content and fun stuff to do, plus plenty of polish, and I'm sure other subsystems and points of interest will come up if there's time.  But when we hit those milestones, that will be a game that is varied and interesting for at least a few hours, if not far more.

That's when we'll start needing player feedback and content.  Right now the code changes enough that we don't really want custom content that will all just have to be changed, anyway.  That stuff needs to stabilize before we invite in a ton of people.  If I was a player giving feedback on this current version of the game, my feedback would be summarized by the above blog post.  Those are the largest "problems" with the game, and they all are simply factors of it not being done yet, like with the Stalfos battle in the Zelda prototype.

Players can't give us meaningful feedback until we actually have things a bit further along, I'm realizing.  Otherwise the incompleteness of the current builds masks everything and adds way too much uncertainty.  That's a new thought for me.

Just Sum It Up, Please: How Is The Project Going?
It's actually going really, really well.  The engine works and is in excellent form, and we've got the beginnings of content development going on.  What's there is cool and fun, but it's not yet nearly enough for anyone to really have a good time with it yet.

It's going to be a while yet before we hit that critical mass.  There's always a tipping point when a game goes from "promising prototype" to "this is an non-final build of a really fun game, and it's already fun."  We'd be stupid to release something that wasn't in the second category; I don't know how the press would react, but I know our biggest fans would be really let down.

What Comes Next?
It's quick for me to do textual posts, and to some extent screenshots or at least sprite samples, and I'm going to try to do more progress reports along those lines.  Videos are vastly slower, and are something I don't think I'm going to try doing any more regularly than biweekly.  Erik, our PR guy, may start doing some videos before too long, to take some of that burden off me.

In terms of when we will hit public beta, the answer is that we don't know.  It's really up in the air.  I would be really surprised if it is in May, honestly, but you never know.  I also would be really surprised if it is in August or after.  My best guess is maybe sometime in June, but that's too far out to really have any degree of accuracy.  We'll know we're ready for beta when we see it; it will have to meet all the criteria on my list above, and then I'll be happy.  Right now we're just chipping away everything that doesn't look like an elephant.

I can't wait for folks to be able to see this game, but we do have to be sensible about these things.  If it makes anyone feel any better, I've been on the anxious-player side of things my whole life.  We're telling you now because we didn't want to string you along with "oh, it will be next month" every month.  As a player, I always found that supremely annoying.  I never quite understood the dynamics of what the developer was going through until lately.  Live and learn! 

4 comments:

EmperorPenguin said...

As much as I would like to get my hands on this, I think that's a wise decision, Chris.

Christopher M. Park said...

Thanks for the support!

Jeppeboy said...

I fully agree with EmperorPenguin. I'm very excited about AVWW but I think it's a wise decision to wait until you have an actual game before you release anything. I'm glad to hear that the progress on the game is going well!

7rigger said...

Well, I won't lie, I am disappointed in this news. But it is your project and I trust in your decisions to make the best it can be.

I wish you the best of luck in getting the game to the level we all know it can reach :)