So, a this point you've built your level: it feels the way it should with encounters and puzzles feeling well rounded and interesting. The design at this point should invoke every aspect you were designing for, all without the artistic flourish of lighting, clutter, and maybe a little tweak here and there. This is where you make things pretty and effectively communicate player path for those unused to your level. This also is going to take several attempts, this isn't like the other passes where you do one, test it, and usually it all checks out. You'll see a little thing out every now and then, you're going to find ways to better optimize that particle effect, or use that rubble in a much better spot. Polish passes likely would never end if you let it, so at some point you will have to declare it "done" at some point or another.
Complete is what it sounds like, with some caveats. This is where fixes start happening, anything you had to do to quickly go through the last pass (see kitbashing) gets finalized with proper art, and aspects beyond the golden/critical path gets some love. What parts do you focus on? Well that depends on what your own tests during the Gameplay pass was. Realistically that should mean "all of it", but keep in mind that aspects that are more art driven, like lighting and clutter may not be in that category. At the end of this pass, you should be able to put the current state of your level in front of a player and only feel bad about it, but not cause them to feel the same. The tests from this should help aim you where to begin your next pass: polish.
This is where rubber really starts hitting the road. Using the results of your tests of the Layout stage, you should have a good idea of space in your environment and build on it. It is going to be ugly kitbash it if you have to, even if it only sort of works. That isn't important, what is important is that it can work and be tested. This is a step that often many folks get hung up on, begin worrying about little details that should not dealt with. That said, I'm going to contradict myself and mention that some basic optimization should be done now. It may seem odd with crude nothings in place, but it is still part of a good pass and should be tackled. It also avoids problems like the image below where aspects like RoomBounds or RenderVolumes just wont fit right in the space you had intended. Next up we're going to be looking at a bastardized 'ship-it' quality tomorrow.
With a solid plan in, now is a time to think about layout and space. At an early stage, this can be a little paralyzing or overly liberating, it shouldn't be either. This is where we take this blog series' namesake into mind and reference the previous step where possible. This should help organize the outline in a more logical fashion with as few questions left to your overly excited brain. Now don't get me wrong, you'll want some of that in there as it will help shape aspects such as the player loop, but rely on it too much and you'll end up with spaces and concepts you cannot build or are too complicated to communicate to the player. Rely on your head too little and you'll end up with a nondescript floor plan filled with spaces that have no sense of self or history. Its a balancing act that does figure out with a little experience and luck.
It all starts with Planning. This is even before layout is thought about, and should be simple and short but lead into the actual building of the layout. You need to be able to summarize what is special about the level: history of the place that may not be surfaced to the player, main kit and size. This will help focus and ground the design, as the summary will do as I said it will: answer questions. If your summary is not a good reference to figure out whether you should put a specific room, then it is either too broad or too specific. (And if you are following the 1-2 paragraph limit, chances are it is too broad) This also stops making any design decisions in any other part of the project which would be harmful in taking up not only your time to then stitch into what is on the screen, but also other team members who then have to create any art or code, should it require it in a different project.
Abandoned Mine has been an interesting experience in diving into the Creation Engine and getting myself familiar with the tools and practicing some design fundamentals. That said, there's elements that I've taken away from the experience that would have helped speed up design and iteration time and would have made it better quicker. These lessons I call 'answering the why': creating an answer to your design question before you actually get the point of asking it. These questions are ones that you generally wouldn't think about if you are both inexperienced and/or getting swept up in your idea: 'is this space the right size?', 'do I have all of the gameplay sections I want to explore?', 'what do I do now?'. The biggest aspect of this is understanding how to plan, take iteration one step at a time, and play test as soon as you can. I'm going to be taking the steps of what makes each progressive pass and breaking them into digestible chunks over the course of the next week.
Well, I've built something that I'm, so far, happy with as a playable level. I still think it will need tweaking after some people get their hands on it and it is missing a playable quest to guide someone to the level itself, so it is version 0.2.
To play it, subscribe to the mod through Steam over here: http://steamcommunity.com/sharedfiles/filedetails/?id=717000496
Doubling the size of this level led to a complete overhaul, and, thus, quite an extra workload. In many ways, I was redoing most everything I had done with repurposing old areas and developing new. So many parts I had built had to be removed, re-created, and reimplemented; this is why you prototype first. But I've caught up to where I left off originally, and am now going through lighting.
There isn't a whole lot of conclusion I've come up with when it comes to lighting, much of it I'm fairly new at when it comes to more enclosed spaces. However: I have a couple of screenshots!
This week was all about learning about game flow of the level. It didn't start right off the bat, but it quickly became the theme. With the initial Navmesh completed, which allows the AI to know where it can go, and thus pathfind around, it was time to test, test, and test again. Adding little details, fixing walking areas, moving clutter around, finding new areas for loot, over and over again I came back to the same conclusion: the play space was too small for these multiple encounters.
Often going from one encounter would trigger another that were in too close proximity. This is an issue for several reasons: first and foremost it lowered the player's control over their experience, which is not a fun prospect. It's not enjoyable to accidentally back away from an enemy into another unseen enemy that you couldn't have guessed as there until its much too late. The other major issue is that it also leads to an inconsistent flow. There's no natural transition from one encounter to another with the natural downtime between encounters lost. This leaves the whole area feeling stilted and confusing to the player. To add to that, finishing the dungeon yielded little sense of rising and falling action. There was no crescendo to accent that a dungeon was cleared.
In my mind, my fix was to take what I already had as a base and add a little more to it, limiting the amount of paths while making the experience a lot more authored and nearly doubling its length. This actually ended up working well with staying with the Bethesda design structure in a few ways. It helped keep the pacing present in their own authored spaces in my own, on top of fixing the problems at hand. I'll take the two critical paths and re-purpose them: one to be closed off and instead act as foreshadowing: allowing the player to peek into the room containing the necromancers/warlocks while keeping the pathing for it as a separate room, the other only temporarily closed off, with it opening through a prefab lever that would be accessible after the player had passed a boss encounter; leading them right back to the entrance after a much more lengthy exploration of the space prior. It's what I will be working towards in the next week.