Designing for single player

[Taylor]

You need to make something before you can iterate it with playtests. The "theory" is all valid as long as it comes with something practical.

[Sentura]

You need to make something before you can iterate it with playtests. The "theory" is all valid as long as it comes with something practical.

it was assumed that at the testing stage you already have something done, obviously. my point was to test sooner in order to get in the early build "feel" so you can build and polish on top of that.

[Bluestrike]

The way I made HL2:Awakening, I just had some general idea's for landmark locations and a rough story.

Then winged it pretty much all together as I went forward and filled in the blanks:

- Filling in the puzzles as I made them up during playtests.

- Connecting the levels.

- Add Enemies where I see fit.

- Even using one of my multiplayer maps to save time.

- Lots of testing and iterations.

And eventually cutting some planned levels because it was taking to long to get everything done :-)

[Furyo]

The first thing I figure out is the kind of game I'm designing the level. Creating an additional level for one game is one thing, creating a level for a game being made at the same time as your level is a whole different thing.

My modding experience is only for MP levels, but the approach is exactly identical because it applies for everything: I figure out the strengths and weaknesses of the game in its current form and try to think of variations of the formula and ways to provide a different vibe to the whole thing. It's risky and harder to do properly than a simple rinse and repeat of the proven concept, but that is the motivation I need to finish the work in the end.

Once I have found that idea, I try and come up with top down layouts that could help provide that experience, but as it turns out I quickly hammer them out in Source to get a proper feel for them and don't look back on paper design. That lack of actual planning and focus is detrimental to the cohesion of the entire level, but I find that it's more exciting for personal projects to try multiple things at once. The downside is that you have to be ready to redo everything multiple times, but that's half the fun.

Once the basic shell is done, I usually have a playtest asap. Try and find out how common my own experience of the level is, and how well I carried over the gameplay I was going for. And if that's positive, I just go full on with the design, making it pretty etc.

When making games at work however, it's an entirely different thing. My experience is mainly in production phases of AAA SP projects, so game design has to be followed instead of experienced against. The twist to make interesting levels is to introduce new mechanics or rythms to the existing formula by relying on LD tricks alone instead of asking for brand new gameplay mechanics no one ever has the time or will to try out. In a platforming game, this can be achieved in the rythm you ask from the player (speed, traps, size of platforms, etc). In AC2, the gameplay mechanic was mine to toy with so that was even more exciting to work on. In Crysis 2 right now, I am also using the existing concept and turn it on its head a bit to drive different emotions and gameplay experience. Environment can be your best friend to achieve that, as well as story twists, but that should obviously be more planned with the story writer ahead of making the level.

If I had to put the different processes down in a bullet point list it would go something like this:

1) Study of game design, strengths and weaknesses

2) Brainstorming of different twists to add in

3) 15 minutes layout drawing session

4) quick hash in a game engine

5) playtest, rince and repeat until formula is found

6) go as far in that direction as you can

[Puddy]

Ehm.

Make a list of fun gameplay ideas. Then plan these around the setting and pacing.

Then create!