Monday, May 7, 2012

A Crusade Against the Arbitrary

I have reached the point in the development of my Pac-Man clone where all of the fundamental game play and meta stuff is implemented and working just swell. Basically, from here I expect that I can just create a huge load of content, implementing additional mechanics as needed, and then, once the game is "playable" from beginning to end, and all of the major issues have been ironed out, I can proceed to add a thin, shiny layer of polish. This is basically the approach I applied to my previous two games, minus the polishing stage, which was substituted with me trying to make sure the first pass on each element was, well, good enough. Such is the case in hyper accelerated game development processes. However, will this work for a large project such as The Pac-Man? I am interested to see.

A problem of entering into the content creation stage is the delicious temptation of creating "cheap" experiences. Or, experiences that provide no growth for the player, present no new ideas, are usually quite "grindy" in an attempt to artificially extend the life of the game, and contain no opportunities for the player to enter to mystical flow state. How are these created?

Creating levels for Pac-Man at first might seem like an incredibly easy job. The designer needs to draw a maze, of whatever shape, and then add in pellets, power pellets, ghosts, and Pac-Man. After this, the work is done; congratulations, you just created a Pac-Man fan game with over +200 levels. Put that on the back of the box (shoddy website).

Except, the experience will be terrible.

In the original Pac-Man arcade game the player had to play through the same maze over and over, to the point that the layout of that is burned into the minds of millions under the fever. Shouldn't a large variety of maze setups provide a better, more surprising experience? In a sense, there is an element of surprise and anticipation of the next maze, but when each maze is essentially the same in nature, except arbitrarily rearranged, these positive attributes quickly fall it despair and death. Haha, alteration is great.

The original Pac-Man maze is designed. It is the perfect size to allow for just enough space between Pac-Man and the ghosts for tension but not frustration. Regarding the ghosts themselves, their artificial intelligence is very carefully designed to take full advantage of that maze, with different sections showing off multiple types of enemy behavior. The symmetry of the maze adds a sense of order and duality, and the placement of the four power pellets makes absolutely certain that they can not be abused. As I mentioned earlier, it is very iconic for an entire generation of gamers, and rightfully so.

Anyways, I have been having to avoid the urges to just simply draw a bunch of randomly conjured mazes, decorate them a bit, and call it good. The option seems so easy to embrace, but it is also ultimately unsatisfying. With the different way that I am handling the enemy AI in this project as opposed to the original it is cloning, each stage can be designed around the wills of a variety of enemies. The goal is to maintain that balance between creating a challenging, engaging experience, but not a frustrating or dull one, which arbitrary level design seems to always lead to.

One last note on arbitrary level design more generally applied to other games. In the past two years, thanks to the success of the indie title Minecraft, it seems that developers have become far more willing and inspired to implement procedurally generated content. There are obviously games that handle this concept better than others, but I think there is a very clear difference in the quality of the experience had in a procedurally generated level design compared to a human crafted level design. Just contrast the experiences had in the dungeons of the Zelda series to those of any game that procedurally creates dungeon. Yes, it is Zelda, but Zelda does design better than almost anything else on the market, and the reputation of the franchise is a testament to that (regardless of fan's feelings on some of the more recent entries, the games still have that designed quality). Even games like Skyrim, whose dungeons are not random generated, but abundant and extremely similar, face this problem. Once the player has entered one dungeon in Skyrim, they have basically seen them all; the only differences are in layout, element arrangement, and story notes that are more often than not completely meaningless or ineffective at providing a constant stream of interesting content.

Since this post is at the breaking point of dissolving into a rant, I will stop here.

No comments:

Post a Comment