Mirror's Edge 2D/3D: Reflection, Review, & Repair pt.1

Before I got a chance to play Mirror's Edge on my PS3, I whet my curiosity for the game by playing the 2D version. After playing the 2D version, I discussed a few things with its creator Brad Borne. At that time I made some predictions as to what I thought would be the design features, or lack thereof, that would hold back the next-gen version of Mirror's Edge from greatness. Since then, I've played and beaten Mirror's Edge as well as dabbled in its time trial mode.

After spending over a week writing over 7.7k words for my feature 2D + 2D = 3D, I decided to take this great opportunity to compare and repair the game design between the 2D and 3D versions of Mirror's Edge. I will cover each game in the first two parts and then move on to the repair in the 3rd.

The 3D Mirror's Edge (3DME) is a fairly unique video game. From the main character, the graphics, the story, to the first person actions actions at the player's command, this game has carved out a unique space for itself. Unfortunately, the final product reflects the rushed culmination of a collection of neat ideas that were not tailored to each other. Somewhere between the obvious first person shooter design and the lack of quality platforming game design, 3DME ends up providing a hazy and fragmented gameplay experience that occasionally manages to focus on what little quality design it has.

Thinking of gameplay systems in layers is the best way to comprehend the complex dynamics that are at work. Since interactivity is at the heart of the video game medium, starting with the player mechanics is the best way to get to the heart of the matter. From there, we look at the core elements of the game that make up the building blocks for challenges. Next, we consider the level design and/or the specific ways the elements are arranged to make the game challenges. And finally, we analyze a game's design elements that work to mix up and freshen the gameplay (level design, challenges, contrary motion, etc). This layer includes elements of transformation, suspension, power ups, and bonus influences.

The next-gen, 3D Mirror's Edge is a game that is built up from weaker and weaker layers design wise. More specifically, the core mechanics are mostly solid, the level elements less so, the enemies even less so, and finally the level design is the worst. Let's start with a close up of the core mechanics.

Core Mechanics

Primary Mechanics: UPACTION, DOWNACTION,

Secondary Mechanics: TURN (90/180), ATTACK, REACTION TIME, HINT

Tertiary Mechanics: INTERACT, PICKUP/DROP

The look and feel of the core mechanics are what most people appreciate most about 3DME. For the most part, UPACTION and DOWNACTION provide contextual platforming moves that smoothly transition between previous actions/states. Momentum is a key element of the player's motion, and the first person perspective of 3DME gives even a relatively simple action like running a fresh feel. Unlike other first person games, Faith's arms and legs are rendered accurately and convincingly.

The ATTACK mechanic is simple and is mainly used to beat up enemies and knock open doors. The mechanic suffers from a bit of ba3D making it difficult to judge if Faith's punches will reach a target in front of her or how much Faith's attacks missed by. This is espeically true for the sliding kick.

The INTERACT mechanic is surprisingly undynamic when it comes to disarms. 3DME is designed so that you can only disarm enemies during a small window of time when they strike at you. This means, when an enemy is stunned and doubled over in pain, the game doesn't allow you to take their guns away from them.

Overall, there are a few quirks with the mechanics of 3DME. The drawbacks aren't too substantial. From this point, it's up to the rest of the design layers to shape the gameplay experience to focus on what Faith does best and avoid the trouble areas in the design of the player mechanics.

 LEVEL ELEMENTS

3DME is a game where players interact directly with the various elements of architecture. Edges can be grabbed and climbed over. Pipes can be climbed. Poles can be swung on. Walls can be kicked off of. Small boxes can be vaulted over. Glass can be smashed. Chords function like tight ropes or zip lines. The list goes on.

The biggest problem I have with the level elements is that they aren't very dynamic or variable. Aside from the glass and perhaps the doors, the stage elements can't move, bend, or break. The exact way the developers have designed the elements to be used is how they can be used. In Mario bricks can be broken. In Sonic walls can be smashed. In Spelunky most objects can be bombed through. And in Jumpman the levels can be rotated. A little transformation can go a long way. 3DME has far to little. And what little it does have isn't incorporated into the level challenges/dynamics.

ENEMY ELEMENTS

There are not a lot of enemy elements in 3DME. Aside from the occasional spinning fan, helicopters, exploding barrels, and the moving trains, all of the other enemy elements in 3DME are human. Unfortunately, most of these human enemies use firearms to attack Faith. As I've explained before, guns are very powerful weapons that can have damaging effects on a game's gameplay. In 3DME, kike in first person shooter games, it's hard to determine where enemies shoot you from because guns are long range weapons that can target you from different heights and positions in a 3D environment. Because guns fire projectiles that fly fairly straight at high speeds, the enemies are constantly engaging the player in a functionally 2D space while the player must maneuver through a 3D space. The result is like playing an FPS like Halo where everyone else is shooting you with Battle Rifles and you can only move around and melee attack.

The gun wielding enemies (Patrol Cop, Riot Cop, Swat Cop, Sniper, and Swat Support) put the same stress on the design of 3DME that first person shooters face. When you don't have a gun to engage a gunman, you're probably going to get shot. The more enemies with guns around, the more likely you are to be shot to death. In Resident Evil 4, when engaging in contextual counter moves against enemy zombies, the player becomes completely invincible throughout the duration of the fixed animations. In 3DME, when the player disarms a cop, during the fixed animation other enemies can pepper you with bullets and even kill the you. The AI/abilities of the enemies are too effective. The game never finds the balance between the enemies being too good or too dumb when you overcome them most likely after many deaths.

Though the enemies don't layer well with the player mechanics and level elements, which I'll talk about in the level design section, the enemies do provide an interesting source for poweups an element of suspension that effectively turns the tables on Faith's ability to maneuver through the environment and attack enemies. Instead of running from enemies through 3D space, players can use an acquired gun to dispense of enemies in a functionally 2D space. The limited ammo for acquired weapons is an element of decay meaning whatever advantages you gain from using a gun won't last forever. Beyond this restrictive design element, Faith's platforming abilities are significantly reduced when holding a gun putting a limit on how far players can carry a gun away from its source. There's a lot more potential in this gun powerup design, which I'll cover in the repair in part 3.

LEVEL DESIGN

Here is where all the potential Mirror's Edge has stored in its mechanics and level/enemy elements is unfortunately lost. Overall, the levels are either linear or alternate path. Though I love the pristine, almost sterile graphical style and the architecture of the game world, the form comes at a cost to the function. By making all of the environments look interesting and somewhat realistic, the functional visual fidelity of the game drops off somewhat. In other words, though the levels are composed of a variety of level elements there are a lot of objects in the environment that players can't interact with as expected according to how these objects look. The result is a game world that is design first around looks and then around play.

All of the levels/areas in 3DME can be broken down into four types: explore, chase/escape, puzzle, battle.

• The explore sections aren't designed to pressure the player with timers or enemy elements. The environment is spread out around you and you simply must make it to some target destination. These sections usually have alternate paths that give the player the ability to test their curiosity and increase their difficulty.
• The chase/escape levels put the pressure on the player as he/she must maneuver through an obstacle course that generally doesn't feature a lot of alternate paths. In these sections, you must either keep up with/chase down a moving target, or run away from oncoming armed forces. These sections are all about moving forward as quickly as possible.
• The puzzle areas are a lot like the explore areas. The player is put into an environment without pressure from timers or enemy elements. In general, the puzzle sections have less alterate  paths. Instead, they force the player to figure out how to get from point A to point B by being observant and using more nuanced mechanics.
• The battle areas are all about taking on enemy units. It's their guns against your maneuvers. In these sections the level elements that were useful for maneuvering around can be even more useful for cover. All of the enemies patrol out in the open while Faith has the ability to leap up and over obstacles and run like the wind.

The problem with 3DME's level design is that each of the four types of levels aren't gone particularly well designed. The explore level design seems open and free until you realize that you're not making your own way through the environment. Without any type of level transformation or ability powerup, you're simply picking the paths the developers have made for you. Without elements of tension or contrary motion, there isn't much point of doing things one way or the other. There are no coins to collect for making trickier jumps, enemies you can avoid by taking a treacherous path, or a time limit to give context/purpose for the exploration.

The chase sections are surprisingly linear. When in a chase, players generally have to maneuver in the same ways as their targets. If you mess up or attempt to experiement and create a short cut, the target will most likely get too far away and the level will reset. The chase sectoins don't have a lot of variation or contrary motion to keep things interesting.

The puzzle sections push the the limits of 3DME's fire person core mechanics. Unlike Portal, a puzzle game that uses the limited first person perspective to enhance the core mechanic that creates bridges between two areas in 3D space, 3DME's puzzles are simply about moving through 3D space. The ba3D that comes from judging distances in the first person coupled with how the environment is mostly obscured from view in the first person perspective are what put too great of a stress on the puzzle platforming design. All of the environments in 3DME are like puzzles that force the player to look around and assemble a bigger picture of the space around him/her. But the puzzle areas specifcially take things too far with areas that are designed to be especially convoluted, cramped, and obscured.

Finally, the enemy units that populate the battle sections aren't designed to work with the dynamics of the player's primary mechanics or the dynamics of the level design. If 3DME were a platformer, the enemies should be designed around the dynamic of gravity and space: what goes up must come down. Instead of moving around in 3D space, most of the enemies move around along the flat surfaces. Instead of engaging the space, they simplify it with guns. Because of their accuracy and persistence, using the 3D environment as cover is somewhat diminished. In other words, even if you can climb and vault all over the rooftops, a bullet can hit you as simply as if you were walking around. 

For these reasons 3DME is caught in an awkward place between trying to be like a shooter, and action game, and a platformer, yet it falls short of every goal. The infinite checkpoints and respawn penalty/difficulty system really held the game back from sustaining any high level of design. Instead of designing a game like Super Mario Brothers, where players can see every upcoming threat with enough time to make informed decisions and to use their player mechanics with a complete sense of contained control, 3DME design is filled with short cuts and holes that, when players fall through, are smoothed over with an easy "try again."

In 3DME gravity isn't a core element of the game's challenges. For this reason 3DME is more of an action game than a platformer. And as an action game, it uses roller coaster level design, which makes it more like Sonic Rush Adventure than Super Mario Brothers. The story is cheesy the quality of which I find more fitting for a "made for TV movie." The animated cut scenes are worse than the Esurance commercials. The time trials mode eliminates the gun issues and focuses the gameplay on moving through an environment and folding the level on itself by moving the checkpoints. While this is a good mode to have in the game, adding a time trials element to any challenge is a simple way of increasing its difficulty and appeal. After all, just about any task/challenge no matter how menial can become interesting when you challenge yourself to do it faster or more efficiently.

In conclusion, 3DME is like...

• Super Mario Brothers without the contrary motion of the enemy design, powerups, coins, suspension, or transformation
• LittleBigPlanet levels that use too much electric material and contain too many "free ride" or automatic level elements like spring pads and car rides.
• Sonic levels without powerups or rings, and that use too many spikes and pit falls making the game more trial and error than need be.

In part 2 I'll detail the 2D Mirror's Edge game.

Jumpman Review

Here at the Critical-Gaming Blog, I've been upfront and clear about my game design theories and preferences. Even after spending such an extensive amount of time thinking and writing about games, I'm discovering new things about my appreciation of video games all the time. It wasn't until I played through Jumpan that I realized how effective the different design layers of Classical game design are (following the framework established by Super Mario Brothers on the NES) and how the inclusion of even a few of these layers can turn a mediocre game into one that I keep coming back to for more. Play it here.

 

Let's start at the core.

The Core Mechanics

Jumpan is a platformer, meaning that the core dynamic of the game involves fighting against gravity and using it as well. Mario, the original Jumpman, was designed around the primary mechanic JUMP. Jumpman gets the job done with the JUMP and ROTATE for primary mechanics. Experienced platforming gamers that have been trained in the ways of Nintendo, will find a few jarring differences between Mario's and Jumpan's JUMP and MOVE mechanics.

Mario's JUMP is dynamic meaning there are a range of interesting effects in the game like stomping enemies, breaking bricks, stoping/kickicking shells. Mario's JUMP is also a direct mechanic in that it changes according to the player input. Tap the button and Mario short hops. Press it lightly and Mario Jumps high. Hold it down and Mario will Jump his maximum height. These variable jumps go a long way in giving the level design the freedom to include variable JUMP height challenges in addition to giving the players the ability to fine tune their aim/timing by taking bigger/smaller jumps when needed.

In Jumpman, there is only one JUMP height. If you want to cleanly JUMP just over an enemy, you simply can't. When trying to make quick successive JUMPs, the pacing of the game slows down as players wait for Jumpman to slowly fall back to the ground from his full JUMP height.

While Mario can MOVE carefully and accurately at constant speeds stopping quickly if need be, Jumpman features an unconventional horizontal MOVE mechanic. In Jumpman, holding left or right does not control the player avatar at an even speed. Instead it accelerates Jumpman. From what feels like a very small effort, Jumpman will rocket across the screen horizontally. Also, instead of stopping quickly on the ground due to friction, all the platforms cause Jumpan to slide as if everything were made of ice. Some have found this "icy" design annoying and hard to deal with. However, it wasn't until I played the game through the end that I understood the purpose for these unconventional core design features. More on that later.

Besides JUMP, the other primary mechanic of Jumpman is ROTATE. This mechanic is the most dynamic part of the game. Depending on the level, the player is able to rotate the stage by different quantified degrees, or not at all. By rotating the level, the direction of gravity changes relative to the player, enemy, and level elements. In this way, floors can be rotated/transformed into slanted surfaces, walls, or ceilings. The level can be rotated to move enemies, balls, or free falling level pieces around. With a significant amount of skill, you can even use the ROTATE mechanic to make Jumpman hover and fly around. Gravity is the core dynamic in Jumpman and rotating the level makes gravity more variable and interactive than in more traditional platformers like Super Mario Brothers.

The design of Jumpman really comes together with the ROTATE mechanic. Understanding just how much this mechanic changes the nature of the platforming challenges comes with practice, experimentation, and experience. I believe that Andrew McClure designed the MOVE and JUMP mechanics with such a high degree of unfamiliarity to guide players subtly into understanding ROTATE. Because Jumpman can be thought of as a Mario style platformer, there is a great tendency for players to to enter a new stage, see the arrangement of platforms/enemies, and try to reach the goal without using the ROTATE mechanic at all. If we could put Mario from New Super Mario Brothers into Jumpman, you wouldn't need to rotate most of the levels at all to reach the goals. But by making the basic movement and platforming abilities less direct and variable, players are gently guided away from "doing the Mario." Instead of JUMPing around to reach the goal (which can be quite tricky), players begin to experiment with the other primary mechanic of the game. ROTATE. For example, the level on the right forces players to ROTATE instead of platform.

Enemy Design

The enemies in Jumpman can't be destroyed by Jumpman. Basically, each enemy is designed to take up a unique design space according to how they move and how gravity affects them. From left to right, Spiny moves back and forth and falls according to gravity. Sticky moves back and forth yet moves regardless of gravity like a sticky bug or the common wall-crawler Metroid enemies. Hunters use horizontal pathfinding to attack the player and is affected by gravity. Swoopy fly around at 45 degree angles to gravity. The Angryball can roll and bounce around due to gravity. Because at the slightest touch, Jumpman is instantly destroyed by these

enemy elements, each element functions as considerable line of contrary motion. In other words, if you fail to avoid them, you'll never beat a single level. In fact, the only way to interact with these enemy elements is by ROTATing the level to move them around, or by guiding a Superball (yellow) or Happy Ball (orange) into them to destroy them. In these ways, the enemy design is very basic, yet effective even though the interplay only reaches level 1 (one counter) at most.

One issue I noticed with the enemy design is that there are many cases where their movement patterns seem to be artificially altered just to create specific level challenges. For example, some Sticky and Swoopy enemies appear to hit invisible walls causing them to change direction. Normally, these enemies can only change direction after running into another enemy or solid surface. These irregularities in enemy movement work against the paired form and function of the enemies. Just think of what would happen if you saw a Goomba that started jumping around in Super Mario Brothers.

Power Ups, Suspension, and Transformation

There are no powerups in Jumpan. You might consider the Happy Ball as a powerup as it allows Jumpman to kill enemies and extend the height of his jumps by bouncing on top, but I like to think of all the balls as enemy/level elements.

There is very little suspension in Jumpan. Each time you beat a level, you are transported to the next level. There's nothing in the game that you take with you from level to level but yourself. The little suspension in the game is contained within each level. Each time you die in Jumpman and you respawn, the rotation, enemy positions, and level element positions remain the same. So, if you rotate the level so that all the enemies pile up at the bottom and you die, you'll respawn with all the enemies gathered at the bottom. In the example image above, touching the orange bomb causes the "cage" to fall apart (1). As the cage is falling (2), I touched the colored sides of the level and died. When I respawned, I landed on the broken "cage" pieces (3). Then I rotated the stage standing the pieces up along the right side of the level (4). From there I moved toward the goal and died. After I died, I respawned close enough to the goal to move left and touch it.

Between rotating a level to change the function of platforms from floors, to slanted surfaces, to walls, to ceilings, rotating the stage to reposition enemies and other level elements, using bombs to break away platforms, and grabbing paint brushes to change the enemy/level elements that Jumpman can and cannot interact with, Jumpman is packed with a considerable about of transformation. These transformative design elements along with the physics engine help create a highly variable and responsive gameplay experience. Essentially, any given level can be rotated 8 ways to create 8 different platforming challenges. On top of this, any given level can be designed with up to 16 layers players can travel between by grabbing paint brushes. In this way, multiple levels can be stacked on top of each other to create a kind of Z axis depth. This design feature significantly increases the potential for challenges in the level design (see image above). On top of all of this level potential, every time you try to control the stage with the ROTATE mechanic, some enemies will change relative to gravity, others with change relative to the stage changing the nature of the level challenges even further.

In the end, Jumpman has a few features that are questionable with its mechanics, enemy design, and difficulty design. However, the clever level design (especially in paths 7-10) and the transformative potential of the dynamic ROTATE mechanic focus the game on what it does best while avoiding the questionable areas in its design. After playing through Jumpman about 2 times and fiddling around with the level editor, the game has won me over. Jumpman is a perfect example of how powerful level design and transformative mechanics are in shaping gameplay to improve dynamics, depth, and replayability. Without power ups, "coins," much suspensions, or much interplay, Andrew McClure has succeeded in creating a quality platforming video game around one dyanmic mechanic, transformation, and suspension. It's not the way Mario would have done it, but it's enough.

I'll end with a few quick looks at some of my favorite levels from Jumpman.

  A folded level using paint brushes. The colors take you floor by floor until the bottom. If you rotate the stage, the enemy positions can be rearranged to make the challenege easier after the crease.

Another folded level using paint brushes. If you turn the level upside down at the start, you can simply walk to the goal for an elegant solution.

Yet another folded level using paint brushes. Tricky falls all the way down. On the way back, all of the purprle Swoopy enemies become threats.

These levels are completely dark. Use the ROTATE mechanic to make the enemies slide around and reveal where the platformers/holes are.

If you try to platform to the goal on this stage, you're in for a challenge. But if you rotate the level, you can turn the floor into a slant and use it as a spring board to soar up the left side and reach the goal for an elegant solution.

Spelunky: A Game Design Gold Mine

Spelunky is an indie platforming game by Derek Yu. I stumbled across it, downloaded it, and was quickly absorbed into the game. The best part about this game isn't that it's free. It's that the design is solid according to the principles of Classical game design. Uncovering the intricacies of Spelunky requires a close examination of the game. Fortunately, we are well equipped with the language and the theory necessary. Let's drop a rope into this investigation and start pointing out the design gems.

Structurally, Spelunky is very similar to Super Mario Brothers. Both are 2D, contain platforming elements, enemies that can be killed by jumping on them, and power ups. And like Super Mario Brothers, the player, enemy, and level elements are designed to create a triangle of contrary motion. To understand how excellent the design is of Spelunky, we must understand how the core of the game works.  First up...

PLAYER MECHANICS

• Primary mechanics: MOVE, JUMP, BOMB, ROPE
• Secondary mechanics: WHIP, USE/ACTION, CLIMB, DUCK, RUN, FLARE, LOOK
• Tertiary mechanics: PICK UP, DROP, THROW

In general the player mechanics are all very intuitive, direct, and independent. This is almost irrelevant considering that Yu gives players the ability to customize their controls or use gamepads. I synched up my Wiimote to my PC and found these controls to work quite nicely.

The most important and impressive part of Spelunky's mechanics design is how dynamic the BOMB and ROPE are. The BOMB simply explodes after a short period of time. Just about anything that gets caught up in the explosion blows up including the player, enemies, and the level. Blow up the sides of ponds to drain the water/lava. THROW a bomb at a shop keeper to set him off in a gun toting rage. THROW a bomb to hit targets like you're throwing a rock. It won't explode until the timer runs out. You can even use the light from the explosion to temporarily illuminate dark areas.

In fact, the entire game features an extremely high level of polish and attention to detail. A fire frog enemy will lose its flame if it happens to jump into water. Leaves from plant enemies flutter down and float on top of the surface of water. Upset a shop keeper and wanted signs will be posted of your face in every shop for the rest of the game. The angry shop keepers will even leave their shops and patrol the level exits hoping to catch you. BOMB shiny areas in the walls or floors and the gold will be freed for the collecting. The list goes on and on.

Another interesting design feature of the BOMBS and ROPE mechanics are that they're limited. Players start the game with 4 of each and, unless they find/buy more, that's all they get. Unlike in games like Super Metroid, players can't simply kill/farm enemies to replenish their supplies or find recharge stations. Supplies are much harder to come by. Working with a limited supply of these powerful and versatile tools (explained below) makes each one more valuable to the player.

ENEMY ELEMENTS

There are a lot of different enemies in Spelunky. Collectively, the enemies vary in size, health, strength, speed, movement pattern, rarity, and special abilities. One of my favorite design features of the enemies in Spelunky is some of the enemies can hurt/affect each other. The man-eating plants can eat cavemen. The fire frogs create explosions like normal bombs. The arrow traps can be triggered by and kill enemies. The spikes can impale shop keepers. The aliens can kill yeti. The list goes on. It's such details that help make the core design of Spelunky so dynamic.

LEVEL ELEMENTS

The world of Spelunky is quantified to the block unit like in Super Mario Brothers. The player avatar is the size of a block. Enemies are designed to take up space by the block. Boxes and chests are a block large. Even the WHIP, BOMB, JUMP, and ROPE mechanic are measured in blocks. Because of this design feature, everything in Spelunky fits together like puzzle pieces.

Spelunky is also features a type of suspension that is completely intuitive, yet has been missing from many games. When exiting one level with an item in your hand you get to take it into the next level. You can even grab a rock on the first level, carry it throughout the whole game, and throw it in the face of the last boss. It's your rock. The game doesn't get in your way of doing whatever you want with it. It's this simple level of suspension that is very effective at connecting the gameplay from one level to the next and across the whole game. Like in Super Mario Bros (more specially SMB3), you can hold onto some of the best powerups forever as long as you don't die/get hurt.

The gold,gems, pots, crates, damsels, and other level elements, like coins in SMB, give players incentives to go out of their way. In Mario's case, coins are arranged to encourage more difficult, interesting, and superfluous JUMPing. With Spelunky, the elements encourage platforming through exploration and terrain transformation.

LEVEL DESIGN

Level design is where the core design of a game is put to the test. Without levels, players wouldn't have a place to stand on, enemies to fight, or an environment to give context and relevance to the mechanics. The levels in Spelunky have a variety of features that accentuate the well designed core (player, enemy, and level elements) as well as keep things fresh for the player.

• Solid vertical gravity design. Spelunky features all the same dynamics of gravity as Super Mario Brothers. Each jump creates a commitment on the player's part. What goes up must come down. If players aren't careful about how they jump before they leave the ground, they might end up suffering for it. In Spelunky, almost everything is affected by gravity; gold, gals, shop keepers, enemies, rocks, arrows, etc. 
• Going beyond Mario, Spelunky uses the gravity as a key dynamic in the exploration. In general, it's easy to jump down from high places. If you're not careful, and the fall distance is too great, you'll end up taking damage and collapsing in a temporary stun state. By holding down, players can use the LOOK mechanic to see far below their position. If the drop is too great to jump, players can climb down on a rope. Otherwise, the only way to keep from getting hurt is to land on top of an enemy, a maneuver that can be very risky. So, though it's easy to go down, there is a dynamic element of danger involved.
• Going up creates a different dynamic. While bombs are great for blowing up holes and dropping down through them, ropes are the most practical way of platforming upwards. Unlike falling down where the player can air control left or right, climbing a rope puts players on a strict, straight path. Though this may seem limiting compared to descending, the rope is very long can stick anywhere (even in mid air). Players can also jump off from the rope. The contrast in the paths of motion possible for ascending versus descending are especially important in that they naturally create areas of folded level design. 

• Pure organic level design. In Spelunky, the objective of each level (except the last boss level) is to get from the start of the level to the end. The possible paths the player can take to accomplish this goal are at first very limited. However, by using bombs and ropes the level can be dynamically transformed. If there's a wall/floor in your way, consider blowing a hole in it. If there's an area out of reach above you or too far below you, consider using a rope to climb up/down. Each time the player uses either of these transformative mechanics, the change is permanent. In this way, new paths can open up, intersect with old paths, or even shut off the viability of old paths. Though the goal is the same, how you get there is very flexible, and dynamic.
• Along the way from point A to B, the player is likely to decide to go after optional elements thus creating their own goals. Chests, gold, damsels, and shop keepers are all examples of elements that the player will likely go out of his/her way for. Going after these new goals naturally create small areas of folded level design. For example, going down into a small pit to get a chest and then climbing back out are two very different platforming maneuvers. This is due to the ascending-descending gravity design. With every element the player seeks adjusting the difficulty of their play experience, the potential for the paths to overlap/intersect increase due to the level transforming mechanics.
• From the start of a level, every level and enemy element are set in place. Enemies don't spawn from some unknown area off screen. For the most part, everything within a level is persistent. Due to the high level of detail put into the interactions of the various game elements, the player is free to cause chain reactions, make traps, and create strategies powered by his/her own curiosity. The game world is alive and everything works for or against everything else.
• Because the player can only carry one item at at time, juggling more than one item creates natural folds in the level design that play off of the core dynamics. For example, in dark levels, the player can use flares to illuminate the surrounding areas. When holding the flare, the player can't hold anything including a highly valued shotgun, damsel, bomb, or rope. To navigate safely through the dark while keeping up with any number of these extra items, the player must juggle putting the flare down and using the items. One time, after setting up my flares in strategic places and escorting the damsel to safely, I went back through large parts of a level just to retrieve my shotgun. It's amazing that so much dynamic gameplay comes from only being able to do one thing at a time.

• Randomly generated levels. Most of the levels in Spelunky are randomly generated. This feature forces the player to learn and apply the game rules instead of memorizing how to get through each level. Players have to think on their toes to survive. The random level generation also gives the player a better chance to experience how rich and deep the emergence/counterpoint is in Spelunky. The more you play, the more unique and interesting combinations you'll run into. And with a game that's as dynamic and detailed as Spelunky, you'll be experiencing new things for a very long time. 

• The biggest problem I have with the randomly generated levels is that the combinations will either randomly work or randomly not work. Sometimes you are forced to play in an area that you need a rope or a bomb to progress. If you don't have either of these items or an platforming powerup (see below), you're out of luck. Unlike in Nintendo games, in Spelunky there are no infinitely repawning pots/enemies to resupply the player. One time, I entered a temple level only to find a Giant Mummy right behind me already spewing killer flies (see above for dramatic reenactment). I had no chance to react. Like in the game of Solitaire, sometimes you just can't win. Ultimately, there's a tradeoff between a level you design and a level that's randomly generated. The former can be carefully crafted and highly tuned, yet remain largely unchanged with multiple playthroughs. The latter is new every time, but can be very unfocused, unforgiving, and unfunctional.

• Permadeath. In Spelunky, there are no free healing stations. Your heath doesn't regenerate if you rest for a bit. And your life doesn't come back at the start of each level. What little life (4 hits) that you start with is absolutely precious. In a world where almost everything can hurt you from your own thrown rocks, falling from too high, to your own bombs, players quickly learn that life in Spelunky is fragile. Not only will you die when playing this game, but you'll die so much, so quickly, and so easily, that you'll laugh at how easy even the most "hardcore" games are these days. When you die in Spelunky, that's it. Game over. Better luck next time.
• A pro for such an extreme penalty system is that any accomplishment, no matter how small, is a significant accomplishment. Like in Perfect Dark, my favorite FPS, beating a level means you accomplished all the objectives without dying once. Because there's no respawning, save stations, save states, or check points the only thing that can get you to the end is your own consistent skill. Like playing a piece of music from beginning to end, there's a distinct level of appreciation that you gain from doing it all in one run.
• The cons for this extreme permadeath design is that the game is more deadly than fun and there's a significant amount of trial and error in the game. To investigate a new enemy/level element, you have to interact with it (assuming you don't check a faq/wiki page). This can be very dangerous. If you die while trying to figure out just how dangerous the a Giant Mummy is, for example, you might be set back to the beginning of the game without having learned enough to keep you alive the next time. Simply getting back to another Giant Mummy is a challenge in and of itself. After a while, the player wishing to succeed is discouraged from experimenting and "playing" around because one wrong step/move can cost him/her everything. Also, whenever the game glitches and kills the player, the amount of work that is unfairly taken from the player is extremely disheartening. I got to the final boss of the game only to have the physics glitch out and kill me. I was squished while standing on top of the head of the Giant Idol. I'll probably never beat the game now.

POWERUP DESIGN

The powerups in Spelunky are almost all designed to augment the players offensive, defensive, and movement (platforming) capabilities. Click on the link above for specifics on each powerup item.

• Decay: bomb, rope, parachute, bow, and the mattock. All of these powerups decay with use.
• Offensive: bomb, machete, bow, pitcher's mit, pistol, and shotgun. 
• Movement: spike shoes, spring shoes, teleporter, jetpack, cape, climbing gloves.

All of the powerups work in unique ways. Many of them can also stack on top of each other to create a super powered spelunker. Once you obtain a powerup, as long as you can wear it or carry it, you're free to take it with you to the end of the game. This kind of suspension is very similar to Super Mario Brothers with the biggest difference being that in Super Mario Brothers, one hit from any kind of enemy will take away your powerup. I guess there's some things that are harder in Super Mario Brothers than in Spelunky after all.

In the end, Derek Yu very much succeeded in creating a fast paced, roguelike platforming game. This game is interesting and stands a very good chance of ending up on my GOTY list by the end of the year. After playing and dying more than 300 times, I have many stories to share. And with a level editor included in the package, I am free to create platforming and puzzle challenges to my heart's content. As evident with Spelunky, the tenets and principles of Classical game design are what fill a game with a certain depth and richness. It is these principles that are of a very high value these days. For those of you who haven't tried Spelunky yet, I wish you happy hunting. You'll find gold no matter if you make it out alive or not.