 |
|
|
| |
Analysis: McMillen On Risk & Reward In Video Games
by Edmund McMillen
|
|
| |
|
May 7, 2010
|
| |
[In this delightfully-illustrated analysis, Super Meat Boy co-creator Edmund McMillen discusses ways to implement retro risk-reward mechanics, now that "lives" no longer matter.]
Last time, we talked about "Difficulty", something I might reference in this article, so you might as well check that one out if you haven't. But this week's topic is risk/reward!
So what is risk/reward?
Risk/reward is a system established by the arcade generation that rewards the player for taking a risk that goes beyond what they are asked to do normally.
Pac-Man uses r/r perfectly in many aspects of its design; not many think about it, but this is actually why the game is good.
So the first and most obvious aspect of r/r in Pac-Man is the blue ghost multiplier. When the player eats a power pellet, the four ghosts chasing you become edible for a small amount of time and attempt to avoid you.
Eating said ghosts will result in score points, and with each ghost eaten after the first, that score is multiplied.
The risk here of course is the fact that the ghosts will turn back to normal very quickly, so eating them becomes a race against the clock, if you're too close to one when they become normal you usually die.
So the risk here is a loss of a life, and the reward is a higher score -- but the substantial aspect of the reward is the fact that Pac-Man rewards the player with extra lives every 10,000 points, and this is what makes the r/r in Pac-Man important.
The player isn't ever required to eat a single ghost; it's simply an optional risk the player can take to rack up score and gain extra lives in the process.
R/R In Super Mario Bros.
Super Mario Bros. was the first console title to reinvent r/r in ways that had less to do with score, which had basically become irrelevant once games left the arcade.
The introduction of coins as things to collect to get an extra life is an r/r mechanic introduced by Mario that was blindly copied by just about every platformer thereafter. The basic idea behind coins in Mario was to use them as r/r that usually put the player in a dangerous situation, and if the player collected 100 coins they got an extra life.
Mario also used coins as a way to instruct the player on how to play giving them reference points on where to jump and also hints on where to explore, but I'll talk more about instruction through level design in another article.
The second example of r/r in Mario was the 1up mushroom.
Hidden within almost every level in Mario was a single block that, when hit, released a 1up mushroom. The introduction of the 1up mushroom was a great way to get the player to take their time and explore the levels a bit more.
Exploring each level for that hidden 1up added more danger, and also ate up the player's time limit. 1up mushrooms were also usually rigged to run from the player and/or into kill zones, requiring the player to take an even bigger risk to acquire them.
But as most staple mechanics in games, coins and 1up mushroom became easier to get and more abundant as the years went by, and the reward for going the extra mile to get them became pointless when the player no longer needed to worry about lives.
Infinite Lives?
So how could r/r be applied to a game that had infinite lives?
Every chapter in Super Meat Boy has 20 bandages to be collected: seven in the main world, six in the dark world, and six in that chapter's warp zones. The r/r formula with bandages is very similar to coins in Mario, with a few minor changes.
So every level in SMB that has a bandage requires the player to collect said bandage, then complete the level without being killed in the process. Bandages are placed in areas that would require more action from the player and put them in much more danger -- most importantly, the player isn't ever required to collect a single bandage to complete the main game.
What about the exploration aspect of 1up mushrooms? How can that be applied to a game like SMB?
So warp zones act as 1up mushrooms, in the sense that the player needs to explore levels to find them and they are usually put in hidden or hard to reach areas of the game. When a warp zone is found it is permanently unlocked in the level menu.
Now bear with me, because this might get a little wordy.
Every chapter of SMB has 4 hidden warp zones: 3 in the main game and 1 in the dark world. There are 2 types of warp zones: retro warps and character warps.
Retro warp zone:
3 levels, 3 lives per level and 2 bandages to be collected.
A retro warp zone is a series of 3 small (usually single screen) levels with a retro visual theme. The player must beat all 3 levels in succession with a limit of 3 lives per level to complete it. Each level set will have 2 bandages to collect, but the player only keeps those bandages IF they complete the warp zone.
This was not only a perfect way to bring back old school r/r, but also a way juxtapose the new established rules of difficulty (see previous article) with the retro formulas of the past.
Character warp zone:
Infinite lives, play as new character, beat the warp and unlock that character.
Each character warp in SMB starts with a cut scene introducing a new playable character from the indie community. The warp zone is again a series of three levels that must be completed to unlock that character to be used in the main game -- but this time, there's no life limit, and the player will only be using that character to complete the levels.
Every unlockable character in SMB plays differently, usually having a unique ability like double jump, floating or flight. The levels in the warp are designed in a way that forces the player to use that character's ability in order to finish them, in turn teaching the player what advantages that character could have over certain levels previously played by Meat Boy.
When the warp is completed, the character is unlocked and can now be used in most main game levels (excluding warp zones, boss fights and other special circumstances)
And that brings us to reward: now that lives don't matter, how can we reward the player in a substantial way that adds more to the experience in the same way lives would in the past?
Reward = Content unlockables = playable characters / new levels
Every 10 bandages collected unlocks something that changes/adds to the current game. and as mentioned above 1 of every 4 warp zones completed also unlocks a playable character.
Substantial reward is very important to help motivate the player to push themselves to get better and try more difficult challenges. The difference between unlocking a new way to play the whole game and unlocking a digital badge or hat for your character is quite large, but adding something as in-depth as a new playable character also comes with its own set of difficulties.
[This analysis was originally posted on SuperMeatBoy.com, which has more information on Edmund's upcoming IGF-nominated XBLA, WiiWare and PC action platform title.]
|
| |
|
|
This comment is not meant as a thread-jack, but as a way to try to get folks to think more broadly about the risk-reward relationship. While I believe that designing with risk-reward in mind is a great way to stimulate and guide player behavior, players are not monolithic in their thinking. This means that they may have different perceptions about the level of risk involved and the desirability of reward.
I played a bunch of SMB (and was a reviewer on the IGF competition). I loved, loved, loved the concept that death no longer had meaning -- hey, Planescape Torment is one of my favorite games of all time. But I had to quit the game before I completed it because the level of risk was close to infinity for someone of my skill and ability and the reward was... well, more of the same content that was maddeningly difficult.
What I did enjoy (and a reason why I still gave the game high marks even though I had to quit) was the great user feedback provided by the game so that I could learn from my mistakes. Blood trails from previous failures provided a great way to "tag" the environment and learn where to place your marks for difficult jumps. The replay feature that showed all previous attempts simultaneously was funny to watch and hugely helpful when it came to analyizing a level and coming up with a strategy.
But, alas, even knowing what I had to do was not enough for me to succeed. I wanted to experience new content and characters, but the problem was that the overall risk-reward curve (kind of a balance/progression/challenge curve) was too steep. I couldn't get to the next plateau (and receive new content) on a schedule that made the experience reinforcing and, in turn, motivated me to continue.
I wonder if the underlying issue is that, really, only the "best players" are rewarded for taking the extra risks. This was true in the old side scrolling Mario games as well as in SMB. Unfortunately, that means that people like myself who aren't the "best players" miss out on certain content and experiences. This raises a whole other issue of balance and tuning. How do you make games for me without diluting the experience for your core audience of "best players"? What would a "casual" mode look like on SMB? Does it even make sense? If it does make sense to try it, how would you do it without spoiling the core game play and making it so easy that players don't get any sense of risk?
Great article!
Unfortunately, that doesn't sound like a workable solution in Super Meat Boy, especially if they have characters that already have floaty effects. The risk/reward system sounds very tight, and change would probably through multiple systems out of wack.
I personally think the only way we could give weight to achievements would be to only award them for things that ask the player to go above and beyond what they are normal asked to do. I believe Achievements were based off the Atari badges they would send out to players who could hit extremely high scores and then send in proof, they were given out to people for going above and beyond. I think a reward like this only holds weight if it something not everyone has, and giving out achievements for simply completing level one end up cheapening their meaning.
That said i do think there is hope in leaderboards as a type of social reward and it something we are currently working on with the xbla version of SMB.
Im glad you liked SMB, but i do have to point out that the build you played was only a few months old and had no real progression unlocking or anything more then the 1st chapter and some of the 2nd.
The point you bring up though is something i talked about in my last article about difficulty. Its my personal goal as a designer to make the main game in SMB something that most casual players will be able to complete if they put in the time and effort. I believe if i do my job i can curve the difficulty and teach them in a way that will hopefully help them grown into the difficulty and finish the main game.
I dont want to make a casual game, i dont want to cheapen the experience by making the game too easy, its important for the game to be hard. We will have to see how this all plays out in the end, but with a ton of testing i think we can get a good idea of how people play and learn and try to fix up things to make sure the mistakes made were at least not on our side of things.