Wednesday, June 26, 2013

More Randomness

Last weekend, Mark Rosewater did his podcast on randomness. (I transcribed it here, but I really recommend listening if possible.) His article on the topic covers roughly the same territory in fewer words but doesn’t mention anybody getting punched in the face, so... you know. Pros and cons.

Rosewater talks about some of the upsides of randomness: namely “It creates surprises,” “It makes the game play differently,” and “It allows players to react.” He mainly focuses on the ways randomness creates excitement and novelty. (If I recall correctly, his guess for why people are interested in the unknown is more or less dead right according to psychologist Mihaly Csikszentmihalyi’s book Creativity.)

Listening to this, my neuropsych side got to thinking about other possible benefits of randomness. I am fascinated by the neuropsychology of games, and how our abilities in various areas affect what games we enjoy and/or are good at.

Recently I was playing Robo Rally, one of my all-time favorite games. For those who aren’t familiar with it, Robo Rally is a game by Richard Garfield (designer of Magic: The Gathering) in which the players race their robots through an obstacle course. The board is full of elements like conveyor belts, pushers, and rotating gears to knock your robots off track. Since you have to plan out your next five moves in your head, taking all of these into account, it is very easy to accidentally wind up somewhere you weren’t expecting. Complicating things further is the fact that other robots can bump into you, spoiling even the best-plotted courses.

Now, I normally go way out of my way to avoid visual-spatial games. Chess, Othello, Connect Four—my brain just doesn’t work that way. I think this may have something to do with the fact that I don’t use pictures to think at all. I think almost entirely in words, with a few vague and tenuous spatial concepts. If I ever drive you somewhere, you’d better be ready to navigate.

However, I love Robo Rally, and I think the reason has something to do with the other major mechanic: the move cards. You’re given nine cards (fewer if you’ve taken damage) that say things like “Move 2” or “Turn Around.” You play five of the cards, and then you’re dealt a new hand of nine. And somehow, this mechanic makes all the difference.

This brings us to one aspect of randomness that Rosewater doesn’t really touch on—the fact that randomness actually limits how much information you need to really need to consider at once. In chess, you’re generally supposed to think ahead by a certain number of moves. Kasparov reportedly said that he thinks three to five moves ahead, and that in certain situations it is possible to go up to twelve or fourteen.

To be able to consider that much information, particularly hypothetical branching paths, requires a strong capacity in what is called executive functioning. Executive functioning is actually many separate but related skills which combine to effectively be the manager of the brain. To give you a sense of what is included in this group, here is a list (bolding is mine) from LDonline, a learning disabilities website:

Normally, features of executive function are seen in our ability to:
·            make plans
·            keep track of time
·            keep track of more than one thing at once
·            meaningfully include past knowledge in discussions
·            engage in group dynamics
·            evaluate ideas
·            reflect on our work
·            change our minds and make mid-course and corrections while thinking, reading and writing
·            finish work on time
·            ask for help
·            wait to speak until we're called on
·            seek more information when we need it.”
When all information (besides your opponent’s choices) is available for the foreseeable future, as is the case with chess, the person who can think way ahead (i.e., the person with strong executive functioning) is going to have a major advantage. In addition, a chess board offers… well, I can’t calculate off the top of my head how many possible moves there are in a given turn, but it is a whole lot. Just comparing all of your options requires holding a huge amount of information in your head at once. Alternately, you may be able to recognize which moves are irrelevant to the task at hand and ignore them, but screening out irrelevant stimuli is also part of the executive functioning package.
As is typically the case for people with ADHD, most aspects of executive functioning are really not my strong suit. While it’s good enough to get me through the day-to-day, games are designed to push the limits of our abilities—and something like chess pushes mine to the breaking point. However, in Robo Rally, drawing a random hand every five moves means the value of planning beyond that is significantly curtailed. When a random component is introduced (as in the case of Robo Rally), weakness in that area is no longer the huge handicap that it was. In addition, mistakes have a random chance of actually being fortunate, reducing the sting of badly chosen moves. Magic is similar in that you can consider what’s in your deck and play to your outs, but mostly you are reacting to what is in front of you.
The limited hand size isn’t technically part of randomness, but the random draw helps the “ridiculous number of moves to consider” problem by limiting the number of options you have to consider. If you’ve ever played Mental Magic you know how overwhelming it gets when you can choose from any card ever. This can also be seen in the fact that things that reduce randomness, such as tutoring and library manipulation, can be satisfying but also multiply the number of options you have to consider in order to make the optimal play.
Rosewater says that people think randomness makes games less skill-testing, and argues that in reality it makes it more skill-testing. The truth is, neither position is right—random and nonrandom games both test skill, they just test different skills. While drafting is a skill, so is being able to read the metagame, learn your deck inside and out, and generally plan ahead for Constructed. They cater to different strengths—careful planning vs. thinking on the fly. The more randomness is introduced, the less planning matters and the more cognitive flexibility matters. (Unsurprisingly, one of my friends called me the most exclusively Limited player she has ever met.)
Randomness is exciting. It can make games more emotionally engaging and intellectually stimulating. And paradoxically it can stop them from being overstimulating, limiting the amount of information we are flooded with so that we can focus on what’s fun and interesting to think about. Like I talked about in my Understanding Complexity article, one of the cool things about Magic is the huge variety of cognitive processes it draws on. As a result, relative strength or weakness in one particular area does not result in the serious beating it would lead to in a game that fully emphasized that area. Randomness is one of many tools that keeps the game accessible to all.


4 comments:

  1. IMHO the main issue with MTG with regard to randomness is not the fact of having to react to something unpredictable, it's more the fact that winning or losing depends so much on your draw. How many times did you ever play a match that was unwinnable because of a crappy draw on your part, or a nut draw on the opponent's part. No matter how you design your deck, you will always have bad draws that are much MUCH worse than your best draws.

    Besides, randomness has not much to do with the branching factor of a game. Here's a purely deterministic casual constructed format. I call it Stacked Deck Magic. It is a format in which instead of shuffling decks and getting them cut by the opponent, you get to put them in whatever order suits the player (the order being known only by the player). Cards which flip coins are banned (so are cards which would be broken in the case of stacking rather than shuffling, such as Unexpected Results), cards which say discard/reveal cards at random disregard this "at ramdom" qualifier. Do you have that many more choices in such a game. During "shuffling", yes, but most of the choices do not matter much (should I put this card on the bottom or next to the bottom). Obviously, such a game would be broken (everybody would combo out at turn 1 - at least in legacy).

    ReplyDelete
  2. While it is true that your choices would matter less and less as you got closer to the bottom in Stacked Deck Magic, I disagree with your projection of how the games would play out. If the person on the play tried to combo out on Turn 1, wouldn't his or her opponent just fill their starting hand with Force of Wills and Mental Missteps?

    It would become a game that's pretty much all about the "shuffling" and a rock-paper-scissors metagame, but preparations would involve considering a huge amount of branching paths.

    ReplyDelete
  3. Yeah, that is something I always thought too. I came up with that answer when one of my friends asked me why I prefer say Dominion to Go. The sheer amount of pressure a game like Go produces is insane... It's cool that my pet theory is confirmed by psychological studies.

    ReplyDelete
  4. Oh yeah, great example! And to take that further, compare Dominion and Ascension.

    ReplyDelete