Data and Player Behavior in Reality Ends Here

One of the first tasks for prepping Reality Ends Here each year is to devise the cards that’ll be used in the coming semester. The task itself is divided into two main content areas: collectibles and gameplay. The former has traditionally taken the most amount of time. Gathering 100+ different types of documentaries, famous directors, influential games, etc. each year, while not difficult, is time-consuming. This holds true this year as well.

11008FB

Card 11008, Playable / Gameplay Side (Front) and Collectible (Back)

This year, however, I decided to monitor the data from the collectible side as I was creating the deck. What started as a means by which to keep the correct “Out” ratios (see the above card on the left — the green sections with out arrows are the “outs”, and cards can have between 0 and 3 of these on a card) soon became a balancing effort, and some observances on how simple data ratios could affect play [if you’re not aware of how to play, check out this site, which explains it well.

Stats + Observed Behaviors

Maker-Property Card Ratios [attempting to] Predict Deal Size

The first stat observed are the ratios of different card types in the deck (between Green Maker cards, Pink Property cards, and Blue Special cards). In the first year of the game, the ratios broke down along these lines:

  • Maker – 50 Cards – 18.1%
  • Property – 221 Cards – 79.8%
  • Special – 6 Cards – 2.2%

Ignoring the special cards, this may imply a ratio of maker-property cards submitted during the first year to be just below 1-5. That is, each deal having 5 property cards and 1 maker card attached to it. Of course, this wasn’t the case. During the first year, each deal had an average of ~17 property cards, with a median of 6. During the second, it rose to ~22.5 and 11. In the end, they don’t predict deal size at all. A better number is needed.

Out and In Ratio to Limit Resources

One statistic that became a bit more important were the distribution of out cards and the resultant In-Out ratio. In 2011, these were the ratios:

  • 0 Out – 82 Cards – 39.6%
  • 1 Out – 60 Cards – 29.0%
  • 2 Out – 49 Cards – 23.7%
  • 3 Out – 16 Cards – 7.7%
  • In-Out Ratio – 191-256 OR there were 65 more outs than ins per 300 card deck.

The last number has huge gameplay implications. Each deck we put out into the game — and we put out multiple decks over the course of the semester — pumped more and more possible connections while not filling their potential. This would imply, and this behavior was observed over the two years of play, that players were not limited by the configuration and physical layout of their deals, but rather their sheer number of cards. Theoretically, without the creative constraints, players could create an infinite web of cards. I’m trying to close the -65 number for this coming year — albeit not as well as I’d like.

The Imbalance of “Alt” Projects

One of our side goals in this game was to introduce students to new forms of media and narrative creation by filling the game with interesting and new forms of media. While we would include common forms of media, such as 5-minute shorts and trailers, we also populated the game with “alt” media forms, such as ‘character artifacts’ and ‘walking cinema’. While we anticipated the more common forms were going to be more popular, we had also hoped that players would experiment wildly with the various maker types, and that at the end, there would be a nice even distribution of media.

The actual breakdown of usage was much more interesting. Firstly, 15 different maker cards were never used during the first year. These were, in general, much more experimental forms of media: city symphony, walking cinema, faux newsreel, to name a few. There are some that are surprising, such as remix video and snippet short, that hold the same point value as their more-often used counterpart but aren’t used. On the other end are the not-so-surprising cards: scene and tableaux are both used 9 times, while things like comedy sketch, silent short, and 30-second short aren’t far behind.

The 1-off used maker cards were where most of the interesting experimentation, and often the most rewarding projects, were made. In here we find two main types of maker cards: an interesting or large application of a classic form, or a somewhat off-the-wall card. Anecdotally, I know that the latter type was a result of someone genuinely curious and wanting to experiment with the form. A great example of this is the drawn-on film, which, by connecting said student with Dino, the archivist, resulted in a really interesting sharpie-based addition to an old film reel.

MakerUsage

Distribution of maker card usage (y-axis: # of occurences of usage frequency, x-axis: frequency of maker card distribution).

Application(s)

Future application of this knowledge, admittedly, is somewhat limited. The biggest problem is the scale of the game: if any of my speculations about data vs. behavior is correct, it would be years before I’d have enough statistical data to back it up. It’s mainly, at this point, driven by intuition, experience, and math. Still, some insight will hopefully be gleamed through this data not only for research (Ben Stokes has some awesome analysis that will be sharable soon), but to reflect and iterate on the design in the future.

Leave a Comment

Your email address will not be published. Required fields are marked *