Friday, September 23, 2016

Roguelike Celebration, Notes on My Talk

I am back, at last, from Roguelike Celebration in San Francisco, and the longest trip I've taken in my whole life.  Thanks to the people who ran it for helping me to attend and for putting up with a hick from rural Georgia.  When I wasn't working on and fretting about the presentation (I was hacking away on it literally minutes before I presented it), I had a great time!

Here are all the talks of the conference, which might be useful to you even if you were there, since they had two tracks going simultaneously.

Other people have discussed the experience of being there. I did that kind of thing last year with IRDC US 2015, so I'll spare you all that this time and, instead, provide an addendum to my own weird little talk.

Carrying the vagueish title "Difficulty in random dungeon games" and noticably not carrying a description like the other talks, mine was, I believe disjoined and weird and rambling, with actual slides marked "Aside." They're not *really* asides, they're there for a reason, the slides being the skeleton that the connective tissue of my talk was supposed to join together, y'see. Whether I succeeded in that or not, well, I leave to you to decide.

Here is the talk (YouTube). Here are my slides (Dropbox link, Powerpoint)

Gachami (Copyright Konami)
Note that I didn't get through the whole thing; the presentation ran out of time right before the Bubble Bobble slide, the point of which was to demonstrate an obscure, but not truly random, play mechanic as an example of the kind of trend-causing behavior I was talking about. (If you get to the end of the slides you'll find a picture of Gachami, who is my favorite video game character of the hour. I heard that the rules of the Internet require me to provide that. Anyway, here's video of the game she's from, the ridiculous Gachagachamp.  It is not a roguelike.)

Speaking of unasked-for personal opinions, my favorite moment of the conference was when one of the Rogue devs (did I mention all three of the creators of Rogue were in the room at the same time, for the first time in 30 years? SO AWESOME!!!) asked the crowd if anyone there knew what Rog-O-Matic was, and of the hundred-or-so people there just two hands went up. One of them was mine; I've even written an article that half of which was on Rog-O-Matic! (I think it's in the book? Here it is on GameSetWatch.)

I am making this post because it is the culmination of ideas and thoughts I've been struggling to render on @Play for literally years. I'm not even convinced the main part of the presentation, on "Knowledge, Logic and Wit," is even useful, but people told me they liked it anyway. I could write several articles on these things, especially "Wit." Yes it's a terrible name, someone suggested "intuition" but that doesn't get to the heart of it I think? One could also call it general skill of playing video games.

Anyway, here are my comments on the talk:

1. I worked feverishly the days before the talk trying to assemble a jumble of thoughts into something coherent. This is why the title doesn't really fit the subject well; it drifted a bit in the days, indeed hours, before I gave it.

2. There is a whole essay to be written on how CRPGs have drifted fron pen-and-paper RPGs, often for the worse. It's the copy-of-a-copy phenomenon: early CRPGs were inspired by PnP RPGs, while later CRPGs were inspired by other CRPGs. These days, it's often CRPGs-inspired-by-JRPGs-inspired-by-older-CRPGs. (Remember: Yuji Horii was inspired to create Dragon Quest from seeing Ultima games on display on a trip to the US.)

3. There are multiple essays to be written elaborating on what exactly I mean by Knowledge, Logic and Wit. One thing I wanted to make clearer in the talk: "Wit," in addition to being not the greatest name, is also kind of a cheat, defined as everything that isn't the other two. It might be dividable into narrower categories. Also note that this classification of skill doesn't map as well to real-time and action games, in which reaction speed and decision making under time contraints also matter.

4. On grind: Now understand, I believe that if something is fun, it is its own justification. People play horribly grindy games of their own free will, so I have to conclude there is something enjoyable they get out of them. That doesn't mean that I think people should pander to this desire, the playing of basically empty games. (Some people who make horribly grindy games often think of game design in behavioral terms, like it were a Skinner box. I couldn't be more against that.)

My definition of "grind" is, something mechanical a player has to do to pass time in a game before he's allowed to get to interesting gameplay. Ideally players will have interesting decisions to make at low levels and high. The attempted elimination of gride is DCSS's aim, and it's admirable. (I even think they go a bit too far in it, but I can't fault them for trying.)

It is worth noting that many players (especially retrogamers) believe that Dungeons & Dragons itself is best played as a low-level game, that characters get too powerful at levels above six or so. Interested readers are directed to search out "E6," a variant of D&D that limits players to 6th level.

But what makes grind, grind? Aren't games played for enjoyment, regardless of the form it takes? I don't think so. Listen this.

One aspect of gaming that I hear no one talk about is their improving aspect. That is, playing games helps build alertness, ability to think under pressure, ability to judge situations, ability to look ahead to future game states, logical thinking, intuition, and a plethora of other skills that have real and direct applications for everybody! People who play strategy games very well are likely to be very smart, and people who strive to learn to play strategy games well will become smarter as a result. I don't mean this in an airy "Brain Training" way, but that well-made games push players to improve themselves as a basic function of their playing. Badly made games reward only luck, time, and/or money put into the playing.

This is why I said the human race will thank you for eliminating grind.  It's not entirely hyperbole; it literally wastes the player's time.

5. Every time I try to pin down exactly what wit is I don't do a great job. But there is something there.
Part of it, as mentioned previously, is general skill at gaming. When you've gotten good at Pac-Man, some of the skill transfers over to other maze-like, and even general action gaming. And some of the skill even transfers, slightly, potentially, to real-life situations. (See "improving aspect," above.) Some of it is putting yourself in the mindset of the game designer; detecting when something seems *too easy*, that sort of thing. Some of it is guessing story direction, like figuring out the villain ahead of time in Scooby-Doo. Some of it is subconscious pattern recognition.

Some is getting better with (player) experience, which I find to be a very interesting process in non real-time games....

6. So why is it useful to name Wit at all? Because, and here is the thing: when the moon is full, I think the purpose of roguelike design is to preserve the function of Wit in a game for as long as possible.
The nature of Wit is that it turns into Knowledge (conscious information about the game) with practice. In a non-randomized game, this is nearly immediate, which is why they aren't replayable. Once you know the game, if the game doesn't change on succeeding plays, the fun is gone.

Randomization obfuscates the processes of the game, changing the world on successive plays, which keeps it in the domain of Wit longer. This is not forever, it's only a delaying process. A walkthrough of NetHack doesn't tell you, step by step, how to win, but there are plenty of spoilers that explain almost everything about the game that can be explained.  If the processes that produced those items were somehow randomized, then those spoilers wouldn't work, but more would arise that detailed the meta-processes used, and so on, without end.  That doesn't mean randomization isn't a (very!) good idea, just that it delays the inevitable.

Even without spoilers and walkthroughs, them you can figure out some things about the game, pick up some Knowledge, but some of that is purposely made invalid when the game ends. But even if all the visually-signifying characteristics of the game changed from play to play, the underlying algorithms remain the same, which keeps the game consistent between plays, which players can subconsciously detect as Wit, but will eventually become Knowledge, which, when applied with Logic, is a surer means of winning.

This is because the nature of thoughtful play involves analysis, which is the process of formalizing successful approaches. This is why I am fascinated by games that cannot be so easily analyzed. Someone came up to me (I forgot who, sorry!) and suggested Go as a game that appeals to Wit very well, and yes that's an extremely good observation. Although analysis provides some headways into Go, the same things that keep computers from playing as well as expert Go players are the same things that keep it in the domain of "Wit," which seems to point at something fundamental. Machine learning folks may have a better name for it than I. Maybe "intractability?"

Anyway, that's it for the moment. Now that the monkey is off my back, maybe I can write @Play columns more frequently again! Until next time....


  1. I think your thoughts on Wit are a very interesting perspective. I'll be mulling them over as we approach this years 7DRL.

  2. Great Article it its really informative and innovative keep us posted with new updates. its was really valuable. thanks a lot.
    Subway surfer hacks