The Icehouse Game Design Competition

Hello Gameshelf viewers! I'm Doug Orleans, Friend of the Show. I have not actually appeared in any episodes, but my copy of Acquire was in Episode #6. I was also mentioned by name because a game I designed, Pylon, was the winner of the Summer 2007 Icehouse Game Design Competition. The IGDC is the topic of this, my first post to the Gameshelf weblog.

The Icehouse Game Design Competition was started by the Icehouse mailing list community in 2004 as a way to encourage people to design (and play, and give feedback on) new games using Icehouse pyramids. It was inspired by similar game design competitions run by the Piecepack and Interactive Fiction communities, which had been pretty successful in generating and maintaining interest in creating new games. I don't remember who actually suggested the IGDC first (the mailing list was not archived back then) but the Gameshelf weblog's own Andrew "Zarf" Plotkin got the ball rolling by serving as the first competition administrator. Zarf ran the first four competitions; in 2005, the fifth competition was run by Jonathan Leistiko, but it was abandoned and the votes were never released.

After a couple years of apparent non-interest, the IGDC was revived in 2007 by David Artman, who ran the sixth and seventh competitions. The eighth competition was recently announced by the new administrator, Dale Sheldon; game submissions will be accepted until June 20th, 2008, which will be followed by a one-month open judging process in which anyone can vote on the games by ranking them all from best to worst. If you want to try your hand at designing an Icehouse game for the competition, create a page for the game rules on the Icehouse Games Wiki and send mail to (But first you should read the full competition rules.)

One of the main goals of the competition was to encourage people to give feedback to game designers about their games. To that end, I am going to post my thoughts here about the games in the Winter 2008 competition. I did manage to play all 8 games, but most of them only once, so take my opinions with a grain of salt. I'll discuss them in the same order I ranked them on my ballot, from best to worst, along with my ratings on the 10-point scale (10 being "outstanding", 1 being "clearly broken").

Virus Fight - 8. The old programming game Core War involves writing computer programs that run in a shared memory space and attack each other by overwriting each other's code. Translating this idea to a turn-based strategy board game turns out to work surprisingly well! Each pyramid represents an instruction based on its color: yellow is "move", green is "write", blue is "jump", and red is "erase". Each player has a program counter, a small pyramid that moves around the board executing instructions one at a time. Unlike Core War, which is completely deterministic after the initial selection of programs, in Virus Fight the players have choices of how to execute the current instruction; for example, a "write" instruction lets the current player place any spare instruction onto any empty square. A player can also choose where to move his program counter after executing an instruction, using the two-dimensional nature of the board instead of the linear memory model of Core War. Despite these differences, much of the spirit of Core War is still here: players' programs start out in separate regions of the board, but can move and expand and merge with each other, allowing players to invade each other by moving their program counters into each other's programs, and the "erase" instruction allows players to remove certain instructions from the board, reducing the other players' options. The goal is to force all your opponents' programs to die by having nowhere to move their program counters.

Aside from the initial simultaneous selection of programs, there is no element of chance in Virus Fight, and when played as a two-player game (which I think is the best way to play, in order to avoid the negotiation and king-making that can occur in multi-player games with no chance) this is a pure abstract strategy game akin to chess or go. I can appreciate most pure strategy games for their elegant mechanics, but in general I don't actually enjoy playing them—I don't have the patience to mentally search the game tree looking for the optimal move. For some reason, though, I really enjoy playing Virus Fight. Maybe it's because it's quite difficult to look more than a couple moves ahead, due to the way the turn order can change based on the relative sizes of the pyramids currently under the program counters, so you have to just play by intuition most of the time. Or maybe it's just that, as a programmer by vocation and avocation, the theme is a natural fit for me. In any case, the game works well, both as a well-balanced strategy game and in capturing the essence of Core War, and for me it was clearly the best game of the competition. Sadly, it did not fare well in the voting, perhaps due to its somewhat intimidatingly complex rules (which are nonetheless quite elegant, in my opinion). But if you like pure abstract strategy games, or programming, give it a try.

WreckTangle - 5. It's very difficult to design a simple abstract board game that has an element of chance without the chance element overwhelming the strategy. Backgammon is the canonical example of this kind of game, and I think it succeeds in part because the chance element only serves to randomly restrict your options at the beginning of your turn rather than to randomly determine if you succeed or fail after performing some game action (like in most wargames). (This distinction is sometimes referred to as "situational luck" vs. "resolution luck"; I can't find a citation for who came up with these terms, but I think I first saw them in The Games Journal.) WreckTangle uses the same idea: on your turn, roll the die (a Treehouse die), then make two moves, one of which is partially determined by the die roll; for example, "dig" lets you move a pyramid, either yours or an opponent's, one space diagonally away from its home row, and this can be done either before or after you move one of your own pyramids. This system works out pretty well: you still have a wealth of options on your turn, but sometimes you have to play the odds and hope that the next player's roll doesn't let him ruin you. This sort of blurs the distinction between the two kinds of luck, however, since these situations can feel more like you're randomly succeeding or failing, especially because, unlike in backgammon, your captured pieces are permanently removed from the game. And a single capturing move (by forming a rectangle—of any size—with four of your pyramids) can capture multiple pyramids, so this can be pretty drastic. Still, I liked the basic idea of WreckTangle, and maybe with some sort of tweak to reduce the potential for drastic swings of fate this could become a solid game.

Hunt - 5. Hunt involves moving a stack of pyramids around a maze of dangerous obstacles, trying to stay alive while positioning the obstacles to kill your opponents' stacks. This game is in the same general class as Wrecktangle, and most of what I wrote about that game applies to Hunt as well—Hunt has slightly less chance of a drastic swing of fate, because sometimes the damage to your stack can be healed, but it also has slightly more of the feel of resolution luck: sometimes your die roll dictates that your stack immediately take damage because you can't make the required move. Fortunately, your stack is immune to damage caused by obstacle pyramids of the same color as the top pyramid in your stack; each time do you take damage, though, you remove the top pyramid of your stack, which means you become immune to a different set of obstacles. There are also ways in which the order of the pyramids in your stack can change, which also changes your immunity. This is a clever mechanism, leading to some tactical positioning options, but you still don't have quite enough control over the element of chance to implement any kind of long-term (or even medium-term) strategy. The rules also seem to be a bit more convoluted than they need to be, which is why I decided to rank this slightly lower than WreckTangle.

Martian 12s - 5. A push-your-luck gambling game, like Blackjack with a few twists: there are essentially three different "decks", and you can choose which one to draw from; also, "card"-counting is not discouraged! This game is simple and to the point, and it works perfectly well for what it is, but it just doesn't excite me at all. I'm not a big fan of push-your-luck games, but that isn't really my problem with Martian 12s; it seems like there isn't much else going on besides calculating the odds and deciding the appropriate level of risk to take. This turned out to be the winner of the competition, which I'm okay with—I would have preferred Virus Fight, but Martian 12s is a solid game that works as intended, and that's really the main goal of game design, isn't it?

Timelock - 3. Here we start getting into the games that had serious problems. Timelock is structurally quite similar to backgammon: your dice roll determines how far you can move your pieces towards the goal, and you can sometimes interfere with your opponent's progress. The problem is granularity: in backgammon, rolling high numbers is generally better, since you can make further progress towards your goal, but sometimes you want a specific smaller number so that you can land on a particular spot; in Timelock, however, rolling higher numers is always better, and for the most part it's simply a matter of who rolls the highest total over the course of the game. There are some slightly non-linear decisions about whether to make progress towards your goal or to block (or undo) your opponent's progress, and some Treehouse die results are better in some situations than others. But most of these decisions are obvious, and there's really not enough of this non-linearity to matter, so it pretty much boils down to a dice-rolling contest.

Timberland - 3. This hybrid of Treehouse and Volcano seemed like it had promise, but turns out to be overwhelmingly random—even moreso than Treehouse, which aims to be the Fluxx of Icehouse games. Volcano is a Nim-style game, where all the pieces on the board are shared between the players and you're trying to arrange for the best captures to be available on your turn. But when you add the randomness of rolling two Treehouse dice on each turn, it's pretty much impossible to arrange a configuration that will have any chance of surviving until your turn, even in a 2-player game; it's difficult just to minimize the next player's chances of making a capture on his turn, even if your options weren't restricted by what you rolled. In theory, I like the idea of adding an element of chance to Volcano, but this isn't the way to do it.

Chicken Run - 2. This game is similar to Timelock (and backgammon), but it has the same problem: higher rolls are always better. In fact, you only get to move at all if you roll higher than your opponent, which makes it boil down to a series of dice-offs, rather like War. The only interaction comes from moving neutral pyramids to block your opponent or unblock yourself, but they can only be moved if both players roll the same number, which only happens one sixth of the time on average. This means they play a very small part in the game, and in fact the first time I played I think we only got matching rolls once before the game was over. I won't go so far as to say this game is broken, but it would need some significant changes before it would really work.

Martian Gunslinger - 2. There are some interesting ideas in here, sort of a resource management/exploration/dueling card game with an intricate Western theme (...on Mars). The problem, once again, is the overwhelming amount of luck, especially resolution luck: when you "attempt a plotpoint", you draw two cards, and you succeed only if the first is higher than the second. Also, it feels like an afterthought that this game involves Icehouse pyramids at all: one part of the rules suggests using them to keep score, encoding numbers based on stacking configurations, while another part talks about using dice to keep score. (There are pyramids on the board as well, but these could just as easily be pawns, or painted miniatures.) The rules are also ridiculously complicated, with each playing card representing a different action, resource, and event, based on three lookup tables filled with text descriptions of what they do. Even if this were a custom deck of cards, there probably wouldn't be room on most of the cards for the explanatory text! There might be the germs of a decent game buried in here, but it's probably not an Icehouse game.

This entry was posted in Uncategorized Bookmark the permalink.

Leave a Reply

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


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>