Carl Muckenhoupt on Best Puzzles

Carl Muckenhoupt was the creator of Baf’s Guide to the Interactive Fiction Archive, one of the first websites devoted to IF during the 1990s and the ancestor of IFDB. In 2001, he wrote The Gostak, one of the more extreme experiments in IF. Today, he works as a programmer for Telltale Games.

The Best Puzzles nominees for 2013 were Captain Verdeterre’s Plunder, Coloratura, and Threediopolis.

Coloratura (Lynnea Glasser)

It’s hard to say anything about the puzzle design in Lynnea Glasser’s Coloratura that the author hasn’t already said, in extensive notes on her blog, but I’ll try to describe it from more of a player’s perspective.

The biggest and most significant puzzle in the game, although not the most difficult, is simply figuring out what’s going on, what your relationship is to the world, and how you can affect things. The player character, a sort of psychic gelatinous thing from the ocean depths, is extremely nonhuman, and both perceives the world and interacts with it in extremely nonhuman ways, which you slowly learn about through the first half or so of the game. The PC is learning during that time too, but mirror-wise, about the human things that the player is already familiar with. There’s a lovely moment early on where the ignorance cuts both ways: at the text’s urging, you “help” a number of sleeping humans by “untethering” their minds, leaving the the player wondering: Did I just kill them? At that point you don’t know, and the PC doesn’t either, due to not understanding how humans work. That kind of summarizes the tone of the piece as a whole.

It’s also an example of a broader technique: alien perceptions and expectations let the author describe things obliquely, not just obscuring what’s going on, but making the player figure out what mundane objects are. For example, a clothes dryer is described with the words “A strange sort of crude metal prison. It hums and vibrates in an inefficient song whose only purpose seems to be excitement and temperature.” It’s a trick as old as Zork (or, in the form of the Literary Riddle, far older), but here, the premise justifies using it pervasively. This is more a matter of atmosphere than puzzle, however: the author takes care to make sure that we’re never stuck as a result of not knowing what things are, and a map provided with the game helps to establish the setting (an ocean-going ship). Also, the room descriptions largely change to more familiar terms once you successfully make contact with a human mind.

Nonhuman abilities, or inabilities, are a stronger factor. Your ability to engage your environment physically is limited, but not nonexistent. You can’t open doors, but you can slink through air ducts. You have no inventory, because you’re incapable of carrying anything. (With a bit of handwaving, you even keep this limitation when possessing a human host.) What you can do instead, as the player discovers after flailing about for a bit, is manipulate people’s auras, mainly by changing their color to produce emotional effects: red is fear, yellow is curiosity, violet is self-assurance and determination, and so on. So the core puzzle type here is in some sense based around interacting with characters instead of manipulating objects, but you interact with characters more as objects than as people. In fact, I’d say it feels a lot like inventory-based puzzles, but with the inventory replaced by your emotional palette. As such, I found myself treating the color list like an inventory, scanning through the list for promising suggestions when temporarily stumped, even occasionally trying every color on every available character — although, in retrospect, I only did that when I was supposed to be doing something else entirely.

What else can you do? Well, as mentioned, you can possess people, or at any rate one person. There’s one bit you can only pass by releasing your possession, which is a bit like closing a door behind you in an adventure game: something you don’t think of doing at first because it’s a clear reduction in convenience. And you can to some extent use your body to transport energy, such as heat or magnetism. There are two puzzles that rely on this, but I don’t think I actually solved either of them. The heat one was a complete fluke where I just happened to perform the right actions before knowing about the puzzle, but I just don’t remember doing what the in-game walkthrough says to do for the magnetism one.

And you know something? I probably didn’t. The author’s notes provide some interesting details about what’s going on under the hood. The game tries to automatically make suggestions about whatever puzzle you’re trying to solve, starting subtle and becoming more obvious the longer you spend on it. That much isn’t uncommon. The interesting part is that it doesn’t decide how long you’ve spent on a puzzle simply by counting turns; it has some built-in notion of what actions constitute failed attempts at solving a particular puzzle, and only counts those. So if you spend multiple turns poking at the ship’s radio in various ways, you’ll get help with the radio puzzle that you wouldn’t get if you were poking at something else in the same room. In some cases, if you spend long enough on a puzzle, it’ll even solve it for you, or automatically skip intermediary steps. Now, the idea of games not letting you solve their puzzles doesn’t appeal to me, but I have to admit that it worked pretty well here, possibly as a result of the way it was linked to specific actions: puzzles only got solved while I was actually trying to solve them. It’s likely that I received more invisible help than I was aware of. At any rate, it kept the story moving, and there were still plenty of moments where I felt like I had done something clever.

There was one and only one moment when I became stuck enough to need the in-game hint system. This happened because no system, however sophisticated, can prevent the player from being an idiot: I had failed to notice a connection between two rooms that was mentioned outright in the room description. I think the map actually hurt me here, because it didn’t show the connection in question, as it wasn’t passable by humans.

Threediopolis (Andrew Schultz)

Threediopolis, by Andrew Schultz, is a clever little word game, although this may not be obvious at first. Much like Shuffling Around, Schultz’s game from last year, it starts off by letting you figure out the central gimmick on your own, then just iterates on that gimmick. Threediopolis is a much purer and simpler expression of that formula, though, dispensing with most of the habitual mechanics of the text adventure, making room descriptions explicitly irrelevant and in fact randomized, getting rid of any form of interaction that isn’t directly related to the central gimmick. There are only a handful of commands allowed, and each can be abbreviated to a single letter. This isn’t just a UI convenience, it’s a major hint about the rules.

Since I can’t meaningfully discuss the game’s puzzles without giving away those rules, I will do so now. The setup: The city of Threediopolis is a three-dimensional 10-by-10-by-10 grid, through which you move freely in the six cardinal directions (north, south, east, west, up down). Your coordinates are given as a three-digit number, starting at 444; you can teleport back to your starting point at will. You have a list of errands, each given as a description followed by some coordinates, for example, “garden startup materials@326” or “sandy beaches@445”. But simply going to those coordinates usually doesn’t do anything, which is puzzling at first, because you basically have no way of interacting with the environment other than by moving. The trick is that it’s not just where you go that’s important, but how. To run any of your errands, you have to spell out a word with your movement. For example, to reach the “garden startup materials”, you move south, east, east, down, and south, spelling out SEEDS.

In fact, once you’ve spelled out enough words this way to convince the game that you know what you’re doing, it tells you that you can dispense with entering them as individual movement commands and just type in the word. This is a very welcome convenience. I know I’ve complained in the past about other word-building games that expect you to go through a whole complicated rigamarole for each word you build. In some types of game, you could justify this as building anticipation, but that assumes the game cares about drama, whereas pure puzzle games of this sort are all about the solving, and the less that stands between thinking of your solution and being rewarded for it, the better.

In a sense, once you’ve figured out the rules, it’s all just one puzzle. Your condition never really changes. You can solve the individual clues in order you like, with no explicit progression other than your inevitably running out of easy clues and moving to the harder ones. But some of the clues are quite hard. Some of the solutions are obscure, and some are even misspelled, like “SEEWEED”, to fit the letter constraints (that is, you can’t go SEAWEED because there’s no direction that starts with A). But you get a lot of help, if you can figure out that it’s there and how to use it. For starters, the words in the list are sorted by length, and that, in combination with the coordinates, says a lot. “Garden startup materials” is in the “kinda near” row, which translates to five letters. Since its coordinates are five steps away from your starting position — two steps south, two steps east, and one step down — you know with certainty that the word is some permutation of the letters “SSEED”. “Sandy beaches@445”, also five letters, is a little harder: the coordinates are only one step away from your starting point, so all you know is that there’s an E in the word, and that the other four letters must come in opposing pairs that cancel out: N/S, E/W, or U/D. Optionally, the task list will be sorted alphabetically by answer, making the words you’ve found provide a little extra information about the words you haven’t, crossword-style. This is the analytical approach, and the sort of thinking you have to do if you want to mop up every last item.

In addition, one location offers a choice of couple of gadgets to help you with the tougher words, similar to Shuffling Around’s anagramming help. Here, the gadgets deal with the question of what can be done from your current path: one reports on the number of words that you can still complete without going back to the starting point, another alerts you when that number becomes zero because you’ve strayed. But both gizmos have limited charges, and using either of them effectively requires using it repeatedly, so I personally never got much benefit from them.

The thing is, though, at some point you’ll start to ignore the clues and just spell out a bunch of words. There are only so many words that can be made from the six allowed letters, and it’s easy to think of a bunch of them without any more prompting than that constraint. You probably won’t be able to do this for every word in the list, but you’re not really required to get them all — you can get the best ending with only 90% completion.

Also, the game actually rewards finding words that aren’t on the list: many of them trigger amusing one-time events that have nothing to do with completing the game, but which are tracked by the SCORE command anyway. For example, if you go east three times, you hear a scream – EEE! After finishing the main game, you’re given access to a sort of New Game Plus mode for the true completist, replacing your errand list with a list of all these special events. This list includes even more misspellings and non-words than the first one. I think one of the kindest choices that the author made was to make this part explicitly optional.

Captain Verdeterre’s Plunder (Ryan Veeder)

Captain Verdeterre’s Plunder, by Ryan Veeder, is about escaping from a sinking pirate ship with a talking rat companion, grabbing as much loot as you can along the way. A single play-through is very short; if you ignore the treasure, you can accomplish your primary goal in five turns. But of course ignoring the treasure would be missing the point. When you reach the safety of your dinghy, the game emphasizes this by tallying up the value of everything you nabbed, then gives you a brief epilogue to tell you just how short of fabulously wealthy you are, encouraging you to try again.

None of the individual treasures are particularly difficult to get. There are some simple object-based puzzles like finding the key to an unmoveable strongbox and getting a loose pocketwatch away from a seagull, but a lot of things are just sitting there waiting for you to pick them up. As always, there’s a good deal of satisfaction in gathering diverse valuables from an environment that’s rich with them, but if that were all there was to it, I doubt it would be a Best Puzzles nominee. The thing that got people’s attention is the planning and optimization aspect created by time limits. Not only do you have only so long before the ship sinks, you also have only so long before each deck disappears under the water. In a particularly nice touch, the water level rises slowly, and things sitting on the floor disappear from the room description before things hanging on the wall or sitting on shelves. Thus, if you dawdle, you can wind up not even being aware of the existence of the treasures you’re losing access to. Getting the most out of the ship requires a number of exploratory sallies, where you’re free to waste turns poking about and trying things, followed by a pass where you put together everything you’ve learned.

Thus, the experience is something like the “accretive” games like Varicella, where you spend multiple play-throughs learning to be the player character. The idea of a treasure hunt within a large-scale optimization puzzle even goes back to venerable Colossal Cave, with its limited lantern battery. But there’s one thing that sets it apart from such predecessors: they were completable. There was always the promise that if you hoarded your battery effectively enough, you could get every single treasure in the cave back to the wellhouse. Captain Verdeterre’s Plunder very specifically and deliberately makes this impossible. You’re not aiming for perfection here, you’re trying to maximize. And that means you have the additional challenge of deciding which treasures are worth even trying to get. If you have to search a container multiple times to find a valuable item at the bottom, is it worth enough to justify the turns spent on it? You won’t know until you try at least once; the value of loot is only revealed at the end of the game. (Some objects turn out to be completely worthless, and those are the ones that your rat companion most eagerly urges you to not leave behind.)

The maximum possible score isn’t even known for sure. The author asked players to submit their best walkthroughs, and the reigning champion more than doubled the highest score found during beta-testing. Interestingly, this kind of takes the pressure off. You can take notes on how valuable everything is and how many turns it takes to get it if you like, but that’s optional. You can also just play through it a few times and try to do a little better each time and wind up with a reasonably good score and an epilogue that you’re satisfied with. There’s no win condition other than making it out alive, completism is impossible, perfectionism just a dream, so all you have to satisfy is yourself.

2 thoughts on “Carl Muckenhoupt on Best Puzzles

  1. Andrew Schultz

    Before I forget, I enjoyed reading about the other games mine was nominated with, too.

    I was very glad to read details about someone exploring the logical side of the game, as I didn’t really have any data on that. Your experience (I think) jibed with what I was trying to get for the player, a-ha moments combined with hopefully not too much logical grinding.

    I myself didn’t really have a grasp on how helpful the availableometer/adrift-a-tron were, or even which is more useful, because I visualized those post-comp & thus my testers couldn’t get to them. I wanted to do more than offer one cheap random hint, but unfortunately, they weren’t useful enough.

    Yeah, the scenery puzzles are tougher–writing the logic/walkthrough document for them without hand-waving took time. I figured it would be a waste of time to work on “advanced” mode for the comp proper, because many people took the full 2 hours for the tasks. I’ll just make an extra mention it’s all in the source code & I wanted a puzzle where you couldn’t quite grep a big file of words. It was sort of a vanity project, and I hope that people are still amused by the easier ones.

  2. Pingback: The Stack » Xyzzy Awards Best Puzzles Writeups

Leave a Reply

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