Quern: Undying Thoughts: design ruminations

Wednesday, December 7, 2016

Tagged: reviews, ruminations, puzzles, quern undying thoughts, islands, donkey puzzle, interactive fiction, if, obduction, adventures, game design, myst


When the Obduction kickstarter fired up in 2013, it seemed like a good moment for adventure games in general. With Unity3D well-established and the Unreal 4 engine coming up, small teams were in a good position to produce really stellar visual environments. Then Cyan got a million dollars out of nostalgic Myst fans. Good sign, right?

Sure enough, a couple of years later, I saw several Myst-inspired projects on Kickstarter and Steam Greenlight.

Of those, I have now played Haven Moon (my notes in this post) and Neptune Flux (didn't have much to say). We're still waiting on Zed and Xing. (To be sure, Xing's Kickstarter predated Obduction's -- plus one point for foresight, minus one point for taking longer. Give the point back for making progress on a KS payout way less than a million dollars.)

And I have played Obduction, and now I have played Quern: Undying Thoughts. Spoiler: those are the two good ones, so far. In fact, the great ones.

(Note: I was a Kickstarter backer on Quern. Also on Obduction and Neptune Flux.)

Just as it was impossible to talk about Obduction without comparing it to Myst, I cannot talk about Quern without comparing it to Obduction. They're both aiming at the same target: a first-person adventure in which the puzzles span every aspect of the environment. They are graphical IF in the sense that I used to talk about: you must engage with them immersively, placing yourself in the world, imagining those objects around you (and in your hands), considering what makes sense to do in that physical reality.

(Note that that "Characterizing IF" post is harsh on CYOA games. That was me writing in 2002. The field has advanced.)

Quern and Obduction are both top-notch adventure games. Both have really great, creatively constructed puzzles. They both take advantage of the 3D world engine, both visually and in their puzzle design. Both are lonely worlds; they avoid human interaction (and thus the high costs of character modeling and animation). And I finished both in roughly 15 hours of play time. So those are obvious similarities.



Now I can talk about the differences -- which are smaller, but more interesting to discuss.

Quern has lots of visual detail, but it's not so good on focus. You will frequently find a workbench full of tools, and it's not at all clear which are the important tools and which are just scenery. So many hammers! There are things in the game that I want to smash! Sorry, no hammer for you. Even more annoying, there's a loose ladder in the very first room, but you can't take it or use it to climb anywhere.

In contrast, Obduction keeps the really tempting tools out of reach. It also avoids puzzles that make you think "if I only had a hammer..." (Or garden shears, or a couple of sticks, or...)

Quern is generally in tune with Cyan's house style, but it misses a few of the details. Obduction is good about showing the difference between a two-way switch (which can be flipped back and forth) and a one-way switch (which locks after you flip it). The control might retract to show that it's locked, or you might see a pin drop into place. Quern tends not to do this. Thus, one-way switches feel arbitrary. It's particularly annoying when the effect of the switch is not directly visible; then then you have no way to experiment to figure it out.

(You might say that every control should be flippable back and forth. That's how real life works! But when designing a game, you often want to simplify. Once the power has been turned on, it stays on. Once door X is open, it stays open for the rest of the game. And so on. This is a useful trick for keeping the player out of stuck-unwinnable states.)

Obduction was built primarily around one puzzle mechanism: the seed machines. There are other sorts of puzzles (starting engines, finding passwords, using the mine cart) -- but they're very much the Lord High Everything Else. I don't mean it's 99% seed puzzles, but you wind up thinking of the puzzles as "seed machines" and "the other stuff".

Quern, in contrast, has lots of puzzle types. It's downright exuberant with them. Slider puzzles, machine puzzles, symbol-finding puzzles, symbol-matching puzzles, letter puzzles, sound puzzles, light puzzles, weight puzzles, alchemy puzzles (yay!). That's not remotely a complete list.

Moreover, Quern mostly adheres to the puzzle design rule of "do everything twice". (Once as a directly-presented puzzle; once in a new context where you have to remember that thing you did earlier in the game.) Obduction does this too, but it has fewer puzzle concepts! With Quern, by the time you're halfway through, you are balancing a mental map of everything you've encountered. Any of the mechanisms or locations could wind up being relevant again. Not to mention a mental map of the island and where every unsolved puzzle is -- because any of them might be next.

The down side of this is that, with so many puzzle types, a few are worn-out hats. There's a Mastermind game. There's a block-slider. (But not the worst block-slider, which you have to pay me $50 to solve these days. Quern's slider was okay.)

There are, as I said, a couple of audio puzzles. I did not see any accomodation for hearing-impaired players. This is not a fatal strike (not like that flippin' Donkey slider!) but you want to ask if an accomodation is possible. In some games, the puzzle is "notice the audio component at all" -- any kind of subtitling would spoil it. The audio puzzles in Quern are different; they're about noticing qualities of sounds. A non-audio indicator could work. But you'd have to think about it.

There is one terrible puzzle. I know puzzles are subjective, but at one point I said "I hit a bad puzzle" and my friend said "There is one very bad puzzle" and we were talking about the same puzzle.

I don't want to rag on that one puzzle, because the developers have said in a Steam forum thread that they're considering ways to fix it. You can read the thread for the details.

However, it's a great example of the perils of puzzle design. So I'm going to dig into it a little. I will try to avoid spoilery specifics, but I will describe some elements of the puzzle. Starting... NOW.

The puzzle has two stages. ("Do everything twice", remember?) The first stage is a straightforward information-matching puzzle. You need to look at two clues, figure out what each diagram means, combine the information, and apply the result to a device. When you push the right buttons -- you're not finished. The device ostentiatiously turns upside down.

It's clear that you have to use the device again, but with a new button sequence. You now have to interpret the clues "upside down". There are a couple of things that could mean, so you try one of them. Then you try a different one. Then you try applying those ideas to the other clue. Then you start trying combinations of interpretations...

(If you look back at the forum thread, one player mentions trying sixteen possible input sequences, based on different combinations of what "upside down" could mean. I went down the same road.)

None of this works, so eventually you give up and go to the forum. Lo, there is a thread explaining what you missed: you have to go to the other side of the island and look in a place marked by a familiar symbol. There you will find a third clue, which supersedes one of the originals. Now the second stage is solvable.

So. What is the design problem here? Missing the third clue, right? I saw players talking about ways to make the marker symbol more visible, or making it easier to extract the third clue.

But I would say that the problem is not missing the third clue; it's believing that the first two clues are sufficient! The ambiguity of the upside-down hint, while a fine puzzle element in itself, misleads players into thinking that that's the entire second stage of the puzzle -- figuring out the right interpretation of "upside down". As I noted, there are several possibilities. Each time one fails, you look harder for another. Nothing points you at the idea that you've got the wrong clues in hand.

This is, of course, why game design is hard. You have to imagine the state of not knowing -- and then keep imagining every stage of figuring out. Including the dead ends.

Also, once someone has run out of patience and looked at hints, they're not likely to appreciate your clever design any more. I went and got the third clue (nice puzzle in itself!); but then I didn't have to energy to work it back through the puzzle logic. I just looked at a walkthrough.

(PS: Video walkthroughs are still terrible, but I admit that this game would have required too many diagrams for a text file.) (PPS: Wait, someone made a text walkthrough with diagrams! See this thread. Thanks!)


I don't want to give you the wrong idea. I've spent a page and a half talking about the worst puzzle in Quern, because it makes a good case study. Quern is packed with puzzles that are much better than that. I recommend this game! You should play it. The designers should make another game.

This is a good time for adventure games.