Adventure Gamers - Forums
Puzzle design in gaming
Hello! I’m pretty new here, so I hope this is the right place to put this post, as I don’t want to limit our discussion to only adventure games.
Anyways I just wanted to talk about notable puzzles that are well designed, or in general what makes a good versus bad puzzle design, and what kind of impact that has on a game (be it video game or even tabletop game). I’ve done a little research into people that are puzzle designers, (I’m jealous of that job, like the people that work in professor layton games. Wow!) their process etc.
(I searched the forum but only found threads on puzzle difficulty.)
My final goal is to use what I learn about puzzles in a tabletop RPG, which presents it’s own challenges as the Player and the character are kind of separate people, but that’s a classy problem for later.
I tend to prefer a nice mix of inventory object puzzles, dialog puzzles, and logic puzzles.
I’ve actually not played a ton of adventure games, and beaten even fewer, so I’m hoping that you all can point me in the direction of well-executed puzzles that have that satisfying feeling oh “oh look how clever I am.”
So I divide puzzles up into these categories:
Logic puzzles; where the player has to use their brain, some guess-work, good ol logic, etc. The Character in the game doesn’t really have to do anything. A good example of that would be the direct cipher letters in broken sword. The player does all the work substituting letters in trial by error until it starts to come together. Ideally, these provide a starting point, a goal, and a clear way to tease out a solution to arrive at your goal through whatever means a parameters the puzzle provides.
Inventory-object puzzles: You’re required to have picked up the thing, and bring to the (probably incongruous) place to use it in a probably novel fashion. In Adventure games this isn’t that bad because the majority of inventories are like puzzle objects, and then junk usually. The problem becomes when a game has a mix of these items, and utility items, and does a poor job of separating them. I can’t tell you how many RPG’s I’ve played where you receive a seemingly useless object, only for it be used once in a very specific place. In RPGS that wouldn’t be so bad except potions, elixir etc all have utility. Have you ever played Earthbound? It stores all key items and utility items in the same inventory (it’s maddening really), where as in a game like final fantasy, they are usually put in a “key items” inventory (and never really require you to pick the item and “use” it anyways). Some inventory puzzles are pretty coherent, and some are cute quirky, or creative, and that’s all good. See it falls apart for me when I start playing something like The Longest Journey. The way the inventory is expected to be used together to solve puzzles make NO sense! I was a big fan of Funcom games, so my heart was broken when I just could not figure out how to pick this thing up off of the train rail, in which you have to get like an inflatable duck, a ring, clamps, and something else and combine them to retrieve something else
which after trying everything with everything else, I just gave up, and looked up a walkthrough which really ruins adventure games for me. The problem is that this logic is never consistent. In the same game (before stark or arcadia, or whatever the fantasy world is called) they recycle one of those reflect the laser beam with a mirror puzzle. The logic there is very straightforward and a little cliche. I’ve never included this type of puzzle in a table top game, but I expect what I will do is allow characters to have a personal inventory for utility written on character sheets, and then a party inventory where puzzle items are kept distinct. (I haven’t decided yet, there might not be any utility items.)
The last one I would call skill check puzzles, and maybe they don’t count at all, but it’s when you can learn information or proceed because of the character’s skill, not the player’s. Your character has a hacking skill, so he gets a hint on the password to open the door, or he can just open it. Other character’s lore skill reminds them of a 2nd entrance long abandoned (and so on). It’s kind of the polar opposite of the first puzzle type where it is completely dependent on the player and not the character. (think QFG series)
I think as mentioned on Josh Brycer’s blog, it’s very important to have a clearly defined goal with a clear understanding of whats preventing my advancement in the game. And then the means to which that is resolved can be presented in varying levels of clarity depending on the puzzle. Ideally, the puzzle pieces you have to work with and the logic of how to arrive at the goal is clear as well, but I’ve solved a lot of puzzles in my day just by playing with numbers, or sliding tiles or whatever.
Have you ever noticed some games have included hint systems? In games like broken sword it’s really helpful, but I get stressed out about getting a bad score, and won’t use it. In games like Theresia, Dear Emile, or even 9 hours 9 persons 9 doors, there are built in hint systems that I love because in those games they are usually context sensitive. In Theresia the protagonist I think flat out says “There’s nothing else in this room” If you’ve activated the hint system in a completed room or will mention the thing they picked up from the room, and how it could be used. This is really useful for when you come back to a game after a long time. 999 does a thing a like with it’s hint system (mix of inventory object and logic puzzles) where the first hint is a gentle nudge in the right direction, the second hint gives you a start on the puzzle, and the third hint goes on to say more explicitly what you should do or how the pieces fit together. I like this hint system because I often need a nudge to follow the developer’s design language, but I don’t want a ton of help.
Occuluncus: No, no. The eye stays closed!
Many taxonomies of puzzles depend on how the player interacts with them. For example: inventory puzzles involve using items, dialogue puzzles require the player to navigate a dialogue tree, etc.
This is fine, but tells us little about what makes a puzzle fun to solve. Let’s turn the problem around and focus on solving.
A vital question: Should the puzzle be solved by gradual tinkering, or by a sudden leap of insight?
Some puzzles encourage gradual progress. At best, they also give nudges and feedback.
** A dialogue tree puzzle may be solved by trying options; often one choice will lead to another branch, suggesting that the player should keep following that path.
** Some mechanical Myst-style puzzles encourage you to try poking and pulling and moving parts around until you get a ‘feel’ for the puzzle.
** An inventory puzzle may involve assembling a compound item piece-by-piece, trying parts until you get somewhere, or may involve a challenge with stages of success, such as distracting 3 guards one-by-one).
** Many “puzzle games” (as opposed to adventure games) focus on systems to tinker with over time. Even Tetris can be considered a timed multi-step tetromino fitting puzzle.
What these examples have in common is that the player can gradually chip away at the problem.
Advantages of Experimentation-Driven Puzzles:
** If a puzzle is rewarding to interact with, and not simply a ‘success/failure’ gate, the player can focus less on asking ‘how do I solve this?’ and more on asking ‘What will X do?’.
** A player can get a real sense of progress. It’s often not a matter of having a brilliant idea, but rather of paying attention and being methodical and thorough.
** Watching a player solve such puzzles during testing is more informative! It’s easier to tweak feedback on a gradual puzzle than to give a fair, yet subtle, clue to a leap of insight.
Disadvantages of Experimentation-Driven Puzzles:
** There is less opportunity for an ‘aha’ moment - no sense of a sudden blinding insight. Solving one of these feels like a “job well done.”
** When the ‘progress steps’ are themselves vague, what the designer really is demanding is multiple sudden leaps of insight! (Or psychic powers.)
** In the absence of good feedback and a system worth experimenting with, these puzzles encourage brute-force thinking, which is tedious. Is it satisfying to solve a puzzle by poking a machine until the switches happen to line up? Or to try every dialogue option until you say exactly the combination a designer had in mind?
In contrast, some puzzles leave little room for experimentation. The steps to solving might be described as:
* 1: Think about what you know and what you’re trying to do.
* 2: Do something clever.
** A puzzle in one mystery game required the player to show the correct combination of four pieces of evidence from a long list, with no feedback on wrong answers. Guessing and checking was out. You had to prove you understood the central ‘trick’ of the mystery!
** Finding the correct combination of item-and-target in many inventory puzzles depends more on insight than random clicking. Or at least, it SHOULD.
** A game may shatter a prior expectation. The player must stop haphazardly applying the rules (or even input methods) already taught and think. Naming examples here would be a spoiler, but I’m sure most adventure gamers can think of at least one case!
**** One ingenious text adventure requires the player-character to say one word at a key point. (Not a magic word, exactly…) There is no prompt for this. The player must realize it is time to ‘SAY X’, and why.
That’s not to say that ‘leap of insight’ puzzles HAVE to be hard. There’s one at the end of Portal 2 that could be brute-forced, but most players make the wonderful leap. Zork: Grand Inquisitor has several “spot the easy answer hiding behind the complicated system” puzzles as well.
Advantages of Leap of Insight Puzzles:
** Done well, they create an ‘aha’ moment in the player’s mind that is one of the greatest pleasures of playing a good adventure game.
** When designed to prevent brute force attacks from working, they add to the challenge of a game.
** The insight itself can serve as a knowledge test - does the player understand what is going on? In mystery games, it is important to keep the player from blundering too far without having some real understanding of the plot, or confusion will reign.
** At best, the insight requires the player to understand what they would do in the character’s position. So these puzzles can be a powerful way to build identification with the protagonist.
Disadvantages of Leap of Insight Puzzles:
** One person’s leap of insight is another person’s moon logic. Test well, and make sure the puzzle makes sense in hindsight and that fair cluing of SOME kind exists.
** It is rewarding to solve one of these, but thanks to the lack of feedback, it is quite frustrating to be flat-out stuck on one.
Neither type of puzzle is inherently better. There are days when I want to tinker and get feedback, and days when I want to say “AHA!”
If I were advising a new designer, I’d suggest to focus on the kind of puzzle that gives lots of feedback and is solved gradually, because it’s an easier target to hit by iteration. If you’re writing a tabletop RPG campaign, be sure to allow for multiple approaches and to reward surprising answers. In a computer adventure game, multiple solutions may be too hard to develop.
Either way, test well!
OKay I think follow in terms of how these two kinds of puzzles are approached and solved differently.
Do you have any advice on how designing them differs?
I have a love-hate relationship with these AHA puzzles. I wish I had a better description of what makes them work, or successful or fun.
Seriously, sometimes I don’t get a new hint or anything, I just like…. It just comes to me after a while. Or most often I think of it when I’m explaining to someone else why I’m stuck.
Recently, I’ve thought of a totally different medium that I might want to use these types of puzzles in.
I started this talking about including tabletop games which is fine, but I don’t really want a group of people sitting around watching us not make any progress.
Let me pivot this by saying that I recently was talking with a friend about doing a mail-based adventure style puzzle that combines like choose your own adventure books with educational parts and puzzles. The idea is that the whole family can be involved. The kids learn stuff, and it includes some art projects, and puzzles that younger or older family members can participate in.
So in that idea, you’re right saying that there’s no feedback knowing why a solution is wrong. But also, in this like play-by-mail type of activity, obviously it takes the immediacy expectation out of the equation, but it gives the player long amounts of time to think about the puzzles, their solutions and the context.
Occuluncus: No, no. The eye stays closed!
Sorry for my delay in getting back - I’m not on here that often!
Do you have any advice on how designing them differs?
Yes. I think there are several differences!
In designing “Aha!” puzzles, the key things are:
* To have a sudden inspiration or insight
* To guide the player to have that sudden inspiration or insight subtly.
* To “stick the landing” - after the insight occurs, the player should be able to make sense of it in hindsight.
Step 1: Coming up with the ‘Aha’.
There is no recipe for inspiration, because if there were, it would not be inspiration. HOWEVER, there are many ways to dramatically increase the possibility of inspiration.
* In Wonderbook, Rosie Weinberg describes an exercise inspired by Guilford’s Alternative Uses Task. You look for household items, such as a paperclip or a magazine, and try to come up with as many nonstandard uses for an item as possible in one minute. You repeat this several times with different items. “Paperclip as lockpick” is trite, but what about paperclip as earring, or paperclip bent into a shape and dipped in ink to make a pattern as a stamp?
* Think through the implications of your setting and characters carefully, asking “What Ifs?” In particular, watch out for cases where you make an assumption that doesn’t hold up on deeper inspection. A misconception that blindsides you can also surprise a player.
Aside: One of my favorite puzzles I used in an RPG came about this way. While planning a mystery that involved time-travel, I was intended to include a standard question about how many times a gun was fired. On reviewing, I realized that I has miscounted the number of firings because I had not taken a ballistics test firing into account - and that this was a huge blind spot that I could turn into an “Aha” moment, which worked out well in play.
* Read something unrelated to your project, then try to relate it to your project. This can encourage lateral thinking.
But I’m not going to sugar-coat this - coming up with inspiration is HARD. I have written many, many puzzles, but I can count the number of really good, genuinely insight-provoking AHA puzzles I’ve come up with on two hands at most, and each time they blindsided me in some way.
2: Guide the player to the same insight.
The biggest danger here is that inspiration can be very individual, and you want this puzzle to be solvable by more people than just you. So you will need to:
*** Make sure that alternative solutions are either recognized or plausibly eliminated. If the player character is going to need to retrieve an object from the top of a tree by chopping the tree down, you need to justify why they can’t just go to a store and buy a ladder, or throw something, or do any other of other reasonable approaches.
*** The more unconventional the approach is, the more explicit the cluing needs to be. In the opening of Infocom’s Trinity, there is a lawn of angry grass stalks that stop you from crossing. The solution is to get in a baby carriage and open an umbrella to let the wind carry you across. To make this even remotely fair, the game repeatedly calls attention to the wind, introduces the umbrella by having it be blown away, and, CRITICALLY, shows that baby carriages rolling across the lawn are not attacked.
Players often discover answers by imitating something else.
*** Make sure the rules of your universe are clear and expectations are consistent. A cartoony solution works in a cartoon universe, but not in Police Quest: Open Season, where the “gritty” earlier gameplay turns into a series of contrived, awkward, game-y puzzles for the endgame.
Testing is generally necessary here - getting a test player to talk through their thinking will help a lot!
3: Sticking the landing
When the solution is found, the reward should be gratifying and the game should make sure the player’s insight is confirmed. Something about the result of the insight should confirm that it was more than a lucky guess! Even better, it would help to re-use that same idea later on in a more complicated puzzle.
I’ll discuss the other kind of puzzle (experimentation-based) later, but I hope this is a helpful start!
Continuing: Experimentation-Based Puzzle design, and how it differs.
The biggest design difference between an experimentation-based puzzle and an insight-based puzzle is that when designing an insight-based puzzle, you generally begin with the realization you want the player to have, and work backwards. In contrast, an experimentation-based puzzle design begins with the type of interaction you want the player to have. In other words, instead of asking, “What do I want the player to realize?” first, you ask “What do I want the player to do?”
An experimentation puzzle lives and dies on how entertaining the process of experimenting is, not how stunning the final step is. For example, a crossword can be seen as a gradual experimentation puzzle built of many small realizations, with each answer giving a clue to the next.
1: Define a mode of interaction and/or a series of system rules.
2: Decide how the player will get feedback of success, failure, and hinting.
3: Experiment with the rules until you get something sufficiently enjoyable, with heavy playtesting.
1: Finding interactions and systems that build up into puzzles
A few kinds of interactions and ways of ‘poking’ at a system to consider:
** Alter position or location. For example, set up a grid, and say that elements on the grid can move and affect each other in specific ways. This describes everything from chess puzzles to Sokoban. If rules then depend on relative position (wires must be connected, for example), and limitations are set, the player can fiddle around with pieces and put them together.
** Change attributes of system elements, such as shrinking, growth, reversal, speed, changing materials, reversing gravity, vulnerabilities to weapons, etc. In a mail-based game, this can even be “changing letters of words” to make new words.
** Trigger responses based on psychology. For example, a system where players interrogate someone and need to get on their good side or get them to reveal information.
Notice that in this case, rather than having a single insight, you instead think, “What might happen if the player does X?” And then you give the player reasons to ask the same question, and try it!
Since you’re interested in play-by-mail, let’s focus there. It really IS possible to give instantaneous feedback without being present. A few ways:
* Progressive correctness checks at intermediate points. One kind of puzzle that is especially suitable for children is a folding puzzle. The parents print a sheet that has lines and symbols on it, and the child follows instructions to cut along lines and fold until they get a certain result. In the end, the folded design makes some cool origami art, and also several symbols are placed in a row to create a code. This is the final answer. Along the way, either images can be shown of what the project should look like (easier), or instructions can provide feedback along the lines of “At this point, the star should be next to the circle. If it is not, you have made a mistake.”
Another example of an interactive correctness check that can be done remotely is by asking a question where a number of blanks are given for an intermediate answer. If the answer should be a 5-letter word, and the player guesses a six-letter word, they know they need to rethink that step.
The MIT Puzzle Hunt uses this sort of system - each puzzle produces a keyword, and if all puzzles are solved correctly, the keywords fit a pattern. If one of them does not, you know that you made a mistake there!
* Progressive cluing when people fall behind. Suppose you were writing an interactive story where a password needed to be given. You could give the instruction: “Add 67 to the number of letters in your answer and go to that page. If that page has a star at the top, keep reading. If it does NOT, you have made an incorrect answer. Turn to page 66 instead in that case.” And then page 66 would offer a hint on where to get the password.
A variant would be to send the players password-protected files, or links to password-locked websites. Incorrect entries could yield clues.
* That said, the most elegant way to do this is to build feedback into the puzzle design itself. Suppose, for example, you give players a page to print with square tiles to cut out. Each tile shows wire segments that can be rotated and placed on another grid. The players must use the segments provided to connect two nodes on the grid. In this case, it is obvious when they are making a mistake - the pieces don’t connect properly!
(A variation on the above would be to have there be exactly one solution, which covers up some specific squares on the grid, and then to have the players read the letters on those squares from the beginning of the pipe to the end to spell a message. Then the message could be used as a password, and would provide yet another source of feedback and reward!)
3: Iterative testing
With this kind of design, it is absolutely vital to test and retest what you’ve made, and make improvements. In the “insight” puzzle, the question is: did you clue well? Here, the question is, are the steps of solving fun, and have you made sure the puzzle is both solvable and has no extraneous solutions you are unprepared for?
I hope this helps!