Thursday 21 February 2013

Puzzles, what are they good for?



I recently came a across this article from AdventureGamers about puzzles, and it got me thinking. The article covers the different ways in which puzzles have been swapped for other activities over the years, something that I am very interested in. There is so much great about adventure games that just seem to be held back by their puzzles. It always seem that they break the flow of the experience. I find that many adventure games are more engaging to play when you have a walkthrough close at hand. Of course, consulting a guide has it own share of problems, and is far from an optimal way to play. Some other solution must exist.

Ever since we started Frictional Games, a big goal has been to try and fix this somehow. With each game we have incorporated new ideas in order to deliver a more streamlined experience; to try and minimize the problems that puzzles tend to cause.

When we started our Super Secret Project our initial idea was to get rid of traditional puzzles entirely. A focus from the start was to have levels where the goal was very clear. We wanted to create "scenes of drama" where the player would be free to role play without worrying about solving puzzles. But as the project has progressed, more and more traditional puzzle design have slipped in. I have been aware of this for quite a while, but the AdventureGamers article slapped me in the face with it. Despite all our efforts to the contrary, we seem unable to remove the puzzles entirely. There is just something that makes them a crucial ingredient.

The three main reasons seem to the be following:
  • Goal. They give the player a goal. When a situation is set up in the form a of a puzzle it is so much easier for the player to understand what to do next. It sets up a framework on how to behave, act and what outcomes to strive for. Actually, it is more accurate to say that setting up situation in a comprehensive manner gives you a puzzle. So the puzzle-element is simply a sort of side effect. (For those interested, here is an entire blog post dedicated to this subject).
  • Structure. It is an excellent way to set up a structural framework and provide flow. It is impossible, and story-telling wise unwanted, to allow the player to go in whichever way or do whatever they please. It is necessary to set up scenes in such a way that it confides the player to a certain path (or paths). Puzzles provide bottlenecks that are implicit and goes along with the narrative. If you want the player to visit rooms A, B and C before going to room D, you can set up a puzzles that achieves this. This system also lets the player drive the story forward. Instead of it the game telling the player when it is time to move on, the player is the ones in control. It also sets up a nice way to control the flow of the narrative. For instance, if the player is required to slow down and remain in an area for a while, you can have them searching for clues or engage in other puzzle related activities.
  • Immersion. Puzzles are a great way for the player to become part of the story. When solving a puzzle players use their knowledge of the game's world in a way that has an effect on the narrative. Players become one with the story and base their decisions on that. The puzzle is not there to test the player's wits and/or hinder progress, but to increase the sense of presence. By having something that requires the player to connect the dots often makes it much more engaging. Like how a description in a book can be more compelling if written in an indirect and/or metaphorical fashion.
I find all of these strong arguments for having puzzles. But at the same time the problems of puzzles remain. The AdventureGamers article point a few ways in which games have worked around puzzles; but the problem is that this mostly also removes what is so good about puzzles. For instance, The Walking Dead uses important dialog options to make the player part of the story. But in order to this, the game needs to have long cut scenes and reduce its scope of interaction. Players no longer push the story forward or get implicit goals. The game simply tells them what to do and when it is time to move on. For all its accomplishments, The Walking Dead fail to deliver a game where you play all the way through. This is not the kind of experiences we want to make at Frictional Games.

Instead of thinking about what to replace puzzles with, it is more rewarding to consider how to evolve them. How to improve them in a way that keeps the good traits and removes the bad. The first step towards this is to consider why we have puzzles at all. I think a major reason many adventure games gets problems with puzzles is because they are never justified. Every puzzle is seen as a "fun challenge", a feature with intrinsic value that should not be questioned.  I think that simply asking the question: "how does this puzzle serve the overall experience" is bound to be a good start.

Once it has been decided that a puzzle is really needed, the next question is what kind of complexity it should have. If you want a game that is about engaging the player in a narrative, you really want the puzzle to be as simple as possible without losing any of the benefits  So what is simple enough? My current gold standard is:

"A puzzle should make players to do something in such a way that they feel they came up it themselves."

This means that the puzzle must give the player some kind of "revelation" and must not feel spoon fed. The path from encountering the obstacle to performing the solution should not be too obvious or simple. However, this often means puzzles become too complex and/or difficult. The solving problems then devolves into "guess the designer" which ruins the intended effect. The player should be kept inside the game's world and never be forced to think outside of that. What follows are some of the ways we try and solve this:
  • Locality. All ingredients for solving a puzzle should be in close proximity to one another. This makes sure the player does not get stuck because of missing a clue or an item at a now distant location.
  • Multiple Solutions. Having many ways to solve a puzzle is often used as a replayability feature. In our games, it is instead used to make sure that the solution feels natural and intuitive to a wide range of players. In many cases we have actually implemented whatever fitting approaches that testers have tried (to the point of even allowing button mashing as a way to progress). 
  • Low Item Density. By making sure there are not too many locations, objects, characters, etc, one can avoid confusing the player and leading them on stray paths. Too few items can also be a problem of course, so one has o have a bit finesse.
  • Coherent Simulation. This means that mechanics work globally and are consistent throughout the game. For instance, a pickax is able to break any object made of ice. Most of the recent great puzzle games like Braid and World of Goo use this approach; however all these games are set in fantastic realms where the mechanics come before the story. In a narrative driven game aimed to have a sense of "reality", it is much harder to be 100% consistent. We have tried it with physics and it comes with all sort of trouble. More info here.
  • See it as an Activity. When possible it is often rewarding to think of puzzles as an activity. This push you out of mindset of just thinking about having clever solutions. If you want to have puzzles that are there to enhance our storytelling, they need to stop being seen as challenges. 
  • Part of the World. The most obvious, and also hardest one: puzzles should always stay consistent with the story. If not, it will be painfully obvious when one is encountered. Resident Evil is a poster child of this; very few of its puzzles make sense in the game's world.
  • Story Coherent Hints. I think the best way to make sure that the player is not stuck is to have protagonist comments, notes, or whatever auxiliary means, show the puzzle from different angles. This in order to make sure that the player has not misunderstood some concept and is seeing the puzzle in the "right way". If players get stuck, the most common cause is that there is some step in the logic that they failed to catch. By having subtle hints it is possible to minimize this from happening
The above tips are meant to facility a smoother experience for the player while trying to solve the puzzle. Another important issue is how to make it clear that there is a goal at all. Player often get stuck in games because they do not realize what their objective is, what puzzle it is that they are supposed to solve.  Here are three ways that can help overcome this problem:
  • A Clear Goal. This is probably what we have tried to use in most of our games. It basically means that you make sure players know where to go next. In Amnesia we always tried to have some obvious obstacle or let some kind of note/vision give a hint. As a back-up we also employed a somewhat immersion consistent todo-list, where further hints where given. 
  • Hidden, but guided. Sometimes it is possible to never tell the player exactly what to do, but guide and/or confine them in such a way that they will stumble upon it eventually. A simple, but effective, example is in Silent Hill 2 where you need to escape a well by finding a loose rock. It is a great way to create a sense of panic, and since the solution is so easily found it never becomes frustrating.
  • Spelled out Solution. This is when you just tell the player front up exactly what they are supposed to be doing. This might seem kind of of boring, but can work really well in some situations. A perfect example is the food rationing in The Walking Dead. Here it is obviously clear what you need to be doing, but a quite hard to decide who to give food.
Despite following all these rules, it is not sure that you come up with a puzzles. It is vital to not see them as stumbling blocks along the players's journey. You want something that enhances the player's time in the game's virtual world. Not something that reduce it.

A very bad example of this is in the remake of Broken Sword. When encountering a locked door, a sliding puzzle pops suddenly pops up. Disregarding that I loath sliding puzzles, this is really bad. It has nothing to do with the game's narrative. I gain nothing in terms of a connection with the story by solving this. It is simply there to hinder my path. What makes it worse is that the obstacle itself, a locked door, is not really interesting. The designer has taken an uninspiring set up and made it worse. This is a bad usage of puzzles.

A good example is found in Call of Cthulhu: Dark Corners of the Earth, where you need to open a passage to a priest's secret hideout. Upon arriving at the church you are not aware of there even being a hide-out and need to read this in a note. Here is also a clue on how to access it; the church bell must be rung in a certain order and opening up a secret passage way downstairs. I think this sort of puzzle is great; it requires a combination of lore, exploration and force the player to make narrative connections. It also lets you interact with the environment in an interesting fashion. By both discovering and opening the secret passageway the player has an active role in the progression of the story. 

There is a catch though. The mechanism for opening the hideout makes little sense. The whole town would notice whenever the priest wants to go to his lair. But in the end, it does not matter. It satisfy enough criteria to still be a good puzzle. This is a really important aspect of the craft.

Coming up with puzzles is hard. Coming up with puzzles that are coherent, engaging and fit with the flow of the narrative is extremely hard. If you want to make an engaging and varied adventure, it is impossible to make every puzzle perfect. Above all else, the puzzle must fit with the experience that you want to create. Players can see past strange mechanics (like the above bell puzzle), live with simplified inventory system (like in The Walking Dead) and other sub-optimal solutions as long as it serves to enhance the experience. This is very important to remember when creating a puzzle. 

The goal is not to make players think you are clever or to do the most complex set up. The goal is to make sure all parts serve the experience as a whole. It is very easy to forget this (I have done so many times myself) and it does not help that puzzles are fiendishly hard to evaluate. But I think that with the right mindset, it should not be an insurmountable challenge.

As mentioned in the start, over the years puzzles have been pushed aside for other mechanics. Games with more progressive design either push the puzzle elements into the background (eg Uncharted) or base all around a specific mechanic (eg Portal). I do not think it is time to give up on the more classical adventure game puzzles yet though. By this I do not mean that we should go back to the interconnected puzzle design of old days (as explained here). Instead we should try and look at puzzles in a different light and see how we might change them and reinterpret their role. This post has been an attempt to do just that, but I think there is a lot more to explore. It would be very bad to abandon the quest to combine storytelling and puzzles just yet.

Those interested in more puzzle discussion might want to take a look at series of articles on puzzle that I wrote a few years back while working on Amnesia. They can be found here. The posts go through some other aspects of puzzles design that should be of interest..

I am also interested in getting your input and/or links to other articles on this subject. It is not easy to come by good writing on puzzles, and even harder to find something that discuss narrative-serving puzzles, so I am very grateful for any feedback and tips!