Wednesday 26 August 2009

Puzzles in horror games. Part 3

In the previous blog some problems with puzzles in adventure games where discussed. It was also mentioned that a major culprit in all of these where that many adventure games do not have a coherent system for doing interactions, like for example a Super Mario game. A nice way of solving this and still allowing a varied set of possible actions might seem to be physics. For the last couple of years, especially after the launch of Half-Life 2, physics have gotten a lot of attention and is pretty much a standard component of any 3D game today. It might therefore seem obvious to start using physics in adventure games in order to bring the genre forward. However, having implemented physics for four games now and working on a fifth a lot of problems have emerged. While adventure games and physics seems like the promised land at first glance, it is far from it.

Difficult controls

Classic adventure games like Myst and Broken Sword got very inntuitive and simple controls and can still give the player a very detailed and varied experience. It is just matter of moving the mouse and clicking to control games, and it is hard to make it more simple than that. For a game using physics, it is much harder. Doing something similar to Broken Sword with physics would be very hard and I have yet to hear an ideas on how it would work. A first person perspective is pretty much demanded in order to have a control system that lets the player perform enough actions to solve varied types of puzzles. This leaves games like Myst and Tex Murphy to use physics. The first person system is also what is used by Penumbra.

However, controls quickly get messy because of something called the z-axis or the third dimension. The problem comes from the main input device, the mouse, working in only two dimension and the actions on screen taking place in three. When manipulating objects this can make things a bit difficult, and in Black Plauge we added the scroll wheel to make up for this. Penumbra was also implemented on a haptic device which allows the user full 3d movement, adding a very nice way of interacting. However, given the current spread of haptic devices, it is not viable to design a system around them. As for other gadgets like the wii-mote, it only has acceleration in 3D, postioning is still in 2D and thus it does not solve the problem.

The problems with controls do not stop there though and the user also wants to rotate objects and do other kinds of manipulation. We added rotation in Black Plauge too and by now the full control system was pretty complex. All of this complexity still only allowed very basic interaction though and attempts to further enhance the interaction could easily lead to something like Trespasser which featured increadibly difficult controls.

The Chaos effect

Even though physics are controlled by a very limited set of rules, because of its complexity, even the smallest varition can cause major differences in outcome. The most striking example of this is the mining cart puzzle in Penumbra Overture. Here the player is supposed to push a cart down a slope to hit a wall, but for some reason, every now and then it would derail and miss the intended target. To fix this, extra forces where added keeping the cart in place and numerous hours was spent at getting it stable. After all this work, the cart can still derail though! In hindsight it would have been a lot easier to just use an animation. This shows how such a simple event can cause tons of problem when physics is involved.

Physics breakdown
Not only does a physical simulation suffer from undeterministic behaviour (chaos) , it can also fail. As hinted in the word "simulation", game physics is not a perfect replica of the real world and can break down at certain points. In the Penumbra games, the best example is that the player can bang objects through the floor if using enough force. Some objects are easier to bang through (because of shape, size, etc) and in Requiem we had make an important object magically appear if it went through.

Sequence breaking

In adventure games, the designer can usually set exactly the kinds of interactions possible with an object and have full control of all possible outcomes. For physics, it is impossible to anticipate all that can happen and one can only test as much as possible, hoping all gaps have been closed. For example: a pit that should not be possible to cross at certain point in the game, might be possible to cross with some ingenius use of objects. In order make sure certain things does not happen we had to add alot of extra checks and hacks in the Penumbra games, sometimes even excluding the player from doing certain actions. This can easily break the sense of immersion, but might be a must in order to have a stable game.

Another type of sequence breaking is doing something "stupid". An example from Penumbra Black Plauge is in a machine room where the player needs a steel pipe as leverage to break open a door. However, there is also a hole in this level and it is possible for the player to throw the pipe down in it. To remedy this, we added several pipes in the level and if the player where to throw them all down the hole, a pipe would magically appear inserted into the door that needed to be opened. It is not very immersive, but at least the game did not break. Luckily, most player seem to not trow imporant objects down in holes.

Player getting stuck
All of the problem listed above can lead to the player getting into a unwinnable situation. Normally physics game have a reset option that lets the player start over. This is crucial for most physics game, for example consider how World of Goo would be if it the player should never be able to end up in an unwinnable state. It would pretty much destroy the entire game.

Adventure games do not have this luxary though and unless one counts some really old games, it has always been possible for the player to continue unless death occured. This restriction ends up making many physics puzzles impossible to implement in an adventure game and when designing the Penumbra games, many puzzles had to be abandoned due to the high risk of putting the player in an unwinnable state.

End notes
I turns out that using physics as a way to bypass the rescritive actions and incoherence of normal adventure game interaction is not possible. Instead adventure games with physics require alot of restrictions and special cases to be added in order to make a working game. It also limits the amount of puzzles that can be used in the game and is not as groundbreaking as it first might seem. This can also be seen in games like Half-life 2 that has not only very few physics puzzles, but also heavily restricted ones. Some physics-like puzzles (for example a large bridge acting as a see-saw) does not even use physics!

Still, given all of the problems with physics, it should not by discarded. It allows for great immersion, which is especially imporant for horror games, and correctly implemented it can make puzzles seem more intuitive and fun to perform. It can also makes it easier to do puzzles with many solutions and make puzzle encounters less frustrating. We will still be using physics for our upcoming game and would like to see more adventure games trying it out!

What is your opion about physics in adventure games? Know any other games with good physics puzzles?


  1. This was certainly an interesting look at physics in adventure games. You brought up a few good points, but also brought up some non-issues in my mind.

    For example, if somebody manages to create some McGyver-esque bridge to some other part of the level, in most cases I don't see that as a game breaking issue. Sure, that person may miss a part of the game, but imagine how they would feel knowing that they were able to do something totally original and have in succeed. Unless this would result in game-crashing scenarios, I'd say don't bother spending too much time on this. Besides, probably less than 1% of the players would even notice this.

    It is true though that untill computers become a lot better, physics simulations will be buggy and produce strange reactions, especially with real time applications- I think this is why not many games use physics as a gameplay element. This makes it so that physics puzzles have to be fairly simple, often resulting in contrived scenarios that wouldn't exist in real life (ala Half Life 2). Puzzles in the Penumbra series were generally pretty good about this, but there will always be limitations on what you can and can't do.

  2. Regarding "McGyver-esque bridge":
    I think this is a very difficult issue.

    First of all as a designer I do not really like the fact that people may skip large chunks of the game. Perhaps some important story developments are missed that makes some other twists meaningless or not understandable. Even if it is just 1 in 100 this happens to, I would consider it too much.

    Secondly, it might be hard to predict what kinda of issues might pop up if the player manages to get past a certain "impassable" obstacle. One would have to carefully test this way of completing the game too, but this can get pretty complex if more similar situations arise.

    I guess it also depends on the game a bit. In some games the mechanics might be more built for the player mess with the world. In Penumbra I feel that it is a pretty linear experience, I would not like to allow this kind of sequence breaking.

  3. Hmm... you might have a point. Thinking it through a bit more, that might be a bigger problem than I thought originally, especially for a linear game.

    I thought of a pretty easy fix for it though. All you have to do is block off certain passages with an invisible wall, and then somewhere along the line, a trigger tells the "wall" to dissapere. That way, even if people do manage to beat the system, they can't go anywhere and have to contenue with the game as planned. This would be a bit of an immersion breaker, but would only affect ~1% of the players, rather than all of them.

    Plus, it seems like it would be easier to code than trying to think up every possible way the player could use an object and trying to stop them from doing something you don't want them to.

  4. Good point! Although, setting up invisible walls is a form of "feature-block", just like not letting the player do some other action. Also, one needs to be careful when setting up blocks, so that the player understands that something was not done correctly. Do not want the player to get stuck trying to get past something impossible.

    The whole gambling with these kinda things is interesting too. If there is a chance of 1% that something not so good will happen, then considering there are about 10 really large review sites it is a 10% risk that an important reviewer will encounter it and perhaps lower the score because of it? From a commercial standpoint this is really important. What kinda odds does one feel comfortable with?

  5. Another possibility is for example: when the player has crossed a pit which wasn't meant to be crossed. To use events. To prevent the player from leaving the "game area", let the floor collapse, or let the player encounter a collapsed roof or wall, so he cannot pass. That makes the game more realistic. In half life 2 i found a way to get on a roof. The result was i fell through the roof, through the whole level. I just hate these things. And there are much, MUCH games where happened the same thing to me.

    Sorry for my english!

    Julien, from belgium

  6. There are already points in Penumbra when the player character simply does not want to go back. In a similar fashion, one could add such an invisible wall the PC simple does not feel comfortable in crossing. Something like "I have a strong feeling that i should not go any further here for now". Maybe even adding something like pain, so if the player continues he will eventually even die. This wont break immersion that much.

    Nevertheless I strongly suggest adding an "unstuck" option to your future games, as I managed to get stuck in Overture with no way out (had to reload a save-game). Things like this will always happen because as a developer you can not check for all possibilities and everyone understands this, i think.

    Beside that, I think you are really doing great with the physics approach. This makes the adventure more interesting and more immerse as interaction appears more natural.

  7. Some things that caught my attention in penumbra.

    - almost no wide, open areas. Always tunnels and caves.
    - not much music in game.
    - no humans to talk to. (besides Red)
    - only dogs, spiders, infected, and the giant worm as enemies.

  8. Again the "McGyver-esque bridge":
    I think preventing players from such crazy stuff might be frustrating as well. The two situations I can imagine are:

    1. The "bridge" leads to further parts of the level the player must enter anyway - so the bridge is a sort of shortcut for skipping puzzles.
    In my opinion this should be only prevented if the player misses some items/hints/mechanics that are needed for *subsequent* puzzles after the gap (because he might not be able to use his construction for getting back again).
    If however he misses nothing in that kind then I would not prevent it. Why? Well, in case of such bridges the user puts a lot of effort in creating weird and funny effects and invests quite some time for searching crates, barrels, for putting inventory items at the right places etc. It is then clearly no accident but a deliberate decision to do this. After all, if the player does not want to go the obvious way I think he is fully aware of skipping some other things - and last but not least no one hinders him to play this passage a second time in the obvious way as intended by the designer...

    2. The "bridge" leads to normally unaccessible parts of the game. I remember such things from Thief1 where some crazy people stacked *all* creates they could find in the city streets to climb all sorts of buildings which were only meant to be decoration. Again in every such case it is clearly visible to the player that this was not intended and may lead to strange behaviour. I think both the developers and the community considered such things being rather funny than annyoing. Because everyone knew that it was a kind of deliberately sabotaging the game for doing crazy things.

    In short: in most cases it is the users descision to fool around including quite hard work, why hinder it - he just should not complain afterwards :)

    Btw, in Thief2 they reacted in an interesting way. In this game you were able to throw your mechanical eye and then looking though it. Of course it was abused again by throwing it at the most impossible locations. So Looking Glass put a dancing zombie behind a wall that you could never cross with your player character but throwing the eye over it revealed this funny easteregg.

  9. I remember playing Black Plague, the part where you had to get over a big pitt in the ventilation system. While the real and most obvious solution would be to simply put the piece of wood over the pitt, making a bridge, i didnt think the wood would be long enough so i decided to do something else.
    I went back and forth and used the rubble from the prevoius room to fill up the pitt until i could walk right over.
    I really like puzzles where you have to do alot of physics interaction - building, smashing etc. They count as "fun" puzzles I guess, but they rock!

  10. I forgot to mention barricading, I love barricading. Sadly there wasnt much of that in Black Plague. The problem with barricading is usually that it serves no goal - the room only has one exit. But giving you enough time to escape or hide or putting a puzzle together is cool!

  11. Regarding the McGuyver Bridge, the argument "I do not really like the fact that people may skip large chunks of the game" holds zero water. I understand, from an artistic and personal perspective, that it means a lot for people to experience the entirety of the game. Their experience should not be yours to dictate. It obviously is to a great degree (as a designer), and i am simplifying a complicated issue, but people should be encouraged to play games to have their own experience, not a predetermined one.

    Artists create their art and leave the interpretation of their work to the individual. The ambiguity of the original 'meaning' of their work is part of what makes it engaging art. An artist must be satisfied themselves with their work, and accept the possibility that no one will see it as they do.

    Erecting invisible walls and other such contrived things is frustrating for a player and produces a feeling of powerlessness. It reminds players that they are playing in someone else's world. I believe that making the walls visible and giving an in-game rationale for such walls, such as memory blocks in assassin's creed, would be a softer breaking of the fourth wall.

  12. I understand, from a practical standpoint, there must be SOME limits. For the sake of the plot some concessions (probably) must be made, like the reappearing steel pipes. My point is that is that we must be careful that the concessions we make are for the game and the player, and not for ourselves

    I apologize for my tone - i've been thinking about this a lot recently and i am expressing my frustration at a gaming convention (which stems from my frustration at not being able to emulate the freedom of real life), not at anyone in particular. I'll get to work right now on a 100% foolproof solution :)

  13. Ian:
    The problem depends of course on how you design the plot. If it is a very linear one, then it will be experience destroying if the player manages to break away from the set out path. If you are interested, I talk about this a bit in this post:

    The things you talk about is a very clear clash between plot (narrative in the post) and gameplay. To be able to support stuff like the "mcguyver bridge", the design of the plot needs to be different from what is normally found in games.

    That said, this is not always a bad thing. Some experiences require linear plot elements and then you have to make sure the user does not do too many crazy things :) Unfortunately, compromises are almost always needed.

  14. A lot of what I have to say on this topic can be summarized by saying "Do more of the same, and then some, with Amnesia."

    I've got to say that the Penumbra series is probably the most important thing to happen to gaming since The 7th Guest. A large amount of that has to do with the implementations of physics as well as the spatial awareness found in the Penumbra games. While other games have made attempts in these areas, I feel they were quite mastered by the Penumbra series.

    On the topic of playing to reality, I believe fully in allowing a player to become hopelessly stuck, as long as they are made aware of that possibility ahead of time and as long as there is some kind of periodic save mechanism to help them recover. If you were placed in some game situations in real life, you indeed could do something stupid and get yourself stuck in real life, so why not let it happen in the game? Coddling the player into always knowing they can never cause a irrecoverable situation just further degrades the genre into just another 'gaming' experience. Giving them a save-state to go back to seems very generous to me as a serious player. Then again, maybe I'm a bit hardened... I've been playing adventure games since some of the first text-adventure games were brand-new.

    More important to me is a bug-free experience, which I am happy to say I got with the Penumbra games. This was the first time I actually experienced a game with no noticeable bugs or glitches that would detract from the game. Other than the Penumbra series, it seems the entire industry is quite happy releasing alpha- or beta-quality code as retail package.

    The spatial awareness of the Penumbra games shows quite a mastery of the auditory connection to emotion. Some of the uses of sound actually made me want to stick my head through my screen to peek around a corner or over an object out of sheer fear-driven curiosity. I also found myself straining to listen for things in the silence, whether there was something near or not. It was a very realistic experience with the proper suspension of disbelief applied.

    On the visual side of spatial awareness, as any developer is aware, the very ability to move and place objects realistically in the game world and a masterful use of lighting and shadow can help paint an emotional canvas and further draw the user in. Creating a visually-realistic world, however, is nearly impossible with the average consumer's computer without risking performance issues, so this aspect always must be carefully balanced with other emotional hooks to enhance the experience. Again, I feel the Penumbra series handled this quite well.

    One thing I see too much in all types of 'gaming' experiences is a vast over-use of music tracks. This is not movie-making, it is adventure-immersion. If I'm looking to be drawn into a game, the first thing I do is turn off the music if given the option, because I can almost be guaranteed that it is overused in just about any game. Sometimes a lack of music can strengthen an emotional connection, but most devs don't seem to understand this.

  15. I'm going to disagree with the statement that sequence breaking takes away from a game. If there is an obvious way that a player can sequence break large or important parts of the game, and that it is obvious for most players how to do this, or that many people might do it accidentally instead of the intended rout, then I think that is a bad example of sequence breaking.

    However, I think often sequence breaking adds to a game. One game I love a lot is Super Metroid, mostly because once you get good at the game and have a good understanding of the mechanics, then there are a lot of areas in the game that you can try to sequence break, and this process of trying to break the game and do things other than play the same game over and over added a lot of replay value to it.

    I think this might be even better for horror games. After the first playthough of a horror game, most (if not all in many games) of the fear factor is gone. Even on games with random events, the fear just isn't there. People are probably going to play through a horror game again for reasons other than the thrill of being scared. Having things in there like sequence breaking, or hidden secrets they may have missed the first time helps keep a game alive longer.


Note: only a member of this blog may post a comment.