Thursday, July 13, 2006

Metagaming: Puzzle design and strategy guides

While recently playing Tomb Raider:Legend I got stuck on a few of the sections in the game because I didn't know what to do next. In times past I would have poked and prodded the environment to find the solution. These days I simply don't have that level of patience anymore. Game solutions should be logical and elegant. I believe many developers are doing the best they can to achieve this but the truth is that sometimes the puzzles and challenges we create don't make sense to others in the same way.

SPOILER ALERT for Tomb Raider: Legend
Here are examples of the puzzles that confused me. They all take place in the England level, King Arthur's tomb. This level made you use a number of objects for the first time without explaining their usage. For example, to get down to the tomb you need to use a forklift to destroy a wall. In previous levels you rarely ever destroyed the environment but if you did it usually involved using your grappling hook or grenades. In addition, your primary reference for thinking about forklifts in real life is for the purpose of picking up heavy objects and moving them to another location. When you first see the forklift in the room you also see many crates stacked on top of each other and notice a potential exit up and above you that you may need to use the crates for. I assumed that I needed to use this mechanic to progress through the level rather than bashing a wall.

This mechanic could have used some type of introduction, such as:
1. Have a worker destroy a wall or other environmental object in a cutscene
2. Use a bulldozer instead. I believe bulldozers make more sense to people as demolition vehicles.
3. Slightly weaker solution - force the player to drive the forklift at high speed and destroy an object. I think the puzzle was initially trying to enforce this on you since I accidentally backed into a tomb and watched it fall apart, but that whole interaction was a little strange to begin with.

A bit later in the level, you enter a room with a statue encased in some sort of glass shell. It seems apparent that you need to break it. When you look around the room you see a bell and a chandelier that you can interact with. There are also lots of tombs about. In a previous part of the level you used a chandelier to break open a part of the environment so your first inclination may be to see if the chandelier can break the glass. Upon grappling it and making it swing you notice it doesn't quite reach the shell. Ok, you think maybe the bell has a role to play in the puzzle. You jump on top of it, hit it, etc. It makes some noise but not much really happens. You decide to explore the room some more. As you walk up to one of the tombs a cutscene ensues and you learn that you can take half of the tomb and push it around like a crate. Previously you had only used crates and boulders to create steps. Once you push the tomb around you notice a broken pillar that you can reach and a rope you can jump to from that pillar. This has the effect of acting as a lever for the bell, effectively moving it upwards. Now you notice that the bell can be struck by the chandelier. The trouble is, you have to walk to a different part of the room, grapple the chandelier, jump on the tomb, jump on the pillar, jump to the rope and hope that the chandelier still has enough momentum to hit the bell. If it works it triggers a crack in the glass and now you can walk outside again.

To keep the puzzle solution similar to the one in game I would have either allowed the chandelier to swing for a longer time, allowed it to be grappled from the pillar, or allowed the bell to hang in the air for a while. My guess as to the reason none of these approaches was taken has to do with the intrinsic physics values assigned to the rope and chandelier. They probably wanted to keep them consistent for all objects in the game, which makes sense.

Soon after, you encounter the boss. Once again, another strange mechanic that had never been introduced before is the key to winning the battle. You have to shoot some pillars over and over to make sounds that attract a water beast to listen, at which point you switch to your grappling hook and latch onto a lever that drops a cage on its head. Up to this point the boss encounters mostly consisted of shooting enemies with your guns. I understand they wanted to use some variety in the battle but this game mechanic was also cited as confusing in the guide that I read to understand the solution. It relied on a new mechanic that is never explained to the player.

This whole level had mechanics that did not adequately tap into knowledge gained from the game or real life and I simply cannot understand why it is assumed the player would figure it out. There is a device you can use to get hints on what you can manipulate in the environment but the comments that Lara or her allies make when using these binoculars give you no indication as to what role it plays in the puzzle. It seems like Metroid Prime's scan visor was the inspiration for this device but the difference is that the scan visor in Metroid Prime gives you very specific clues on what type of weapon you may need to use to destroy something or alludes to the fact that in order to solve the puzzle you may need to use other objects in conjunction with the object being scanned. One final note: this game is actually one of the most enjoyable games I've played in a while, so my criticisms here are only cited to give a more detailed explanation of where I'm coming from.

END SPOILERS

So I went over to GameFAQs to expedite my journey. This is one of my favorite resources for finding hints on games. I'm not a big retail strategy guide fan. I like the fact that there is a community devoted to making awesome strategy documents for just about every game. For many games I find myself checking out this site at least a few times to either find solutions or track down any hidden Easter eggs that I missed. When searching for hints this time, I was struck with a slightly disturbing thought. Are we, as developers, counting on strategy guides and FAQs to make up for any shortcomings in our puzzle design?

I doubt that any developer intends for this to happen. We'd like to believe that our puzzles make sense to the majority of players. The problem is the ease with which gamers can turn to outside resources for any solution they may need. I certainly get annoyed when I feel like I have to do this, but I have gotten used to it. I don't really think twice about checking a resource unless it's a game I really enjoy and absolutely feel the need to find everything on my own. Metroid: Zero Mission was one such game where I didn't necessarily know the method for acquiring each item but at least I could visibly see them and develop strategies for their acquisition. Some games may stump me but if I really think abut what the game is trying to teach I can generally come to a solution. Zelda games fall in this category. More often than not the solution involves using a recently acquired item, but not always in the way I had originally expected.

What about those games where the solution is not so apparent? I see this problem more often in games with realistic graphics and environments with objects that don't stand out a whole lot from the area around them. In many cases the way you are asked to use the object is not obvious from its appearance. Certainly there are ways to highlight these objects more and to give the player hints as to how to proceed. I imagine that for well polished games what tends to happen is that the designers decide on what they feel is logical and focus test the puzzle to confirm that their groups solve it in a reasonable time span. The trouble with these focus tests is that they represent a small sample size. What about the rest of the potentially millions of gamers that haven't sampled the puzzle?

They either get bored, frustrated, or seek out a solution. I tend to hit the outside resources more often towards the end of the game when I feel close to completing it. Sometimes I even find the right solution but I didn't push a box all the way in or align every aspect of the puzzle just so. Those situations are unfortunate because the solution may be mostly correct but having a binary feedback system makes me question my strategy, verify it from an outside source, then go back to trying to make the game understand what I wanted.

I sincerely hope that someone at each company is tracking which puzzles stump players the most and is sending that feedback to the team. While the sharp, experienced developers may have already learned a lot of these lessons we need to pass on the information to the upcoming designers who may create puzzles they find seemingly logical but don't quite work as well in practice. I also encourage developers to try finishing their own game, even if they are an artist or engineer so that the designers have more feedback on how intuitive the puzzle design is. I've seen situations where game elements needed to be simplified but no one had said anything to the person in charge of that section.

There are several steps I believe we can take to help with puzzles becoming more intuitive. The first step is providing a tiered feedback system. Why force players to run patterns over and over without giving them any advice on the approach they are taking? This feedback could be in the form of sound effects that range from a tone indicating total failure to one indicating total success. Zelda does a fantastic job demonstrating what an iconic success tone sounds like. Another mechanism we can use is including a hint gathering tool. I don't like the ones that force you to drop out of the game into a menu, but the ones that are a natural part of the interface (such as the scan visor in Metroid Prime) provide an elegant immersive solution to providing a hint system. Another important way to highlight puzzle solutions is through the graphical style we use to represent the objects in our games. While we may want to have the sharpest lighting and most realistic textures, we should think carefully about the way we point out objects the player can interact with. Can placing an object in a different part of the room set off a clue as to its utility in a given situation? Absolutely - Tomb Raider: Legend had several cages and blocks placed at just the right spot to force you to see both the object and its intended use in the same screen. These are just a few ways I believe we can make puzzle solutions a bit more natural. Focus testing should also be done on puzzle areas with measurements taken on how long it takes a player to get through a given section. Ideally these tests would be done with several different groups of players of varying skill levels and game knowledge.

Strategy guides and FAQs should serve as guides and not be a near necessity for many players to complete games. As I find myself reaching for these resources more often I take notes and build up my library of thoughts on level design. I propose aiming for the goal of not needing a FAQ to complete a game. Ideally, for me, each game that uses puzzles would stump the player on several occasions but wouldn't push too many players into a FAQ hunt for clues if they are trying to finish the game on a first play through. For hidden goodies and unlockables I'm all for making trickier problems but to reach a broad audience I think it's important to keep players moving through a game at a reasonable pace. Puzzle solutions need not be simple, but shooting for elegant and intuitive seems feasible.