Puzzles vs. Problems

Okay so I made it past the steamboat level in Hitman 4 and on to the final level, which I hope to finish today. But Blood Money (and Warren’s latest blog post) got me thinking again about the idea of problem-oriented game design. I’d like to think the idea is mine, but first of all it’s doubtlessly been formulated differently by many other people before, and secondly even if I were the first person to put these ideas into words, entire games seem to have already been based on them. Hitman 4 being the best example I can think of right now. Better, even, than Deus Ex.

The idea is that you can approach the creation of a challenge for a game in two ways: You can treat it as a puzzle, where there is only one way for your player to overcome the obstacle you place for him, or you can treat it as a problem where no specific solution is given, just a premise and a set of “tools”. Let me exemplify.

Half-Life 2 is a fantastic game which is heavily puzzle-oriented on several levels. Most obviously, the ordinary run ‘n’ gun shooter gameplay is frequently interrupted by an honest-to-God puzzle – generally based on some aspect of Havok’s impressive physics engine – which you have to solve in order to proceed through the game. On a somewhat higher level, the whole game is extremely linear and gives you only one choice in how to handle enemies: Kill them. The only choices you have in Half-Life 2 is whether to search for hidden items in certain nooks and crannies or just run past them, and which weapon to kill your enemies with. It’s a great shooter, but it’s definitely not going to ask you what you want to do.

By contrast, Hitman 4 puts you in a situation and presents you with a huge array of tools and tells you simply “Kill this guy and find the microfilm he’s hiding.” The tools at your disposal include going in guns blazing, sneaking past guards and security cameras, distracting guards into leaving their posts, climbing over walls and through windows a’la Splinter Cell, conning your way past obstacles with good old social engineering, neutralizing enemies or civilians more or less discretely and donning their clothes to pass effortlessly through restricted areas, or poisoning food and letting servants do the rest of the job. To take the optimal route through each level, you’ll usually have to use many of these techniques at different times in a mission, but often it’s possible to choose a single technique and stick with it, it’ll just be a lot harder. Additionally, Blood Money has a scoring system to steer you in the direction of playing the game like the developers want you to (so if you go in guns blazing, you’ll barely get paid), but once you’ve completed a mission by the book, you can always go back and play it again as a regular terrorist, blasting through to your target with full kevlar and a pimped-out assault rifle.

That’s what I mean by problem-oriented gameplay. Your problem is that you need to kill somebody, but he’s protected by approximately 20 well-armed and paranoid guards and sitting in a Las Vegas casino surrounded by civilians. You have a whole arsenal of weapons at your disposal and more tricks up your sleeve than David Copperfield. The game’s artificial intelligence is equipped to handle anything you can possibly think to throw at it, and it’s now up to you to devise a strategy and combine all these options to take out your target and not get yourself killed in the process. Even if you play the game by the book, laying low and carefully getting close enough to your target that you can safely take him or her out, you still have many many options for how to make the kill. Personally I enjoy creating “accidents” whenever possible – they look particularly good on the post-mission score board.

I want more games like this. I want more games that don’t tell you exactly how to complete your objectives, but just give you a bunch of options and drop you into a complicated situation. In fact I would like to try and take it further. I would like an entire game constructed as one big problem with an overall objective and no predefined path through the game. This game would simply consist of a large world, a micro-cosmos a’la Black Mesa or Citadel Station but structured more like Cyrodiil, filled with interesting situations and complicated, wide-open problems for the player to solve as part of his or her efforts to achieve the final objectives of the game.