I’ve been looking for a good definition of a puzzle that is satisfactory to me and have been unable to find one. So I went through the highly unscientific method of reverse-engineering what I thought a puzzle is with my colleagues and came up with the following. But I still think it is missing something. Any help is appreciated.

A puzzle is a game that uses cognitive reasoning (puzzle-solving skills) to get from an unsolved state to a solved state, with some exceptions:

  • The puzzle cannot be trivial: if I give you a picture of a light switch and ask you to solve it, you do not have to use any puzzle-solving skills to flip the only switch that is there. In the same way, Tic-Tac-Toe is not a puzzle for any adult because any adult can “solve” it and tie or win as long as he goes first. This does make the definition of a puzzle subjective for the audience it covers, but making a trivial puzzle and saying it is for toddlers is a cop-out.
  • A puzzle must involve intellectual effort to get from unsolved to solved. If the only way to solve your puzzle is by brute force, it is not a puzzle. “Press the right combination of buttons” with no other prompting or clues is not a puzzle. “I am thinking of a number from 1 to 64” is not a puzzle if you respond simply with yes/no, but is a (IMO, fairly weak) puzzle if you respond with “higher” or “lower” and only get six guesses.
  • Solving the puzzle must be the same as winning the game. Checkers is a “solved” game. But solving checkers is a different intellectual exercise than winning the game. The goal of solving checkers is to create a strategy that always wins, but winning checkers is about jumping all of the opponent’s pieces. Solving a jigsaw puzzle is the same as completing the jigsaw puzzle.
  • A puzzle can be generated randomly, but must be deterministic once the player encounters it. A board of Sudoku can be generated pseudo-randomly, but once the player starts the puzzle, every player that makes the same moves will experience the puzzle in the same way. If you and I get the same Minesweeper board and uncover the same squares in the same order, we will have identical experiences. If you and I play tennis and make the exact same movements, we will have a very different game experience. Chess, unless against an AI specifically designed for this purpose, is not determinsitic. If I make the same five opening moves, five different players may play it in five (or more) different ways. You can make chess a puzzle by giving a set of moves and deterministic rules – there are examples in puzzle magazines where you must mate in X moves with rules for how the opponent will move.

Sports Simulation Design in a Nutshell

Why sports games have to reinvent their controls annually, I don’t know. Puppeteering a guy dribbling or throwing a baseball would seem to be a basically constant thing in a video game compared to a guy hucking a grenade or acrobatically slashing his foe across his torso. Yet the latter two acts have undergone less change on a video game controller in their respective series, going back more than five years.

This was on Kotaku yesterday and it is an excellent point. As a former disgruntled sports game designer, I will try to explain exactly why.

The tipping point was marketing, lead by EA: “If it’s in the game, it’s in the game.” From that point on, the focus for both producers and consumers wasn’t about making a fun game (although that was still the secondary goal), it was about simulating the real experience.

And so sports games spend every yearly iteration adding more and more to get it closer and closer to the real thing. Awesome, right? Let’s take American football as an example. Sometime in that chain of Maddens, they added custom playbooks. Cool. They added audibles. Cool. They added custom packages. Alright, I guess? Now it is getting harder to come up with low-hanging fruit that isn’t in the game. What else can we simulate? Maybe the hand-play between a receiver and a defensive back downfield on a pass? Maybe we should simulate the QB’s eyes as he looks off receivers? (Don’t be ridiculous.)

Football video game players are generally pretty familiar with the real world mechanics of playing football. There are certainly a lot of them to choose from: running, blocking, passing, play selection, time management, coverage schemes, shifts, choosing plays and formations, and so on and so on. How do we fit all of that into the five chunks of working memory that we have?

The short answer is that we can’t.

As long as we are trying to simulate a thousand little things, only the people who have deep understanding of not only those thousand little things but how we’ve chosen to represent them in our game will be able to actually appreciate it. And it will always feel off because each of us will have a different idea about how that should work on-screen. We never had that problem in Tecmo Bowl because we weren’t as close to the Uncanny Valley. We had some sprites and two buttons; we imagined the hand-play, we imagined the blocking battles that didn’t really exist.

Sports game designers have a panacea that they always fall back on. The problem to them is never that you are asking too much of the player, just that you are asking the player in the wrong way. That’s why you see controller redesigns every year. If we just put juke moves on the right stick, then maybe everyone will get it this year!

Call of Duty doesn’t have that problem. Call of Duty doesn’t have to fulfill the expectations of a firefight, it has to fulfill the expectations of a first-person shooter. It would be moronic for the new Call of Duty to say “Now including wind that changes the trajectory of your bullets!” or “Now including cleaning your gun after battles!” More realistic maybe, but not fulfilling the expectations of the audience. Madden, for instance, isn’t simulating Tecmo Bowl. It is simulating the NFL – a real experience. If NBA 2K was simulating NBA Jam, you probably wouldn’t see changes to the control scheme every year.

Sports games are stuck marketing to the same, shrinking audience every year. As what’s “in the game” increases, the learning curve steepens. Because most of the designers are decades-long experts at playing these games, they don’t notice how utterly confounding it is. They scoff at notions of accessibility, pandering to concerns with a never-viewed tutorial or a condescending “beginner” mode that no one in their right mind would choose when playing against their buddies. The full super-hardcore mode is aspirational. Maybe we will never use offensive line shifts, but we like to know that they are there.

When developers do their consumer research, it starts with the loudest forum nerds. “WHY AREN’T THERE MEDICAL REDSHIRTS IN NCAA,” they scream as if anyone but them actually gave half a damn. The lack of medical redshirtting is absolutely breaking that one guy’s mental model of the simulation. And so designers and producers see this enthusiasm and go “Yeah, that would be pretty easy. I guess that’s what people want.”

But no one is putting down the game because there isn’t a perfect recreation of some byzantine NCAA rule. They are putting the game down because it is offering to fulfill that player’s mental model and failing. Playtest feedback never says that. I should know; I’ve been in a lot of playtest sessions. It’s never that lucid. Playtesters will either say or insinuate “I don’t get the controls.” Because obviously the controls are to blame! They are the interface between the player’s desired outcome and the actual outcome!

So what do you think the game makers try to fix next year?