Our board game was called Gautar Hero (!) and named after the ancient kingdom of Gautar in Sweden. It was a clever asymmetric four-player game, set in a Viking town where a man-eating snake is on the loose. All the judges thought it was a very good game, and we gave it a special mention during the awards ceremony. (You can download the rules, and the graphics for the board and cards, at the Global Game Jam Leeuwarden web site.)
Reviewing the video games, even though I'm a designer, I gave more weight to completeness than to design innovation. I have seen so many student projects (and commercial projects, too, for that matter) fail through overambition and inability to scope correctly that I give major props to anybody who can build a complete and functional game on time.
I've also worked with people who were all big creative talk and never got anything done because they couldn't commit to a decision. So to me, a small finished game is better than a big broken one. Of course, a game still has to be reasonably imaginative; even if it were finished, I wouldn't have rated a Tetris clone very highly.
Two of our games included the concept of health points but had no health bar, so the player couldn't tell how much damage he was taking. The developers knew they needed one; they just didn't have time to implement it. I considered this a demerit, but the other judges didn't.
I think I was the only judge who took programming difficulty into account. One of our games required some tricky coding, but either the other judges didn't notice or didn't consider that an important criterion. Our artist judge was more concerned about design innovation than he was about artistic quality, which surprised me. Our audio engineer judge rated one game strongest that I rated weakest.
We also had very different experiences playing the games. Some judges observed gameplay problems that never occurred while I was playing. Games that I thought were fun, others thought were boring, and vice versa. Each judge played them by himself rather than with the other judges watching, which may have been a mistake.
I'm not complaining about the outcome; I'm happy to stand behind our collective decision, and so are the other judges. Our winning game, which was entitled We couldn't think of a funny title so we call it 'Walkabout' or 'The Great Balance.' Pick one, included artistic, audio and gameplay innovations, and genuinely deserved to win.
It seems to me that game judges have four major areas for discussion: programming, graphics, sound, and gameplay. You can judge each of these on three metrics: innovation, quality, and completeness. You might also consider a fourth, size, but in the context of a game jam I don't think there's much point. Other things being equal, a bigger game requires more work and deserves more kudos than a smaller one, but other things never really are equal, and it would be a shame for participants to sacrifice anything else just for size.
Programming innovation will be difficult to determine because the judges are unlikely to see the code. If the developers include some innovative but entirely hidden algorithm (in a combat model, for example), the judges will never notice unless the programmers tell them.
However, if the game tries to do something more difficult than most of the other games it's competing with -- it includes AI, for example, or more realistic physics -- then I think that qualifies as innovation in the context of a 48-hour jam. Quality means no bugs or crashes, obviously. Completeness means that the whole intended feature set is in and the game plays from a beginning to an end (even if the end comes soon -- most of our games only had one level).
Graphics innovation is easy to spot: are the images and animation different from the things we've seen a thousand times before? Quality likewise is fairly straightforward by conventional video game standards, although if a game goes for an avant-garde or retro look the judges will start debating their aesthetic preferences, which could get tricky. Completeness means that all the intended graphical elements are there, including user interface elements (such as health bars!).