For no particular reason I was puzzling over Batman again this morning. Its combat was simplistic but really engaging, and one reason is because they nail the basics of action -- thinking about position and pressing different buttons in response to the enemy -- but another is how much movement is incorporated in combat. Over the years, I've come to believe that movement is a real boon to most avatar move set design, and just the simple addition of scooting Batman towards his next target before nailing him (perhaps initially added just to reduce frustration with targeting) really makes you feel more exhilarated.
Nowadays, I consider "movement good" a core design principle. Novel movement, like flying, grappling, gliding, running on walls, grinding on rails, skiing down hills (Tribes!) or curling up into a little ball and rolling through tubes are really fun. Even if you're doing something that doesn't normally require a lot of movement (like attacking someone with a sword swing) is better if it has a slight scoot. And things that slow you down, like blocking, slow climbing, slow ledge-grabbing, slow tight-rope walking, or slow scooting around the edge of a wall can really bother the player.
A lot still needs to be done with any avatar move set to get it feeling great. Just adding movement has little value unless it feels responsive and gives you proper feedback (e.g., the run in Gears of War is not much greater than normal speed, but it feels much greater because of the shaky cam and wider field of view). And like any rule, there are good times to break it (e.g., hiding in the shadows in a stealth or horror game as a guard passes by you). But in my opinion, movement is an extremely useful tool for gauging interesting or uninteresting avatar actions, and a good thought to have in my design toolbox.