I missed this article from two days ago, but there's a good preview of Dawn of War II over at IGN (link).

I'm not a huge fan of innovation in video games. I don't think it's bad, just dangerous. In a lot of ways, innovation is like a scorpion holding a gun loaded with a stick of dynamite. In many ways extremely awesome, in many ways extremely dangerous.

So I'm all about using existing mechanics in new ways. Mechanic X works really well in genre A. Will it work in genre B?

So I was pretty excited to see this :

The scene looks more like something from Diablo II than a real-time strategy game. The warboss even has a boss health meter visible at the screen's top and special attacks, like a shockwave move and a ground attack that lights up terrain in a line before a string of explosions triggered. The tactics of taking him down involve moving troops out of the way of his ground shockwave, steering clear of his melee attacks, and keeping a squad focused on clearing out the Ork rabble that keeps spawning. Once he is finally killed, he drops his gun, which is labeled as a "blue" item.

You'll find iconic boss fights in pretty much every genre of combat-related video games -- except the RTS. There might be a bigger, stronger enemy to kill at the end of the game, but they are generally more of the same, and the final battle rarely requires anything out of the ordinary.

So kudos to Relic for using a well-understood, fun mechanic in an entirely new way. I consider this "safe" innovation. You can add something new and fun to your game with perhaps a lot of effort but without a lot of risk.

And any game developers out there : I'm still waiting on my RPG that uses pinball as the conflict mechanic. It's entirely safe to do (since pinball is already fun) and it would be freaking awesome.