Design by committee is definitely the way to go wrong.
Depends on the size of the committee. The magic number seems to be 7 people or less in software. I find game design to be a similar process. Some folks do their best work solo, of course.
Regarding combat. It's been a hotly discussed topic on this board. I'm a little reluctant to dive in (again), but I have rather charitable thoughts about T5.0 combat. I ran 7 combats (2 fistfights and 5 gun fights) in the two years I ran my campaign. I found STAMP to be good for gunfights -- even a rather complex one, but fist fights were better done with opposing task rolls.
This all comes with a big fat caveat. I much prefer guidelines to "rules" and I found the guidelines of T5.0 combat (and to a greater extent, the whole of T5.0) agreed with me and my group. I never tried T5.09 combat. At the point I was at in my game when T5.09 showed up, I wasn't ready to learn and teach a new combat system.
After two years of running T5 and going through a roster of about a dozen players in that time I found the system to be one that's very good for tinkerers -- both players and referees. With the aid of a computer, T5 is great for sandboxing since you can make very detailed locations very quickly. My players tended to max out designer along with trade skills (Biologics, Hydraulics, etc) and really put the makers to the test, along with robots, brains, AI and nanotech. We explored transhumanism when one of the characters secretly cloned another character and took a personality scan, made several copies and then released them to fight each other Thunderdome-style when the original was caught up in a grey-goo mishap. It got weird. Then it got weirder.
My point is that it was T5 that allowed us this two years of shared, undistilled sci-fi weirdness. We had zero-G gun fights, and powered armor warriors assaulting AI-controlled talking space ships, ancient tech, and clone duels to the death. It's the best game I've run to date. It was glorious. T5 is indeed, awesome.