Probably the legacy of 'not quite as advertised' console hardware.  The reason behind "palette ramps with only 2 or 3 shades per ramp" is that - although systems like the NES could theoretically display a roomy ~64colors, in practice, they had wacky additional limitations (like  "24 colors per scanline"), and all but the best devs (already running a huge risk in just doing game dev) played it safe by just sticking to that - just staying lower than 24 colors so they knew it'd always work.  Few companies could afford to take their chances with designing around the hacks; rare examples include Henk Nieborg's "Misadventures of Flink" for the Genesis, which apparently used a weird trick of changing the colors every other frame to get more colors.

I'm glad I wasn't involved in game dev back then.  Ick. 

Note for drazelic;  8-bit and 16-bit consoles is referring to the CPU, -not- the "bits per pixel".  Although they had higher theoretical limitations, in practice the limitations of "what could be onscreen without crazy hacks" were something like:  NES=4bpp, Sega=5bpp, SNES=8bpp, Genesis=6bpp.  Source= http://en.wikipedia.org/wiki/List_of_video_game_console_palettes

I suppose it's more stylistic concerns. 8-bit stuff tends towards higher contrast between shades, though I suppose looking at it from an outside technical perspective it complies to the palette requirements of 8-bit pixelart. To me, 8-bit has always meant palette ramps with two or three shades to a single ramp max.

The shades are all there on the bottom, excepting the green glowy trim. Surprisingly, on other forums people are telling me the bottom is better than the top. I suppose the different techniques attract different opinions?

How is this not 8-bit?

Hands and arms are the best part. Even if you went fluid on the lower half, the highlights would still be needed.