[alanfalcon]

Proper planning is key, both to ensure that you don't find yourself suddenly and unexpectedly needing 500 new images, and to find areas for optimization.

What's the value in having enemies have slight differences in appearance as they get close to death? Will players actually notice or care about this? In a 2D games, it probably makes a lot more economical sense to add some kind of procedural effect to show this, such as tinting them red, making them flash, or showing a bright red outline.

[jarin]

The value is polish.

He's not making a data-driven business app, where a shadow here or a custom icon there are nice but ultimately unnecessary. He's making a piece of software in a field where the small details can often be the difference between obscurity and massive success.

Of course, there are exceptions to the rule (Dungeon Raid, Doodle God, etc), but they became successful in spite of their lack of polish by having excellent and novel gameplay. When you're "borrowing" someone else's successful formula, you're competing with all of the crappy clones out there and need to put in a little extra effort to stand out.

Plus, he's making the game for fun, and it's a hell of a lot more fun to make something you're really proud of.

[SergeDavid]

That can work (as I'll dot that for frozen aka slow characters or rage aka fast characters) However certain things just can't be replaced, like a big bite out of a plant or a black eye on another character with blood on the corner of their mouth.

[bazookaBen]

from the shoes of an indie dev, we can't afford 500 new images so we tend to use a lot of editing (i.e particles can have different colors, shapes, sizes ). Can't introduce so many 'new'/'groundbreaking' images.