The problem you described in your second quote is exactly the thing SimC is trying to solve. You have a predetermined set of buffs, you pick a target. You have a go at it, write down your dps. Change one item (or talent, glyph, rotation, etc.) and go again. The difference is 50k dps favoring your new thing... what does this mean? You go again it's 20k worse than the old one... what does this mean? SimC allows you to recreate so many fights played out with the same behavior that bad or good luck become statistically irrelevant which allows you to make a judgement call on whether you keep your new item/talent/glyph/rotation or revert to the old one.
The published SimulationCraft "class rankings" were exposed to provoke the exact same discussion and innovative thinking that's going on in this (and this) thread: simc says I should do 300k, I do 220k. Please help!. Without the published rankings many more DPS players would be unaware of their shortcomings and would never get enough kills on all bosses to become retroactively involved in fixing them (even the most hardcore audience gets only 15-25 kills max per boss per difficulty when that raid is the current tier).
So imagine SimC as a research tool. If your mentality is "I think I could do more dps than this. Lets find out!" then you're on a healthy path to improve and maximize your performance (not necessarily only dps) in a raiding guild or other environment and SimC will be a huge time saver. On the other hand, if your primary attitude is "my dps is fine, the SimC ranks are full of !@#$%" you'll most likely become stuck with mediocre performance with very limited options to improve.