To throw my perspective on the subject into the ring:
I use fibonacci retracements and extensions as suggestions for where future S/R will occur (I don't personally put a lot of faith in traditional S/R, e.g. swing highs/lows as entry points). Although I draw these lines several times during a trading session, I rarely enter blindly with a limit order at a line touch during a pullback.
I mostly use fib levels as a tool for objectively measuring the strength of a trend, pullback, extention or reversal. I also use them as targets when counter-trend trading (retracements) or when trend-trading (extensions).
Most of my trades are entered on stops after a price action setup (e.g. what you're calling confirmation I think), and I enter tight price-based exit stops immediately after entry. For only two setups will I enter blindly enter at a fib line... the first 23% retracement of a sharp reversal, or the 100% extension of the third leg of an old trend (I do the 2nd very rarely). I use very wide stops for both scenarios, as I want the market to prove me wrong first. The key is to be stubborn when you're right, without being stupid.
The first pullback after a sudden change of momentum is the best scenario I know of in trading... the distance the market has to move against you to prove you wrong is minimal, the distance it can travel in your direction is hypothetically unlimited (although you can't go wrong targeting the next major S/R level after a trend extension). However, the first pullback after a significant momentum shift can be very choppy. I can't tell you how many times I've entered on a price action setup only to be whipped out with a tight stop, just to see price go my way again (I usually re-enter, but now part of my profits have to pay for that loss).
I don't personally understand how guys can enter blindly at a S/R level and also use a tight stop. I always assume the market wants to find out what's on the other side of the S/R level, and only after it founds out there's a whole lot of nothing there (e.g. no selling/buying interest) does it tend to move in the expected direction. That's why I use wide stops when fading a pullback or extension. The extra price you theoretically pay with a wider stop is reimbursed by the extra profit you'll make if you were right (than if you waited for "confirmation"). Given the fact that you'll also experience less whipsaws, I believe the wide stop in these scenarios has greater expectancy than a tight stop, but only if you have the discipline to manually take yourself out at a small loss on the next retracement in your favor should the market prove you wrong (before the wide stop is hit).
In summary, I believe both entry techniques are valid and appropriate when used at the right time. Just like everthing in trading, it's all about context and experience with your edge.