I'm not a fast thinker .  Luckily otb is 2 hours per person, so it is not a problem .

The impression I had with blitz is that it is much more about patterns recognition. What really annoys me in blitz is that players (at least in my level) are counting with the stupidness of the oponent (e4 Qh5, really?) and silly traps that would be easily detected & prevented (& laughteable) in standard chess. Many things are about memorization these sequences and patterns = not fun or even interesting for me.

Correct me if I am wrong, but I suppose that blitz and standard games are two different worlds, and I am not even considering to compare rapid chess with standard OTB (reason why I've mentioned only standard and online ratings on my original post)

If I am not going insane, I remember a Chess Mentor Lesson saying that this opening is bad (you play black against this opening) and it explains why it often is bad.

I can't find it though, so maybe I was dreaming or something. You can check the openings explorer though, and you will see that normally white either draws or lose, except for one blunder game.

Granting ratings won't match, there is certainly no reason for them not to correlate, and given a sufficient sample of players who have played a sufficient sample of games under both systems correlation would be very much expected.  For there to be no correlation, one or both of the systems would have to be meaningless, which is almost certainly not the case.  it is even possible to show (has been shown) a correlation between for example the British Chess Federation system and FIDE/USCF, even though BCF uses a three digit number instead of a four digit number.  For example:    ECF x 8 + 650 = FIDE and   (FIDE - 650) / 8 = ECF.  Now, will that be exact?  Of course not.  But predictive correlation is not the same thing as exact matching.