First point: No, EOline is not objectively less fingertricky than cross. I can consistently sub 2 my EOlines, and about 80-85% of the time, sub 1.5 them. I can say with absolute certainty, they are often quite fingertrick friendly. If you were to do the completely optimal solution, they wouldn't be. But this is like comparing BH to 3-style for BLD. You sacrifice 1-2 moves on average for a vastly more fingertrick friendly EOline.
Second point: Again, this is completely false. If you switch directly from CFOP, maybe for the first week or two, it'll be worse. However, each method has a distinct turning style. Like CFOP users quickly regrip and can do RUF moves very quickly and Roux users are able to switch between R and r for 2nd block and typically turnslower and more fluidly, ZZ users have a distinct solving style as well. If you look at someone like Phil Yu, and watch his solves, you will see one of the best examples of a ZZ turning style out there. He shows how working with ZZ yields turning that utilizes both hands working independently, and turning with both hands at a very similar speed. If you master being able to use the left hand as fluidly as the right, saying "The RUL moveset makes fingertricks way worse" is totally untrue. As well, I would say regrips are vastly less common in ZZ than even CFOP, because you never have to rotate, so you only regrip if an alg depends on it. Also, because you are doing the 2 F2L pairs in the back from the front rather than rotating, that cuts out even more regrips because a ZZ user can fluidly go directly into the pairs, whereas a CFOP user would typically rotate to gain ergonomics that they otherwise lack due to the lack of EO. And the whole "it also makes the F2L too restricted" point is just totally insane. If anything, CFOP f2l is too restrictive because you always need to reverse any L or R move so as to not disturb the cross pieces, whereas with ZZ, the L and R layers are completely free to move. You don't do F or B moves because the best ways to do every case don't require it. Nearly all the cases that are best solved with F and B moves are the ones with a misoriented edge. I'm not going to cover the whole multislotting thing, because Pyjam already covered why ZZ has arguably better multislotting.
Third point: This is probably the point with the most legitimacy, but it matters far less than you make it out to be. I generally plan out a 3/4 cross to help with lookahead, which typically only adds 1-2 moves onto the EOline. This cuts the potential number of L2/R2s in half. And even with that, the few that I do are slightly worse than just an L or an R, but this is mainly seen as a major con by CFOP users simply because they aren't used to them. While rotations may not be a big deal to you, L2s and R2s aren't a big deal to me, and they tend to not really make the f2l any worse. The worst part about the unsolved D edges is typically the "blind spot" it creates, and if you either take note in inspection to see what that piece will be, it gets rid of the problem entirely.
Now, I don't think ZZ is without its flaws. I think that it requires much more experience in ZZ than CFOP to be able to effectively look into F2L in inspection time. It took me 2 years to be able to do it fairly well. However, the amazing LL, and the rotationless and really ergonomic (once you get used to RUL) F2L makes up for any shortcomings that it has compared to CFOP, and imo, makes it have a slight edge. CFOP has maintained popularity due to how many people are fast with it, as well as its easy transition from LBL. When you look at the grand scheme of things, CFOP is incredibly old compared to ZZ. It has been around for less time than some top speedcubers, and it's been truly developed for even less time than that. CFOP was developed throughout the 80s and published officially in 95, whereas ZZ was first published in 2006, and was used almost solely by Polish speedcubers for a long time due to it being published in Polish. It was only when Conrad Rider popularized it with his English guide did it pick up any steam, and that was in 2009. When you compare 8 years vs 18 years, it's obvious that CFOP is more developed. Even Roux was developed in 2003, and there are still vastly fewer people using it. ZZ is still an incredibly new method, and it still needs more time and more users to catch up with CFOP.
(Final note about the opening sentences, that's awesome, you should totally use this in some type of school debate paper. Debate freaking rocks. Sorry this was so long btw, just wanted to get all my points across. In retrospect, like 900 words may have been a bit much)
