While I support ZZ and consider it equal if not better than CFOP in speedsolving potential, I don't support your points.
1. RULD is significantly better that RUFLDy; Only ZZ users can appreciate that. There really aren't any regrips (you'll know if you try ZZ). Plus every rotation could be counted as, say, two moves. This number will only go up as you get faster and faster at F2L. This would completely mitigate the extra moves required to do EOCross. Also, things like psuedoslotting are so easy in ZZ because there's no way you can accidentally get a flipped edge.
RUL reduction for F2L is not a significant improvement. ZZ avoids rotations, but suffer from gripshifts between R/L turning (CFOP, on the other hand, has general hand dominance). Overall these faults roughly cancel, in my opinion. Pseudoslotting and multislotting is not as great of an asset in ZZ as it is often made to be in the ZZ community (yes, I am also referring to myself) and is not a tangible advantage.
2. Tymon and Max both know a ton of ZBLL. That's just another reason to use ZZ, as you can 1LLL every solve if you are willing to learn the algs. Also, it's a lot more feasible for the average cuber to learn 493 methods than 3915. Even learning 2000 algorithms of 1LLL will give you a 1LLL just half the time. Plus, for 2LLL, you only have to learn 28 algs instead of the usual 78. By the way, ZZ has a lot of variants and flexibility for last slot + last layer that you just can't find in CFOP.
Regarding LSLL in general and not your point specifically. I don't have much faith in 1LLL and I think it has no future. There are too many cases and the recognition and recall are pretty bad. A clearly better approach for CFOP would be ZBLS/ZBLL -- about the same movecount and way more developed and trusted, even used at the high level. More about that later. ZZ has no relevant LSLL approaches besides LS ZBLL.
3. I would like to know what you mean by saying that more skips isn't really better? Since there are essentially half the F2L cases in ZZ, lookahead is made much easier and free pairs are twice as common. In CFOP, an OLL skip is 1/216, but in ZZ, it's just 1/27.
Comparing skips is irrelevant, as they are present in both methods in some form or another. Lookahead is not necessarily easier in ZZ. More on that later.
4. EOCross isn't really that hard. All you have to do is plan around 10 moves each time in advance, and in CFOP you have to plan around 7-8.
Both of those numbers are wrong, but that is unimportant. You must account for the planning *after* cross. See below.
The fundamental question when comparing CFOP to eocross ZZ is this: is it worth it to be a full pair behind to have EO solved? xcross and eocross are roughly comparable, which illustrates the issue with ZZ: the extra moves spent on solving EO during cross could have been used for a pair. So for ZZ, we have 3 pairs to LSLL whereas CFOP is 2 pairs to LSLL. I don't think there's much benefit in F2L from ZZ's EO being solved, because while is promotes consistency it does not allow for great flexibility like CFOP. Therefore, I consider each ZZ pair before LS (ZZ LS is stupid fast) to be just as bad (or good) as a CFOP pair. So ZZ chugs along a clean pair behind, but you get easy ZBLL. (I don't think there's any question that heading for ZBLL is the meta, as yoy can see from top cubers.) In CFOP, you must perform influencing during LSLL: ZBLS. Now ZBLS is actually quite bad compared to ZZ's LS: ZZ LS is literally like sub 0.3 on average with no theoretical recognition time because of prediction from the previous pair, whereas ZBLS, having a barely sub 10 movecount avg with questionable movesets and even higher with normal movesets, proabably is barely sub 1 including the recognition time (cases which cannot be predicted)
So the objective comparison between ZZ and CFOP in my view is between ZZ last 2 pairs (first one being an average pair and the second being predicted and stupidly fast) and ZBLS (algorithmic, ruins flow, either similar movecount as ZZ last 2 pairs or questionable moveset). The other differences that subjectively benefit one solver or another are eocross vs xcross (very different skill sets) and CFOP f2l dynamics and flexibility vs ZZ principled and consistent f2l.
Perhaps I should make a thread so i don't have to keep reiterating these points.
