I wasn't intending to post further, but since you've put this in a particular way I felt compelled to share my views on this matter... 4.x cards.
In your post you suggest that the old 4.x iteration of AST rewarded player skill more, but this statement is simply untrue. In fact many players in JP (and NA/EU as far as I can observe) that play this game at the highest level voice a very strong preference towards the 5.x card kit because its optimization factors are magnitudes deeper than the 4.x iteration of the kit.
Put simple:
4.x optimum: Spam AoE on trick. Very heavily luck dependent.
5.x optimum: Spread every party member's potency timeline and choose the maximum. Far smaller luck factors.
A more detailed explanation:
The 4.x optimum was spam AoE on trick, which meant you want two cards/60s, RR and the balance. Because draw and play shared cooldowns, there was nothing a player can do aside from hit draw every 30s and immediately get rid of it praying its the one you wanted. You could almost never draw balance + RR per 60s for the entire fight, so you'd find yourself doing a million compromises during pull; playing Spear, Arrow, non AoE cards, skipping a buff window for the next... the best possible option was blatantly obvious, yet you were only praying you could play that way. And MA was just even more RNG disaster.
On the other hand, I will quote the original post for the 5.x version:
Do you see that the number of options and the level of control we have toward achieving optimum are magnitudes higher than the previous iteration? We can pick the optimum target and oGCD window for every card in the fight. There are still luck factors (melee/ranged and Seals), but we can account for this and play pretty much optimally every pull.
----------
That said, I am in total agreement that RNG curse of looping between two cards feels terrible. I'll point to a post of my own regarding this, as an alternative solution to reverting to 4.x cards. I hope this post overall clarifies a few things for you.