I'm kind of on the same page as monk regarding 4 v 5. I didn't hate 4 so much as I had zero interest in playing. I'm finding 5 more interesting (at least if I can get a coherent plan together for what I want to do in a particular game).

As for besieging cities, I don't think it's worth it unless you're willing to become a warmonger. In a current game as the Greeks I started out hoping to ally with all the citystates (something I'm slowly doing now) but decided to take a nearby city (the Babylonians put a new city exactly where I wanted to expand to). It took maybe 30 turns to build up enough hoplites and cavalry to take it (strangely, cav is better than inf for attacking cities, at least in the classical/medieval eras). However, I now had a big enough force to finish off the Babylonian capital and destroy another city the English dropped near me. So now I'm behind on cities, tech, production, and culture but I've got a potent land army. Unfortunately the nearest other civ is hostile and has loads of units around me. Fittingly it's Darius and the Persians. I do have one military ally and a maritime ally though, so I will probably be able to deal with them.

But yeah, long story short, to get an army capable of challenging cities I needed to stunt most of my other areas of growth. Which makes sense, gameplay-wise. I think siege units proper are too cumbersome though. They take forever to get anywhere and their range isn't any better than a standard ranged unit. I think a better mechanic would be to allow units to build stationary siege units (and the build time would be based on unit strength). So the defender has to work on killing attackers as they build siege units. And it would take roughly as long as dragging siege units along does with less frustration.