I've seen a poll somewhere that has rated the various Civs. I think the results pretty much agreed with how I feel on the subject, having played each one.
1. Civ II
2. Civ IV (with BTS)
3. Civ I
4. Civ V (with G&K, no rating out yet for BNW)
5. Civ III (with the various added packages)
Each game has obviously progressed in graphics and complexity. Even when though I liked Civ II better, I would tend to play Civ III more when it was current because I did like the look and feel of many new elements. But it's clear that some efforts to tweak game balance, among other things, were misguided.
Civ III introduced changes that created the Stacks of Doom issue that continued through IV. It also introduced a much more punitive system for wide vs. tall civs. In the name of game balance, it produced game play that was at once difficult and boring during much of the mid-history portion of the game.
Civ IV failed to correct some of these issues entirely, but was a huge advance with its specialist system and its intricate policies board. The resources system was also much-improved, and the religion system was at least interesting.
Civ V nixed the Stacks of Doom issue with the extreme innovation of one unit of a general type (one army, one boat, one worker/settler/general/religious unit) per hex. The new approach to the policy tree and culture is also interesting and creates new layers of strategy, but may result in a game that requires too much inside-game logic rather than real flexible strategy options. Much like Civ III, the mid-game often seems hard and boring. Certainly there are expert players who figure out how to master the system, but should one have to be a master player to succeed at and enjoy the game? It's still a fun game, but new thought is needed for Civ VI, which I predict we'll see by 2016 at the latest.
Take a look at years of game releases Civ I -- 1991, Civ II -- 1996, Civ III -- 2001, Civ IV, 2005, Civ V, 2010; that's 5 years for every game except Civ III, which was only out for 4 years. Why? Because marketers knew it sucked in comparison to the rest of the series and was damaging the overall brand!!! No word that Civ VI is in development, as they're still trying to salvage Civ V with the BNW expansion. But I wouldn't be shocked to see Civ VI as soon as late 2014 and most likely by sometime in 2015, certainly by sometime in 2016. You can tell something is different with Civ V when you compare strategy posts from Civ V to those for Civ IV. The volume is much lower, with a far smaller number of members participating. Civ V is successful by the conventions of most video game franchises and it's not a bad game. But it hasn't resulted in the passion and enthusiasm that the best games in the series -- Civ II, Civ IV and Civ I -- generated.
So what should Civ VI do differently? A few suggestions (note that I'm saying this without having played BNW, but have read the material and early reviews) --
1. Bring back some policy tree elements similar to Civ IV's while retaining some of the cultural growth strategy of Civ V. In a real world, civilizations mix and match their policies, as in Civ IV. The 21st century American civilization, for instance, includes aspects of every single policy tree, including Autocracy and Order. Make it more viable to select items from each tree with less cost for simply opening the tree. Insteaad of making the tree's absolutely exclusive, perhaps just the final policies or finishers. So you can be moderately orderly, autocratic (wrong name -- call it "Security,") and free, but you have to pass up on some of your Freedom if you want to be a Police State or Communist.
2. We need a happy medium between stacks of doom and the 1-unit limit. A five-unit limit with clear rules about who defends during an attack would be appropriate.
3. The game is too biased in favor of tall empires. The happiness, culture and money systems need to be tweaked to adjust this. Simply introducing some key modifiers in the policy trees isn't adequate, as it rewards or punishes civs for these choices too much. All policy tree options should have their pros and cons.
4. At the higher difficulty levels, very aggressive NPC's tend to make less experienced players' miserable and/or dead very early. This has resulted in all but the best players hovering around the mid-level Prince and King settings. Tweak a few settings in the early game so that even at the highest levels the early game is mostly survivable and only becomes truly brutal once the player has gotten a clear foothold. This will encourage more players to tackle the higher levels and give them an overall better experience.