Do you want random events?

[johnmcd]

Why would a natural disaster on crops encourage research on agriculture any more than having the crops in the first place?

If the game introduces too many factors to be factored into good decision making quickly it starts to play itself to a great degree. The human hand stops controlling and starts guiding. Now that might actually be quite good fun (and I'd welcome something akin to a spectator mode where human intervention was slight) but it would change the game out of all recognition.

I just can't get past the issue that if events are important to success or failure then their randomness mean victory or defeat is not necessarily contingent on how you use your starting position. If they aren't important to success or failure then there are more interesting ways to add 'colour' to the game, palace, newspapers, info on who your great people are etc.

[xenophile58]

Most random events in the real world are, while disastrous for those affected, limited in scope. Very few events are actually beneficial to those affected - but bystanders often benefit from a sudden comparative advantage as their rivals lose ground. Provided the game incorporated disasters that reflected those characteristics I think they would be interesting and challenging. If city improvements did mitigate the effects then that just adds another consideration to those we have.
My main reason for liking the idea of adding more but limited chance into the game would be to reduce the reliability of narrow strategies. CS Slingshot? Honestly, however smart (and it is smart) that's pretty gamey, isn't it? The chance of your focussed timing being thrown off by losing a couple of citizens to starvation because the city hadn't a granary might encourage more robust approaches. Provided the mechanism wasn't skewed towards the AI, the effect actually might to be move the game balance a little the player's way.