Revolutions and the early government civics punish you for overexpanding early, but the current rubber band mechanisms don't seem to me to be quite so effective once you start dominating the AI (for me I feel this is when I can grab a tech anytime I see an AI try to get it-usually to rob them of a religion or wonders). There are some ideas for disasters floating around that may do this once implemented, but I wanted to suggest something more mundane that occurred to me when I started running away with my last game.
The basic idea here is that once a civ gets far enough ahead to 'feel' ahead, it's people get soft and start taking it all for granted, forgetting the hard work it took to get there and, as a consequence, shirking the effort required to remain there.
Basically, once a civ has some lead established (say, 20% lead for 160 turns-which seems to be a magic number on Eternity), the Complacency effect would kick in. Each following turn, the Complacency Counter would increase proportional to the lead (guestimating in my head: {[lead score/second place score] - 1} x .01). The Complacency Counter would be applied as a penalty increasing all costs (maintenance, build, research, etc.) and at certain thresholds increasing unhappiness and unrest (say + 1 unhappy per .2 in all cities, and something similar for national unrest though I don't understand that enough to suggest specifics). The lead could be versus some average of leading competition (maybe the top fifth or something like that) alternatively.
Eventually this should pull the leader back to the pack, which is where complacency could suddenly turn to panic or resolve. Once a complacent civ falls to second (or worse) in score, two things happen. Some Complacency is reduced (this should be fairly rapid, maybe [complacent civ score/leader score]-1) each turn. Or the civ may panic and fall apart. The Panic chance is 1-1/(Complacency Counter + 1) [So it starts higher the longer and more dominant a lead you had]. This is checked each turn until either the Counter is reduced to 0 or the civ returns to first place. If a check fails, an additional check is made against every city (reduced by culture levels and ignored by capitals and similar maintenance distance reducing buildings) and every unit with additional failures assigning those objects to a new civ or the one resulting from a previous panic (pretty much the way revolutions seems to handle it but more catastrophic), with each such loss treated as a loss in war for whatever effects those have (weariness and unrest in remaining cities).
To keep the overall game winnable, the contributions to Complacency should be reduced in some way for each such collapse that has already occurred in the game.
The basic idea here is that once a civ gets far enough ahead to 'feel' ahead, it's people get soft and start taking it all for granted, forgetting the hard work it took to get there and, as a consequence, shirking the effort required to remain there.
Basically, once a civ has some lead established (say, 20% lead for 160 turns-which seems to be a magic number on Eternity), the Complacency effect would kick in. Each following turn, the Complacency Counter would increase proportional to the lead (guestimating in my head: {[lead score/second place score] - 1} x .01). The Complacency Counter would be applied as a penalty increasing all costs (maintenance, build, research, etc.) and at certain thresholds increasing unhappiness and unrest (say + 1 unhappy per .2 in all cities, and something similar for national unrest though I don't understand that enough to suggest specifics). The lead could be versus some average of leading competition (maybe the top fifth or something like that) alternatively.
Eventually this should pull the leader back to the pack, which is where complacency could suddenly turn to panic or resolve. Once a complacent civ falls to second (or worse) in score, two things happen. Some Complacency is reduced (this should be fairly rapid, maybe [complacent civ score/leader score]-1) each turn. Or the civ may panic and fall apart. The Panic chance is 1-1/(Complacency Counter + 1) [So it starts higher the longer and more dominant a lead you had]. This is checked each turn until either the Counter is reduced to 0 or the civ returns to first place. If a check fails, an additional check is made against every city (reduced by culture levels and ignored by capitals and similar maintenance distance reducing buildings) and every unit with additional failures assigning those objects to a new civ or the one resulting from a previous panic (pretty much the way revolutions seems to handle it but more catastrophic), with each such loss treated as a loss in war for whatever effects those have (weariness and unrest in remaining cities).
To keep the overall game winnable, the contributions to Complacency should be reduced in some way for each such collapse that has already occurred in the game.
