Players are not always for win, or may not even care about winning since there can only be one winner, but rather random and emotional. The special rules are here to prevent the warmonger from snow balling too easily (which should be stopped with the whole world's effort). Consider the warmonger has one opportunity to expand through war by annexing the first victim civ, but further aggressive expansion should be more cautious. This is similar to SP, where aggressive play style tends to result in bad relation with AIs.
I take issue with the proposal of this rule, but before challenging it I really truly want to understand why it has been proposed, because I know there can't be any understanding until that happens. I have no understanding of why the idea of this rule seems good to you.
I hope this post is not seen as a wall of text, but as a carefully written invitation to an explanation. I hope the attention I gave to it shows my earnesty in actual dialogue with you.
So, in this part of the post, you just modeled ideal Human play on the behaviour of A.I.s. While the A.I.s do have a rationality that I know you see and many don't, you do know that in the end, the A.I.'s behaviour ought to be modeled on ideal Human play. That was backwards.
I recognize the statement that competitive play and casual play are different things, and keeping these camps separate is best for everyone. So compare the following two scenarios:
Players can play however they want within sporting conditions (not being abusive, playing to the end, possibly communication rules etc.), plus the vow to "try to win." The actions of any players, plus the static world conditions, combine to grant players who detect weaknesses and openings certain opportunities to advance on the world stage. Asymmetries that appear in strength are dealt with under the complete freedom of every other player, who continue in this way until the game rules declare an end.
Players play within sporting conditions, plus the vow to "try to win," and additionally obey a compulsion to oppose a "leader", defined in a computational sense, to the exclusion of any would-be conflicting game choice made for one's civ, whether based on that player's appraisal of competitive merit or otherwise. Every asymmetry has been preplotted, the exact time for this "joint opposition" to be triggered is established before any game begins, and players continue this way until the game rules declare an end.
Your view seems to be to prefer games played on the second model, not just that, but that having a league of games played of that sort is more constructive than a league of games played of the first. I don't get it. Firstly, I assume you do not propose to have players specifically do things not to their game advantage. Given that, surely you do not propose that any machine can tell when a broad change in play strategy policy is the the best strategic move. When it is absolutely required for victory? Maybe.
There is a point where it is obvious, yes. Suppose victory is a known number of moves away. Everyone will see that. But obviously, there are times when the move is not obvious. Either the victory is not an obvious number of moves away, or there is doubt about the reality of the victory being there. By 'not obvious', I mean some player might not see it.
Competitive strategy games are generally about the difference in players' ability to see the good moves. This is what I don't understand about the commonality of this kind of policy and thinking for Civilization league games. If all the moves players should do really can be plotted, why are the players playing it instead of machines? In my view, I learn things when I play games with people. And if they swear on their honour they are trying honestly to win, I learn things about strategy too.
How do you propose that you would ever already know what the players 'have' to do, but would want to play it out with free willed players anyway? What is their role in being there?