Originally posted by DrSpike
Careful though.........is your civ also random? I'm not sure how the cultural link works; if the first civ had a slightly higher than population probability of being American that could snowball with cultural linking on.
Can you be more specific on what you did?
Careful though.........is your civ also random? I'm not sure how the cultural link works; if the first civ had a slightly higher than population probability of being American that could snowball with cultural linking on.
Can you be more specific on what you did?
Here's what I did: launch a new game, standard map size, random barbs, random landmass, water coverage, age, climate, temperature, standard 7 AI opponents - all random, random human player, difficulty happened to be demi-god, default rules (which includes culturally linked starts *on*), and normal aggressiveness setting. When presented with the start spot, click F10, note how many civs, both human and AI, were American, hit ctrl-shft-Q, and select Quick Start from the menu options. Repeat.
Unfortunately, I did not differentiate between how many starts involved human American civs versus AI American civs -- I just noted the aggregate number of American civ appearances (i.e., a column looking something like 3, 2, 0, 1, 4, etc.).
Culturally linked starts obviously is a non-random influencer on random civ selection -- one could assume that the human civ is chosen randomly, and if true, an additional number of civs (between 1 and 7) are then chosen less than randomly. All of which led me to be cautious about only 600 total trials without explaining the reason why I was cautious -- is that 600 real trials (with an expectation that the non-random effects of culturally linked starts is spread roughly evenly among the American and non-American civ appearences) or is it really only 75 trials with a view that only the initial civ chosen is random and all others are determined in a less-than-random fashion thereafter? If it's a de facto 75 trials (or something significantly less than 600 trials), the divergence between actual and expected loses a lot of bite.
And then, of course, there's the point you make, that perhaps the human civ choice is not in fact random and that American civs appear as human-playable civs more often than expected -- culturally linked starts would definitely exacerbate the overall appearence of American civs if such were the case. (Purely anecdotal, but most complaining publicly about the distribution seem to be complaining about the frequency of facing American opponents as opposed to the frequency of being given an American civ). And if it my tests constituted only 75 trials, they were also far too small to support a view as to whether the initial civ choice for the human really is random.
In any event, I am inclined to believe that there is something slightly awry, but quite open to being shown otherwise by someone who wants to conduct a more extensive evaluation.
Catt
Comment