Tribal Village outcomes

[Carrot]

This is my first attempt at Civ4 so I might have missed something here, but...

Is it just me, or are they not entirely random? In GOTM1 I can consistently recreate game starts, including huts, provided I take the same actions. Also, by playing slightly different I get other, reproducable results. As an example: moving the warrior SW, settling on the spot, selecting a worker as first build and researching Agriculture allways gives me 33g from the hut. If I start with warrior SE, all other equal, I allways get a map. What gives?

 

[Pfeffersack]

Yes it is this way...unless you check the option "new random seed on reload".It is just the way a omputer generates random numbers.tThey aren't really random, the are just calculated in a arbitrary and very sophisticated way, in which the same ingredients always produce the same results.

 

[Carrot]

Ah, so it's an option. That's peculiar since most rand functions use a seed based on a non-static variable, such as system time. This must mean that Civ4 uses a constant seed of some sort that appears to be based on game position.

Thank you!

 

[BeefontheBone]

The seed is fixed very deliberatley, in order to prevent the save scumming exploit (although they put an option in the last patch to reenable it).

 

[Glinka]

As an example: moving the warrior SW, settling on the spot, selecting a worker as first build and researching Agriculture allways gives me 33g from the hut. If I start with warrior SE, all other equal, I allways get a map. What gives?

I've never had that happen. Yes, if you reload a game from right before moving a unit onto a village, the outcome will be the same, but if you approach a different city from the same direction, you won't necessarily get identical results. In addition, if you wait a turn and move that unit over the village, the results should be different.

I remember one of Civ2's developers stating at one point that there are "good" and "bad" villages, meaning that the range of randomized results inclines more to the higher or lower end of the scale. The very best possible results for a good village would be a settler, worker or a tech. (In earlier Civs, the village could become one of your cities, but I'm glad they removed that in Civ 4, for strategic reasons.) On the lower end, you have maps or hostility. In addition, a scout will skew results higher than a warrior.

 

[Briareos]

Ah, so it's an option. That's peculiar since most rand functions use a seed based on a non-static variable, such as system time. This must mean that Civ4 uses a constant seed of some sort that appears to be based on game position.

Oh, Civ 4 definitely uses the system time or something similar as a seed, as otherwise you'd always get the same maps from the map generator. It just doesn't re-seed during the game unless you choose the new option added in v1.52 so the seed is saved within the savegame, and the same actions in the same order will always have the same outcome (i.e. the rand() function will produce the same pseudo-random value each time it's used in the code) - but if you shuffle your orders you'll of course get a different outcome.

np: Pass Into Silence - Iceblink (Pop Ambient 2006)