How the game selects planets

[AeoN2]

I have just finished a small experiment (which is similar to one i did earlier in one of my Posts)

To find out how the game selects planets I first created a large-grid game, omniscient, all systems seen, with no other opponents (which had the interesting by-effect of being my quickest game ever, winning before the game started )

Then I counted the number of systems which had planets (78 systems) and counted the number of huge gas giants in the game (90 total).

Then I went into secttype.txt and removed all but one huge gas giant of each atmosphere (each has 6 different in the secttype file) and started an identically made game.

When I counted the results were 70 systems with planets, and only 20 huge gas giants total.

What this implies is that when the game is asked to produce a Any/Any/Any planet in systemtypes.txt, it apparantly selects a random planet from secttype.txt that matches its criteria.

However, there is another statistical anomaly here, because the drop of the huge gas giants were only about one-fourth (when calculating for the number of systems), and not one-sixth as would be expected... That might just be coincidence though...

This means we can make mods that increase number of ice worlds (to make ice more usable)...

The best would ofcourse being able to read the source or let the author tell us how the routine works so we could know it in detail...

--
AeoN2
--
Shirley's Law:
Most people deserve each other.

[geoschmo]

I would guess that discrepancy is simply due to the small size of your sample. If you did several maps of each, the average would porbably be closer to what you were expecting to see, assuming you calculated the one sixth numnber correctly. I didn't try to check that one.

[Suicide Junkie]

Did you take into account the calls that are specifically for a gas giant, instead of Any?

[AeoN2]

There are no calls for gas giants specifically in systemtypes.txt

[geoschmo]

No, I think what SJ was thinking about was the specific calls for Huge planets.

Geoschmo

[AeoN2]

hmm... that is true, I haven't noticed that before... *sighs* well then that experiment was tainted too, but it might atleast explain why it wasn't closer to the expected number... so I'm pretty convinced by the theory...

The big question is if anybody cares? and if anyone actually want a planet/system mod...

I might try to get some time in the next few days to atleast summary what planned changes I would come with, and maybe get some input then...

EDIT: wohoo, finally someone rated me

[ June 20, 2002, 00:11: Message edited by: AeoN2 ]