Home » Stars! 2.6/7 » The Academy » Hab correlation
Hab correlation |
Tue, 28 March 2006 06:51 |
|
PricklyPea | | Lieutenant | Messages: 534
Registered: February 2005 | |
|
posting this from my cell phone on the ski slopes of switzerland!
just wondering: is there any link between different hab environments. e.g. if planet has high temp is it more likely to have high grav?
i remember creating a race which had some habs high and others low and had probs finding greens.
or was this my imagination?
Report message to a moderator
|
|
| | | | | | | |
Re: Hab correlation |
Tue, 28 March 2006 23:31 |
|
LEit | | Lt. Commander | Messages: 879
Registered: April 2003 Location: CT | |
|
yoey wrote on Tue, 28 March 2006 21:25 | I am not sure if this is linked to either of these 2 topics, but why does the race wizard give you more points for moving the bars to the right as opposed to moving them to the left?
|
I've been looking at ConstB's reverse engineering of the RW hab point costs, and I'm pretty sure it was a bug. There is a bit of code that tries to see how habitible the universe is for your hab settings, and charge points accordingly, I think it's got an endpoint problem. However, they either didn't think it made a big difference, or liked the effect. I kind of like the effect, there is a RW advantage to going high in habs, and therefore probably a game play advantage in going low.
- LEitReport message to a moderator
|
|
| | | | | |
Re: Hab correlation |
Wed, 29 March 2006 08:08 |
|
|
Hi!
I did not see any correlation between different habs.
I got, however, really a lot of times "clustered" picture of my empire (Galaxy Clumping was off). It means that for narrow hab, I got green planets in one area and no a single green planet in another - something like a "clusters" of green planets. Usually I get this in the medium or large universes, and when taking habs shifted to right edge with one field immune. So, if you like a solid clusters of green planets in small territory to minimize transporting between them and mine laying to defend them, it seems better to take all higher hab ranges.
BTW, does anybody knows a reason why Race Wizard gives more points when you shift narrow range to right compare to shifting to left? It does not looks like shifting to left gives more planets than shifting to right...
WBR, Vlad
Report message to a moderator
|
|
| | |
Re: Hab correlation |
Wed, 29 March 2006 08:23 |
|
mazda | | Lieutenant | Messages: 655
Registered: April 2003 Location: Reading, UK | |
|
Tomasoid wrote on Wed, 29 March 2006 14:08 | I got, however, really a lot of times "clustered" picture of my empire (Galaxy Clumping was off). It means that for narrow hab, I got green planets in one area and no a single green planet in another - something like a "clusters" of green planets. Usually I get this in the medium or large universes, and when taking habs shifted to right edge with one field immune. So, if you like a solid clusters of green planets in small territory to minimize transporting between them and mine laying to defend them, it seems better to take all higher hab ranges.
|
I don't think this is restricted to higher hab ranges.
In my experience any small habset will tend to get planets that are grouped together rather than randomly spread throughout the galaxy.
Again, it would seem like a lot of hard work to generate the galaxy in this manner (for little gameplay benefit).
So maybe it is just a visual statistical illusion.
Report message to a moderator
|
|
|
Re: Hab correlation |
Wed, 29 March 2006 08:33 |
|
|
mazda wrote on Wed, 29 March 2006 16:23 |
Tomasoid wrote on Wed, 29 March 2006 14:08 | I got, however, really a lot of times "clustered" picture of my empire (Galaxy Clumping was off). It means that for narrow hab, I got green planets in one area and no a single green planet in another - something like a "clusters" of green planets. Usually I get this in the medium or large universes, and when taking habs shifted to right edge with one field immune. So, if you like a solid clusters of green planets in small territory to minimize transporting between them and mine laying to defend them, it seems better to take all higher hab ranges.
|
I don't think this is restricted to higher hab ranges.
In my experience any small habset will tend to get planets that are grouped together rather than randomly spread throughout the galaxy.
Again, it would seem like a lot of hard work to generate the galaxy in this manner (for little gameplay benefit).
So maybe it is just a visual statistical illusion.
|
You may be right. I'm using high narrow hab ranges most of the time, so it only looks like that for me. However, in some games I get clear equal distribution throughout the space, and in other games I get clear clustering. So it certainly depends on something. It may be just another random number generator correlations pattern, and in such case it depends only on the game starting conditions. It would be good to know for sure though, and also would be good to know how often you can expect to get clustered empire vs equal spread through space.
WBR, Vlad
Report message to a moderator
|
|
|
Re: Hab correlation |
Sun, 02 April 2006 23:11 |
|
|
OK, some results.
Alan Kolaga was kind enough to share the scripts and data he used to create the graphs showing the relationships between habs and mineral concentrations.
Once I understood the fundamentals of how his scripts and worksheets accessed and presented the data, then I made up my own set so I could compare how the habs interact with each other.
I graphed the distribution of Temp over Rad, Grav over Rad, and Temp over Grav. The graphs show that there is no correlation between any pair of habs.
I used a sample of 1,000,000 planets from Alan's data file of 10,000,000 planets. When I go home from work I'll leave my office computer generating a completely fresh set of 10,000,000 planets and graphing these, just to be sure.
Report message to a moderator
|
|
|
Re: Hab correlation |
Sun, 02 April 2006 23:16 |
|
|
BTW, the planets are generated from random Large Packed universes, with a tri-immune JOAT.
I suppose I could generate a set of planets from Tiny Sparse universes to see if this makes any difference, but I seriously doubt it and it would take far longer to generate the worlds.
I next plan to generate data files using a couple of narrow hab races, to confirm whether or not race design has any impact on the hab distribution in the universe (I seriously doubt that it does.)
Report message to a moderator
|
|
|
Re: Hab correlation |
Mon, 03 April 2006 01:11 |
|
|
Dogthinkers wrote on Mon, 03 April 2006 13:16 | I next plan to generate data files using a couple of narrow hab races, to confirm whether or not race design has any impact on the hab distribution in the universe (I seriously doubt that it does.)
|
Ran a quick test with OWW habs. I only used a 200,000 planet count sample, but it seems clear enough from this that race designs do not impact the habs of non-HW planets. The only obvious difference between this and the 3-imm data was a small spike at the exact HW habs, of about 0.1%. Which is exactly what you would expect when you consider that the HW made up about 0.1% (200 planets) of the sample. I'll rerun this test with a few million planets in a couple of days, when the computer is idle.
Report message to a moderator
|
|
|
Re: Hab correlation |
Mon, 03 April 2006 20:20 |
|
|
Dogthinkers wrote on Mon, 03 April 2006 13:11 | I graphed the distribution of Temp over Rad, Grav over Rad, and Temp over Grav. The graphs show that there is no correlation between any pair of habs.
I used a sample of 1,000,000 planets from Alan's data file of 10,000,000 planets. When I go home from work I'll leave my office computer generating a completely fresh set of 10,000,000 planets and graphing these, just to be sure.
|
10,000,000 planet sample confirmed the result. Each hab value has has no effect whatsoever on the others.
As a seperate issue, I can also confirm that the mineral concentration advantage seen at extreme high rad is NOT matched by an advantage at the extreme low rads (the graphs on the website listed above do not discriminate between them.) Over a sample of 1,000,000 stars (non-ACCBSS,) rads of 90+ had an average concentration of 61.2 in each mineral, compared to an average of 57.2 for normal rad values and an average of 57.2 for extreme low rad values. So you not only get more points by taking rad high rather than low, you get noticably more minerals too. Of course if everybody uses this there isn't much room left for hab sharing
Report message to a moderator
|
|
| |
Re: Hab correlation |
Tue, 04 April 2006 18:17 |
|
|
[email | m.a@stars[/email] wrote on Tue, 04 April 2006 19:07]Outstanding work!
|
Alan Kolaga did the hard bit. All I did was understand what he did years ago, then apply the same techniques to find out what I wanted.
Quote: |
Quote: | As a seperate issue, I can also confirm that the mineral concentration advantage seen at extreme high rad is NOT matched by an advantage at the extreme low rads
|
Any idea why the universe-generating code would have such a quirk?
|
No idea. Although after I told him, Alan said he is considering hiring bodyguards to protect himself against all the players over the years that took races with extreme *low* rad thinking they were getting a mineral concentrations advantage
Quote: | Perhaps it could be interesting to (ab)use the option for some scenario games, if it could be tied to "special" hab settings and not just hi-Rad.
|
I think there are a few people around capable of customising the stars in any given universe. Making an identical change to all stars is obviously easiest (i.e. all planets ideal for OWW, all planets are HW,) but there's no reason something more complex couldn't be done. I daresay you'd want something 'scriptable' though.
(i.e. It shoulds plausible that you could make universe where the habs were inter-related (i.e. high rad makes grav more likely etc.) You'd just make a new algorithm for determining habs, then generate a random set of stars, then apply that data as a block to replace the hab information in a universe.)
[Updated on: Tue, 04 April 2006 18:18] Report message to a moderator
|
|
|
Goto Forum:
Current Time: Mon May 13 01:51:59 EDT 2024
|