[Space] How habitable is my universe?

[Agemegos]

This project is in abeyance

[thrash]

First response is, what happened to Habitability 7-8? Did you really not have even 0.1% of systems with worlds of Standard or Dense (but not Marginal) atmosphere, hydrographic 60-90%, and Chilly, Cool, Normal, Warm, or Tropical climate? Given that you report 10.5% with Habitability 2-6, I find this hard to believe. Earth is Habitability 8, not 7 -- did you miss the +1 for non-Marginal atmospheres?

Second answer is, look at the Affinity Modifiers Table on p. 91. RVM averages to 0, so on average Affinity is equal to Habitability. All other things being equal, a world with Habitability 8 (like Earth) can support ~62.5 times as many people as one with Habitability 2. You could interpret this several ways, depending on the specifics:

* Only 1.6% as much land is actually habitable (mountain-tops, polar islands, microclimates)

* Only 1.6% of the potential population is able to adapt to conditions, either physically or culturally

* The land surface as a whole is only 1.6% as productive.

The actual answer is probably some combination which results in 1.6%, e.g. 10% of the land is habitable, 40% of the population can adapt, but the land is only 40% as productive.

[TheDS]

He only ran it 200 times; 0.5% is the increment for that.

Let me also suggest to Agemegos: If you can arrange your data into a more manageable format - when I did a sheet for First In world and system generation, I made it a point to put all my system data (minus world and moon data) into a single column, allowing me to easily grab the data and copy it where I wanted, or even generate numerous systems at once by grabbing the handle and dragging it over until I had as many columns as I wanted.

So my advice to you is, put your data into a single column (or row), then select it all and do a massive copy and paste (should take about 3 seconds). Then you can generate data for up to about 250 systems at a time (or 65000 if you go for rows). Your final columns (or rows) will be to count up each result. A simple average is easy to find, but if you wanna know the numbers, you'll have to use something like the COUNTIF function, and may need numerous additional columns.

But with judicious use of copy/paste and some thought about how to count up the results, you could probably generate 10,000 systems at a time (for data-mining purposes, as you did above) with a LOT less total effort and time spent.

[TheDS]

I am also wondering why there wsn't more of a curve in the results, with low-hab worlds being more numerous, but I suppose you may have manually adjusted the curve so we weren't wasting so much time on crap worlds. (I don't have the new Space yet.)

[thrash]

Originally Posted by Agemegos

The sample was only 200. Those figures are good to 0.5%, not 0.1%. And even that is subject to sampling deviation.

Need to pay more attention, obviously. I tend to approach the problem analytically, by working through the probabilities, rather than empirically.

What are your own results?

http://www.io.com/~thrash/firstin.html

The procedure in GT: First In is conceptually similar to GURPS Space 4/e. The key observation is that the definition of "habitable" is based on Earth, so each characteristic forms a distribution around that central value. It may be a tailed distribution (e.g., because atmospheric density is correlated with size), but over a large enough sample the combination should be a bell curve with a peak in the mid-range (maybe 6-7?).

I understand that. I'm trying to get a qualitative feel of how forbidding conditions are. When it comes to generating my universe I will need to know which planets people will go to and which are just too unattractive, given that travel is expensive and one-way... I don't think people would go to settle a 1.

Ideal population distribution is a function of diffusion and affinity: people spread out as far as possible, consistent with the resources available.

The simplest method is to take the figures at face value, calculate carrying capacity directly, divide by transportation costs, and distribute population (or at least, initial colonies) proportionately. There are very few places on Earth so harsh that no one has ever settled there; even places like Antarctica might have been settled by Inuit-analogs, if (e.g.) the process of Polynesian expansion had continued for a few thousand years. Habitable planets are similar: that breathable atmosphere is a huge advantage.

[balzacq]

I have a PHP script to generate systems as per GURPS Space, with a very few minor tweaks (Agemegos' revised tidal formulae and aging the system in 0.1Gy blocks to track changes over stellar evolutionary timescales). I use the chart in G:S to generate stellar types instead of real-universe data. Also, I use all the rules except for brown dwarf and flare star generation.

I've generated a total of 16977 systems this way (I have a perl script that creates whole thousand-system sectors at once). The habitability breakdown is as follows:

Hab.   Systems    %
8  	136  	0.80
7 	300 	1.77
6 	257 	1.51
5 	338 	1.99
4 	303 	1.78
3 	180 	1.06
2 	105 	0.62
1 	28 	0.16
0 	15317 	90.22
NULL 	13 	0.08

(Without diving into the tables right now, I suspect that the NULL values are from systems with nothing but gas giants and belts.)

I'm at work and without my books, but I was under the impression that you didn't even generate carrying capacity for Hab 0 planets -- the assumption was that Hab 0 systems would only ever have outpost-type residents (miners, military, etc.).

[balzacq]

Originally Posted by Agemegos

The Affinity Modifiers Table on p.91 gives carrying capacity multipliers for affinities down to -5. And the maximum RVM is -5 (-3 for worlds that are not asteroid belts), implying non-zero carrying capacity on worlds with habitability -2.

I am perhaps confusing the population rules in Space with those in GT: Interstellar Wars.

[thrash]

Originally Posted by Agemegos

So do I. But my capacity to do this fails when faced with a massively contingent system like this.

Which is why I haven't done it, either.

It's the Anna Karenina Principle. There is only one way to be spot on the ideal, but as for being off by one category in one dimension, there are twice as many possibilities as dimensions...

Some sort of extreme-value distribution, maybe? The thing is that we aren't looking as a measure but an index.

Yes and no. The range for each variable is already constrained because we're necessarily talking about habitable worlds. As long as the bins are wide enough, the combinations will still result in a jackpot fairly frequently.

But what we have got takes that metric and chops it range up into a bunch of unequal-width bands, with the central peak somewhere in the category called 'habitability 0'.

I don't think so, unless I missed something important.

What we have is two categories -- hostile (habitability 0) and habitable (habitability >0) -- with a probability function to choose between them, e.g. f(0) = 0.9. Within the habitable category, there is a roughly bell-shaped distribution of habitability with a central peak defined by how wide are the bins (i.e., how likely are the optimum results) in each variable.

[thrash]

Originally Posted by Agemegos

I'm not necessarily talking about habitable planets. A lot of the time the highest 'habitability' score in a system comes from a planet with suffocating atmosphere, no water, and unendurable temperatures.

Hmmm. On review, you're right. This is a significant departure from earlier uses of "habitability," and a flaw in the system in my opinion. I don't see the point of discussing or measuring the "habitability" of any world which cannot support human life in the open somewhere on its surface. Sealed habitats and life support are pretty much the same, no matter where one goes, and it's pointless to assign them a "natural" carrying capacity.

As for worlds with Toxic or Corrosive atmospheres, the answer is that one would not "colonize" them at all, when it would be simpler and safer to construct an orbital habitat or one on a nearby asteroid or moon. If there were valuable resources to exploit or interesting aliens to contact, one would build an outpost instead, which does not depend on carrying capacity (p. 93), and support it from a colony somewhere more convenient.

[Anthony]

Originally Posted by thrash

Hmmm. On review, you're right. This is a significant departure from earlier uses of "habitability," and a flaw in the system in my opinion. I don't see the point of discussing or measuring the "habitability" of any world which cannot support human life in the open somewhere on its surface. Sealed habitats and life support are pretty much the same, no matter where one goes, and it's pointless to assign them a "natural" carrying capacity.

It's not totally pointless; which resources and hazards are present in the environment makes a big deal on how challenging life support is.