UPDATE frequencies on broken worlds in MGT generation

Rikki Tikki Traveller said:
Captain, I don't think I have a zero tolerance to anomolies, although I will freely admit that it is lower than yours (and maybe most other peoples too...).

well, how bout a "tolerance not statistically different from zero" ?

Yes, that was a stat joke. Sad, ennit ?

Good points, though. Its important to remember that the flavor and feel of the method is crucial, too. Starting to generate goofball worlds (technical term for small rockball with thick atm) would concern me , too....

I will suggest that it's becoming plain to me that the OTU worlds are not what worldgen was meant to create; those, by this time, are created by the writers...by design or default.
 
Now where was I ? Right, in other threads and stuff.

I suppose the best way to get back on topic is to summarize.

The issue was, many types of worlds are considered broken, and occur too frequently for people's taste. But how often do they actually occur ?

Common wisdom points at several examples as way too frequent.

Thick ATM, small SIZ (world: Impossible !)
Hi pop and horrible atm. (suffocating and hostile hellholes)
Hi grade starport, low population &/or tech (potemkin and phantom ports)
unpopulated garden worlds (empty lots in manhattan)

One crystal clear point shines thru on this board: we all have our own ideas for what too much is.

So here are some numbers as to how frequent those anaomalies are, as a basis for personal decision.
Now remember, common wisdom (often espoused by non traveller gamers), is that these happen "all the time" or, at best, way too often.

Also, remember that I am presenting these numbers, not stating that they are okay, too much, acceptable, unacceptable whatever.

And the counts (for classic Traveller worldgen) are:
Code:
Impossible ATM worlds : 26% 
Overpopulated hellholes          
    of which are suffocating (A & 3-)  and pop 9+   2.7%
    of which are hostile (B+) and pop 9+            0.8%
        
    of which are suffocating (A & 3-)  and pop 6+   15.7%
    of which are hostile (B+) and pop 6+             2.2%

9+ was chosen as one cut point, as it defnes a hi population planet
6+ as a second cutpoint, as it is the first digit on the upper side of the 
median. (roughly meaning "more populous than average")

One important notes: these numbers [i]include[/i] the low tech worlds that would be actually be zeroed  by strict CT rules.

Potemkin Starports (TECH 6-, port B+) 12%
(I'll check this one - I cant remember if tis accounts for the posrtxtech dependency; if not, this is the maximum.)

Phantom Starports  (POP 0 & A)      0.2 %
              (POP 0 & B)           0.6%
              (POP 0 & A-D)         2.4%

Porkbarrel Starports:(pop 3-, Port B+)  7.8%
              of which are type A       2.3%   
Garden planets within jump 3 of a crap world         20%

Note.  this is the probability of a garden world having a crap world within j3.
See the earlier post for details on this one. The final number depends on the pop you use to define empty.  If it's simply below median, one gets 10%. If its 0, one gets .6%

Tech too low to 
     survive atmosphere   by CT rules, 0.0%
     (sub minimumtech worlds are zeroed, or raised to the minimum by lenient GMs


What these mean.

A subsector on average has 40 planets
a quadrant 160
a sector 640

Multiplying the percent from the table by the nmber of planets in your chosen group gives one the average number of exaples one expects to encounter. Note that this average need not (and will generally not) actually be a possible result. If the average is 4.5, it means, roughly 4 or 5 examples are expected.
One has to remember that the average includes all the results up to the total...thus, any given sector with a 4.5 expected may have 4 or 5, but it certainly can have 0,1,2,6,8,&10. The liklihood of a non-average result decreases as the group grows larger; at lower levels the average is increasingly unstable. I note that all three levels above will show notable instability (ie how well the mean describes the data)

Wow, that was a boring way of saying, "who knows what the average means in relation to what I'll get ?"

So I like computing the probability that a given result will be found one or more times. The math involves power calculations , so close your eyes if square roots bother you....

(1 - %/100) times itself as many times as you have elements you are checking.

Thus, a 3% prob across a subsector computes to (1-.03)^40 = a 70% chance of one or more....or, if you like, a 30% chance of zero occurances.

to make it a bit easier:
Code:
%      chance of 1+ per subsector
5%      87%
3%      70%
2%      55%
1%      34% 
0.5%    19%
0.3%    11%
0.2%     8%
0.1%     4%
0.05%    2%
0.03%    1%
0.01 % 0.4%
 
Back
Top