UPDATE frequencies on broken worlds in MGT generation

[Rikki Tikki Traveller]

Well, from my point of view 2% is still too much.

A typical sector has 650-ish worlds, so 2% represents about 12-13 worlds. IN MY OPINION that is still too many to have to handwave away. And that is only 1 type of anomoly.

The biggest problem to me is the published sectors (Spinward Marches and Solomani Rim). They were hand made to some extent and seem to have an even higher rate of weird/unrealistic worlds.

Yes, I would include low population "garden" worlds as a "weird" category.

Say POP 3- and ATM 4-9 to give you search criteria.

 

[EDG]

I was hoping to stay away from subjective (or logical) discussions of the actual numbers here. As I finish what I was wanting to present, I'll move the info to its own thread, so we can have some harder observations to discuss the nature and frequency of nonsenseland.

Well I've already presented hard numbers for what I'm defining as "broken worlds" here. You'll have to be clear about what definition of "broken worlds" you're using in the other thread, because if you're not using the definitions I used here then obviously you're going to get different results.

What I will say, is that I'm surprised at the differences between perceived frequency, and actual frequency....both for others , and for myself....and in both directions.

Without being specific, that's pretty meaningless . What sort of differences are you talking about?

 

[captainjack23]

I was hoping to stay away from subjective (or logical) discussions of the actual numbers here. As I finish what I was wanting to present, I'll move the info to its own thread, so we can have some harder observations to discuss the nature and frequency of nonsenseland.

Well I've already presented hard numbers for what I'm defining as "broken worlds" here. You'll have to be clear about what definition of "broken worlds" you're using in the other thread, because if you're not using the definitions I used here then obviously you're going to get different results.

I'm pretty sure I've posted the criterea in each case - you may want to look at the primary posts again.

What I will say, is that I'm surprised at the differences between perceived frequency, and actual frequency....both for others , and for myself....and in both directions.

Without being specific, that's pretty meaningless . What sort of differences are you talking about?

Some situations occur more often than I expected or imagnined would be the case, some less. I'll enumerate them in another thread eventually - for now, I'm trying to avoid making specific statements about my preference or opinions of the numbers.
Again, this is discussed in the primary posts....and is why I think reposting it all together will help clarity for the reader - most of us tend to pay most attention to the bottom of the thread.

 

[captainjack23]

BTW, the primary posts have been extensively updated as of 08-02-08 (DMY)

 

[captainjack23]

Just for comparison...

For the 250k CT run I've got 67128 worlds (26.85%) that have atmospheres that are impossible in the habitable zone (size 1-4, atm 2-9). Of those, 23078 worlds (9.23% of the 250k) are size 1 or 2 with atm 1+ (also impossible in the habitable zone).

Quick clarification: you needed to assume that worlds were in the hab zone for your output, sets correct ?

Given no data on hab zones, is it fair to simplify the impossible range to 1-4 size and 2+ ATM (excluding your A category, of course)
<edit>
never mind. I found your criterea for worldgen. Don;t know why I didn;t look there first.DUUUHHHHH.

 

[captainjack23]

Well, from my point of view 2% is still too much.

A typical sector has 650-ish worlds, so 2% represents about 12-13 worlds. IN MY OPINION that is still too many to have to handwave away. And that is only 1 type of anomoly.

Sorry, seem to have missed this post !

Without commenting pro or con, I'd say that yes, 2% is 12-13 worlds per sector, but pretty well spread out. The explaination could be a handwave (Ancients ! Robots! Magic !) or somthing a bit more systematic and consistent: Coaling station, prison planets, etc. Obviously it depends on the type of anomaly, and the type of explanation. In my experience, a handwave is a solution that essentially says: don't question, just accept; it does nothing to deal with the issue of stressing willful suspension of disbelief.

Also in my experience, the difference between a handwave and a reasonable explanation is just that- it gives some coherent and consistent explaination that isn't just "cope", and that seems subjectively about as likely as other reasons in the story.

Generally simple is good, but consistent is better, and frequency has to be considered. As a hypothetical example, "those worlds with a UWP code of XYZZYX , while they are odd when compared with most worlds, are in fact secret research bases" is fine for one (or two, perhaps) but a bit far fetched for 12-13 examples - but "they're prison planets with a roughly one subsector catchment area" may well make more sense.

he biggest problem to me is the published sectors (Spinward Marches and Solomani Rim). They were hand made to some extent and seem to have an even higher rate of weird/unrealistic worlds.

Agreed. No idea how best to deal with that, though. And, as I've mentioned, I am beginning to strongly suspect tha a good chunk of prevailing opinion about worldgen is actually based on writer whimsy in those publications.

 

[captainjack23]

Yes, I would include low population "garden" worlds as a "weird" category.

Say POP 3- and ATM 4-9 to give you search criteria.

Damn, not at at work. Luckily that's not too hard to run by hand to get probabilities, if not actual counts.

Pop 3- = unmodified 5 or less on 2d6: 1+2+3+4 or 10/36
ATM 4-9 is a bit trickier due to the dependency on SIZ, but I'm going to ballpark it at 60% since it straddles the mean and a bit more on one side - it's max value will be (5+6+5+4+3+2) =25/36...and since I think more SIZ/ATM rolls than not can result in in a 4-9, I say that will knock it down by no more than 50%.

So.....25/36 *1/2 = 25/72, call it 1/3 .

10/36 * 1/3 ~~ 9% conditional prob.

Now if the main anomaly is simply the profile, done. (which is to say, the anomalous question is "In a civilized and developed area, with people living on crap planets, why the hell don't more people live there")

call it 4/subsector on the average, and one certainly will roll up at least one, likely at least two in an given subsector.

If the issue is of being right close to (say) a hi pop hellhole world.....a bit trickier. assume a max of J-2 separation for the sake of ease. Start with the 4 garden worlds, we'll get 1/10 hexes., plenty of room to avoid overlap and complication. So we can worry about the single case -start with one vacant garden.

It has 6+12 hexes around it which qualify.
a simple hi pop hellhole of any type is a good match to define them as an anomalous pair. (ie Why the hell don't more people live there instead of there ?) Luckily, I have the stats for that above: hellholes, all atm types, hi pop = 4%.

So...if the pair is what you're interested in, the basic probability of at least on hellhole and one garden within 2 hexes it is ROUGHLY the probability of having at least one hellhole within the 18 surrounding J2 hexes of any given empty garden * the liklihood of at least one empty garden in that area. . ( This is an lower estimate, as for sanities sake, it ignores the possibility of of multiple examples occurring of either in the defined space).

The ugly math is thus:

(((1 - ((1- .04) ^18 )) * (1- ((1-.09) ^17)) ) / 2

The 17 is because you can't have two stars in the same space, and once you have the low value member, you go for the high value to maximize the liklihood estimate; the * 1/2 is because you can arrange two items in two ways, but in this case, order doesn't matter -so they both are equiv

I leave the actual calculation as an exercise, unless I brought my real calculator home.

Not much help, I admit........

[edit: Okay, the answer is .208]

Please remember THE NUMBERS FOR THE GARDEN WORLD ARE ESTIMATED BY REAL PROBABILITY TIMES BALLPARK PROBABILITY. So caveat emptor...and if someone runs it, and gets different, Bully ! Use that.

 

[EDG]

Agreed. No idea how best to deal with that, though. And, as I've mentioned, I am beginning to strongly suspect tha a good chunk of prevailing opinion about worldgen is actually based on writer whimsy in those publications.

That's why I made a big deal about how people need to go by the CT results when saying "this is what the OTU should be like", because the published sectors don't actually conform to the results from those rules. In fact, if you're right strict about it, the OTU should really look like it was generated using Book 6's full star system generation rules, which means you'd have to account for star type as well since worlds in the habitable zones of M V stars are (in CT) necessarily smaller due to negative DMs to size because of orbit number.

i.e. if you have a habitable world around an M V star, it's in Orbit 0 at most - orbit 0 has a DM-5 to size, and there's an extra DM-2 to size because it's an M V primary, so the maximum size that world can be is 3. So any world generated around an M V primary can't be bigger than that - and if you apply the small worlds rules here to that, then obviously M V stars can't have habitable planets, because they're all size 3 or less. And these are the most common stars that can be rolled in CT, so it'd REALLY change what people think the OTU should look like... but this is all right there in black and white in the official rules.

And let's not mention that all the orbit zone tables in CT Book 6 are incorrectly calculated - the reality is that not every star has a habitable zone that falls conveniently within those numbered orbits that CT uses (for M V stars, only M0-M2 V stars have a habitable zone at Orbit 0 - the innermost that Traveller allows). The habitable zone for M3-M9 V stars is within Orbit 0, so anything placed in Orbit 0 for those stars is going to be in the Middle or Outer Zones, and thus not habitable anyway).

 

[EDG]

Yes, I would include low population "garden" worlds as a "weird" category.

Say POP 3- and ATM 4-9 to give you search criteria.

In the EDG 1.1 worldgen, 29843 (11.9%) of worlds fit those criteria (pop 0-3, atm 4-9). 5972 (2.4%) are atm 4-9 and pop 0 - most likely those are they're tainted atmospheres and aren't Temperate.

6696 worlds (2.7%) are atm 5,6 or 8 and pop 0-3. Some of those are pop 3 and Temperate (the min pop for such worlds is 3). Many are pop 0-2 because the temperature is Frozen/Cold/Hot/Roasting (only 144 worlds (0.05%) are atm 5/6/8 and pop 0, and they're all Frozen/Roasting).

I'm OK with these worlds - the worlds that are Temperate and have untainted atmospheres could easily just be new colonies that haven't had a chance to grow yet. As for the rest, the tainted atmospheres and/or the temperature is making the worlds a lot less attractive to people despite being broadly habitable otherwise. So I don't consider these to be "broken worlds" - in the EDG system the reason they're low population makes sense (it's because of the tainted atmospheres and/or temperature), and it's not like they're a huge percentage anyway.


For comparison, in the MGT run, 38413 (15.4%) of worlds are Pop 0-3 and Atm 4-9. 5972 (2.4%) are atm 4-9 and pop 0 - most likely those are they're tainted atmospheres and aren't Temperate. 3885 worlds (1.5%) are Pop 0 and atm 4-9. But 20326 worlds (8.1%) are atm 5/6/8 and pop 0-3 - there are a lot more low pop true garden worlds in MGT than EDG (2078 (0.83%) of the atm 5/6/8 worlds are pop 0). But in MGT there's no correlation between population, atmosphere type, and temperature, so the distribution of these low pop garden worlds doesn't have an explanation that makes sense - so you get some perfectly habitable, Temperate worlds that are completely barren, and some Frozen worlds are completely barren.

 

[captainjack23]

I'm starting to think that use of the word "broken" is part of the problem with discussion these days (not just here or on the internet, BTW). It really sasy that the issue is physical and provavble, which it often isn't.
Anomolous or odd seems to work better in this context. Broken is a physical state, and could be provable. The ATM/SIZ planets are broken in that they can't exist.
The other categories are much more subjective, and more qualify as unbelievable, or, since some people do believe them, I'll stick with calling them an anomaly.
Like I said, and I'm hoping to stick to this myself, and keep it to the numbers. I'd really hope that this doesn't turn into another "my idea is right and yours sucks" argument.

 

[EDG]

The other categories are much more subjective, and more qualify as unbelievable, or, since some people do believe them, I'll stick with calling them an anomaly.

Eh, it's a valid point I guess.

I'd say that are "Broken worlds" are :

- impossible size/atm combinations
- the tech level is too low to allow the population to survive.
- the tech level is too low to allow the starport to run its facilities (e.g. shipyard)

... and "Anomalous worlds" are:
- high pop/hostile environment.
- population too low to justify starport.

One of the aims of the EDG worldgen is to completely eradicate the Broken worlds, and hopefully completely remove those "anomalous" worlds that CT and MGT generate in large numbers - it should be left up to the GM to add the odd anomaly if he wants.

That said, one idea that I am toying with though is the idea of a "Random Events" table at the end of worldgen. On an unmodified roll of a 12 on 2d6, you'd roll once on a table that had results like "Recent War", or "Starport Upgrade" or "Environmental disaster" or something that would modify the UWP in some way. That way about 3% of the worlds generated would be "anomalous" somehow, but in a much more controlled manner and with a built-in reason for their existence.

 

[captainjack23]

The other categories are much more subjective, and more qualify as unbelievable, or, since some people do believe them, I'll stick with calling them an anomaly.

Eh, it's a valid point I guess.

I'd say that are "Broken worlds" are :

- impossible size/atm combinations
- the tech level is too low to allow the population to survive.
- the tech level is too low to allow the starport to run its facilities (e.g. shipyard)

I'd agree with the first two - and am amazed that the second point hasn't come up for me before. Thanks for that.

The third really depends on what the nature of the starport is vis a vis the population. In one situation its broken, in another anomolous.

Since the categories are pretty artificial, its exact membership is not all that important as long as it gets looked at.

... and "Anomalous worlds" are:
- high pop/hostile environment.
- population too low to justify starport.

One of the aims of the EDG worldgen is to completely eradicate the Broken worlds, and hopefully completely remove those "anomalous" worlds that CT and MGT generate in large numbers - it should be left up to the GM to add the odd anomaly if he wants.

What I'm hoping to look at is what those numbers are -and possibly what can be tolerated.

That said, one idea that I am toying with though is the idea of a "Random Events" table at the end of worldgen. On an unmodified roll of a 12 on 2d6, you'd roll once on a table that had results like "Recent War", or "Starport Upgrade" or "Environmental disaster" or something that would modify the UWP in some way. That way about 3% of the worlds generated would be "anomalous" somehow, but in a much more controlled manner and with a built-in reason for their existence.

Well, it is one more table.....but if I can be absolved of being contradictory, I'd like to say I like the idea.

The 3% is good, about one/sector, about 70% chance of atleast one. I'm getting he feeling that that level of anomaly (one per subsector, for all types) is about what the forum members seem comfortable with.

 

[EDG]

What I'm hoping to look at is what those numbers are -and possibly what can be tolerated.

Well I've already posted those numbers here for CT, MGT and EDG 1.1, based on the criteria I defined here as well. Ideally I'm trying to get all the numbers down to zero (and most are already at zero in EDG 1.1 currently), but you may have higher tolerance to them than me.

Well, it is one more table.....but if I can be absolved of being contradictory, I'd like to say I like the idea.

*gasp!*

The 3% is good, about one/sector, about 70% chance of atleast one. I'm getting he feeling that that level of anomaly (one per subsector, for all types) is about what the forum members seem comfortable with.

I definitely feel it's worth investigating... I'll see what I can come up with (and you can blame TNE/1248 for this, they have a random events table in the Collapse/Recovery process).

 

[Rikki Tikki Traveller]

Also, regarding what is "Broken" depends on the setting.

If you take the Third Imperium (year 1105) as your base and ANY sector in the Imperium, then there should be NO garden worlds with low populations.

The borders of the Imperium have been settled for hundreds of years. Dawnworld in District 268 makes NO SENSE. WHY would an essentially garden world be uninhabited when all the worlds around it have been settled for several hundreds of years.

MAYBE you could come up with an explanation for that one planet, but there are a lot of low pop garden worlds in the Spinward Marches and the history just doesn't support any explanation for why they would ALL be ignored by the Imperium.

Imagine a situation where an isolated planet has just reached TL 9 and developed Jump-1 ships. They spread out 2-3 Parsecs in every direction. Probably every system will have at least an outpost in it, but assuming they find 5 planets with ATM 4-9, why would a reasonable person expect those 5 worlds to have lower populations than the rock balls around them? Colonists from Earth would go to the garden worlds, scientists and miners would go to the rock balls. In 100 years, which of those worlds will have the highest populations? Common sense says the garden worlds will be higher population.

Even if the native life is deadly to humans, we would scour a good size island and plant all Earth life. That is still better than a Mars like world where you can't even breath the air, or go outside without a vacc suit.

Going back to the basic assumptions and assuming a Long Night. Where are those people going to survive without outside help for a couple thousand years? On an airless rock ball or on a garden world?

To me, 2% is not an exception, it is statistically part of every sector and is too high for most random generation methods.

OK, that probably came on a bit strong, but as you can see, I am pretty passionate about getting this, to me, common sense settlement pattern into the core book.

 

[captainjack23]

Also, regarding what is "Broken" depends on the setting.
<snip for brevity>

To me, 2% is not an exception, it is statistically part of every sector and is too high for most random generation methods.

OK, that probably came on a bit strong, but as you can see, I am pretty passionate about getting this, to me, common sense settlement pattern into the core book.

Okay, you've got a zero tolerance approach to anomalies....I don't, and take a significance level approach. Science (hard and otherwise) has been debating this (with pitchforks and torches style of debate, quite often) for a long long time, and probably will for a long time* Not being inflammatory in the least, just a comment on where we will likely jam up in a discussion if we aren't careful.

Rumination over


*heck , even what a good level of significance is is contentious:
".05 ? for alchemists, idiots and witch doctors, maybe...01 is the only scientific standard";
"for hidebound frauds defending outdated dogma by perverting statistics, perhaps.."
"oh yeah ? "
"Yeah."
"Oh yeah ?"
"yeah."


"....blood everywhere, officer, it's horrible !"

 

[EDG]

Also, regarding what is "Broken" depends on the setting.
<snip for brevity>

To me, 2% is not an exception, it is statistically part of every sector and is too high for most random generation methods.

OK, that probably came on a bit strong, but as you can see, I am pretty passionate about getting this, to me, common sense settlement pattern into the core book.

Okay, you've got a zero tolerance approach to anomalies....I don't, and take a significance level approach.

Well I have a zero tolerance approach to anomalies too. What I want is a system that generates interesting worlds that make sense, not interesting worlds that don't make sense.

I mean,in the EDG 1.2 quadrant there's a world that's A100776-J - that's surely an interesting world. But it's all self-consistent and makes sense (though it may have a hard time fitting into the OTU with its high TL - but then the CT system has always been capable of generating up to TL L (20)). But it's not anomalous.

There's some worlds that are gov F and really high law level (e.g. A8947FK-A ) - again, interesting, but not anomalous.


The fact is, we really don't need anomalies to be able to come up with interesting worlds. If the GM wants to change something (or if we have a "Random Events" table to shake things up after UWP generation) then he can and should do that - but the aim of this system is to churn out worlds that make sense. Anomalies - worlds that don't make sense - aren't acceptable, because that shows that there's a flaw in the system, which means it's broken. And I see no reason to settle for a few anomalies when we can engineer the system to not produce any at all.

 

[captainjack23]

Well I have a zero tolerance approach to anomalies too. What I want is a system that generates interesting worlds that make sense, not interesting worlds that don't make sense.

I mean,in the EDG 1.2 quadrant there's a world that's A100776-J - that's surely an interesting world. But it's all self-consistent and makes sense (though it may have a hard time fitting into the OTU with its high TL - but then the CT system has always been capable of generating up to TL L (20)). But it's not anomalous.

There's some worlds that are gov F and really high law level (e.g. A8947FK-A ) - again, interesting, but not anomalous.


The fact is, we really don't need anomalies to be able to come up with interesting worlds. If the GM wants to change something (or if we have a "Random Events" table to shake things up after UWP generation) then he can and should do that - but the aim of this system is to churn out worlds that make sense. Anomalies - worlds that don't make sense - aren't acceptable, because that shows that there's a flaw in the system, which means it's broken. And I see no reason to settle for a few anomalies when we can engineer the system to not produce any at all.

Well, I'm not sure where I said that anomalies are needed for interesting world generation......oh wait... were you talking to me ?

:wink:

 

[EDG]

Well, I'm not sure where I said that anomalies are needed for interesting world generation......oh wait... were you talking to me ?

You have mentioned before that a "good enough" system that generated a few anomalies was OK for you.

 

[captainjack23]

Well, I'm not sure where I said that anomalies are needed for interesting world generation......oh wait... were you talking to me ?

You have mentioned before that a "good enough" system that generated a few anomalies was OK for you.

Ah. my interpretation of what I was saying (as opposed to what I said) is this: a good enough system for me is one that is unlikely to generate anomalies...kind of a difference there.

And of course, they'r only anomalies for me, if....well, we all know where this goes......

 

[Rikki Tikki Traveller]

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...).

I tend to look at things from the OTU point of view. So if something appears once per subsector, then it is not an anomoly, it is part of the setting. Once a Sector is a MAYBE for me. But, when you have 4-5 different types of anomolies all at 1 per sector, then you are back to what I consider "part of the setting" and I don't like it.

I don't have a statistical background and what stats I did have in college has been lost in the intervening decades, so I am not about to argue the details with you. I believe you when you say something is statistically likely or not likely; and I didn't take it personally, I am also trying to explain how I FEEL about this, and I am the first to admit it IS a feeling, not something that HAS to change.

If I generated a quadrant and saw a world A10776-J, I would change it. I might drop the TL to G and leave the rest, or I might throw it out and put something completely different in it's place, depending on how I felt that day.

The fact that it showed up in just a single quadrant tells me that there MIGHT be a problem with the generation method. Without the statistical background, I would be concerned that it would show up in every quadrant, which means 4 per Sector, so it is now a setting issue, and TL J worlds all over the OTU is totally unacceptable to me. ONE in the entire volume of Charted Space is PROBABLY OK, but not much more than that.