All posts by Nils

By day, Nils works as a project manager in the IT industry. By night, he explores strange new worlds, protects the downtrodden, and practices conquering the world. To follow his exploits you can subscribe to his YouTube channel, Have Joystick, Will Travel and follow him on Twitter.

World Size

Traveller system generation starts with determining the size of the main world. Roll 2d6-2. The result is read in 1000s of miles, though of course we use metric units around here. Like so:


Diameter (km)

Surface Gravity (g)





Asteroids, small moons




Ceres, Orcus, Makemake

























Venus, Earth, Kepler 20f




10 (A)



Worthless Rocks Revisited

As we found out in the last post on the subject, almost 42% of the worlds we generate this way will be size 4 or lower and thus fall into the special “hard science” rules that almost ensure that their atmosphere is “none” or “traces”. They all fall into the gaping huge category of “worthless rocks” that will make up 40% of the systems.

One of the ideas of mitigating this is to tamper with the size rolls. We can’t really add any modifiers to the roll since the size roll is the first datum we generate about our new system.

Instead, we could use 3d6-3. The probabilities of getting certain world sizes change quite a bit, and results go up to 15:

Blue is 2d6-2, Red is 3d6-3

With this method, only 16.2% of all main worlds will be size 4 or less. However, another 16.2% will be larger than size 10. Does this open up a new can of worms?

Just how big are Exoplanets anyway?

Although we know a lot more about exoplanets than we did in 1977 (we didn’t know anything, then, after all) our data is still severely limited because detecting exoplanets is no trivial task.

However, on the upper end it seems that planet sizes run smoothly all the way up to Jupiter and beyond. This doesn’t say anything about probability, just whether or not such worlds exist.

NASA Kepler Planet Sizes Compraison
NASA Kepler Planet Sizes Compraison

We don’t know much about the makeup of exoplanets. However, we have been able to determine both mass and radius of some exoplanets. From this data, it seems that worlds with a mass over 1.5 Earths will not get rid of the hydrogen they collected during planetary formation, and are likely mini gas giants.

Potential 3d6-3 results

An extended Traveller world size table up to 15 (F) would yield:


Diameter (km)

Diameter (Earths)





Kepler 10b








Kepler-21b, Kepler-9d







Kepler-20b, Kepler-23b

Kepler 10b is an earthlike world (just too close to its primary); 33b has a high surface gravity (3.6g) and is thus useless for colonization; 21b is hot enough to melt iron; 9d is another insanely hot world; 20b is the “largest planet with an earth-like density”; and not much is known about 23b.

Count size 11 and 12 as superearths, which may be unpleasant but still basically habitable. Size 13-15 are mini-gas giants. If they are inhabited at all, people will live in floating cities akin to Cloud City on Bespin.


Barring playtest results to the contrary, it is feasible to use this method. We have effectively reduced the amount of fringe-y, useless worlds from 42% to 32% and we have split the group itself into half, achieving the goal of “more variety”.

There will of course be side effects because world size feeds back into Atmosphere and Hydrographics, but I think this is not a problem. If we can have blanket filters to set worlds to vacuum on the low end of the spectrum, we can do so at the high end as well.

I will deal with atmospheres in the next post on World Generation matters.

Update: 3d6-4

After some trial runs, it became apparent that 3d6-4 is a better fit for what I want. Not much of a difference, right? But the probabilities do shift a bit.


Worthless Rocks

When rolling up worlds for my sector, I used the “hard science” options in the Mongoose Traveller world generation chapter. I do not obsess about perfect realism – this is entertainment and not a simulation – but I prefer my settings to be plausible and consistent enough to allow suspension of disbelief.

However, there is one issue that I ran into almost at once. The rules specify some hard limiting factors on atmosphere. To quote the SRD:

  • If Size is 0–2, Atmosphere is set to 0. The world is too small to retain an atmosphere.

  • If Size is 3–4 and Atmosphere is 0–2, set Atmosphere to 0.

  • If Size is 3–4 and Atmosphere is 3–5, set Atmosphere to 1.

  • If Size is 3–4 and Atmopshere is 6+, set Atmosphere to A.

The problem is that 41.67% of all generated worlds will be size 4 or smaller, and the vast majority of these will be vacuum worlds or have trace atmospheres:


At the same time, these worlds only suffer from a -2 DM on population at most, meaning that 83.33% of them end up being inhabited with up to “hundreds of thousands” of people.

In my test sector, I ended up with 16 worlds of size 4 or lower with atmosphere A, and 167 with atmosphere 0 or 1 – out of a total of 424. That’s 39%. None of these 167 worlds ended up with a hydrographics score other than 0.

This might be realistic. We simply don’t know at this point, though it does seem that larger world are abundant in the universe. However I do think it’s an issue when 40% of the “Best” worlds of all systems basically look the same. They may be physically different, but they will all be “humans living underground or in domes”. Not inspiring.

Potential Fixes

Assuming I don’t want to maintain the status quo and do not want to regress to the regular, non-“hard science”, system, I have some ideas that might produce better results.

  • Allow denser atmospheres types, perhaps category 2 and 3 (very thin), for worlds of size 3-4. Unfortunately there are no examples for this that we know of. Titan is small (size 3) and has a thicker atmosphere than Earth does. Most In addition, Mars (size 4) is thought to once have had a thicker atmosphere. Both should really be type A though.
  • Allow E and F atmospheric types. Both read like reasonable examples; for example if Europa (size 2) was closer to the sun it could potentially be classified as a class F atmosphere.
  • Lower the chance for human population. People don’t move anywhere without a good reason; and there’s perhaps none to move to a worthless rock. The benefit of these “crappy systems nobody wants and nobody goes to” is that you can hide things there that in any inhabited system would quickly be discovered and salvaged.
  • Change the way size is generated: While there is no reason not to extend the size scale upward a bit (super-earths are quite common) there is no way to do so with 2d6. It might be possible to change to 3d6-3 and either accept the change in distribution or use a lookup-table.

I think I prefer a combination of changing the size roll and lowering chances of human habitation on unsuitable worlds. I’ll take a look at all options in later posts; meanwhile, is there another approach that I missed?


When I first started working on the science fiction setting, I created several maps covering the Orion Spur and even the Milky Way galaxy. I quickly worked out that these settings were much too big for me to handle, even if I only worked on selected “important” systems.

I was in the process of working out the “correct” size out when Realmwright finally convinced me to watch Firefly – I had been hesitant because I knew that it had been cancelled. I was surprised how much Firefly “felt” like a Traveller game. More importantly, it made me think how important recurring characters and locations are, and I finally decided to bring the setting all the way down to the very small scale of a Traveller sector. (This was, incidentally, when I began using Traveller to model the setting.)

I ended up creating an entire sector, but I did not really think about the process first and found out, later, that it was flawed – one of the key lessons that led me to the creation of Contact Light.

How “big” is a sector really?

Obviously, a Traveller sector consists of 4×4 subsectors, each measuring 8×10 parsecs. But 36×40 parsecs doesn’t actually tell us much about the scale of a setting.

Travel times: Once I had decided on using a Traveller sector for my setting, I worked out that a constant-speed hyperdrive with a velocity of 0.5 parsecs would be ideal. A subsector is 14 hexes across (diagonally), so that means 28 days – four weeks. For comparison, it took Colonial Age settlers four to five weeks to cross the Atlantic ocean.

Number of systems: As it turns out, a Traveller sector is quite a crowded place. The Mongoose basic rules book recommends a 50% system density (and if memory serves me right this has always been more or less standard). Since a subsector has 80 hexes, this means forty worlds per subsector or 640 worlds per sector. That’s a lot. Incidentally, my current Sector contains 424 worlds – some subsectors are more sparse – and that feels overwhelming.

What’s “realistic”?

The stellar density near our solar system is 0.004 stars per cubic lightyear, or about 0.14 stars per cubic parsec. The nearest we can get with one or two dice is a 1-in-6, or  16.67%, chance. This would result in ~ 13 systems for a subsector, or 208 in a sector. (Anecdotal data: The Sol sector in Traveller contains 18 systems.)

What can we do with 2d6?

In reality, I can’t imagine why anybody would want to roll even 80 hexes manually, but let’s still stick to the dice conventions. So what can we actually do with 2d6? This quite easily calculated, too:




Per Subsector

Per Sector

2 or less





3 or less





4 or less


Near Earth



5 or less







6 or less








7 or less








In other words, my sector worked out spot on at the 33% mark, or twice the stellar density near our Earth.

And I frankly can’t imagine what a sector at 58% or higher will look like. Messy, for sure.

What’s “good”?

I don’t think there’s a single number you can cite and say “this is what you should use”, since it’s too intimately tied to other assumptions in your setting.

Looking at these numbers and at my first sector, I have a gut feeling that 27.78% is the sweet spot for what I have in mind. But as we noted, it’s not only the number of systems that matters, but how interesting they are; let’s keep the numbers in mind and revisit them when we’re further along in the world generation system.

Setting Assumptions

Since I have a very specific setting I am building, this setting will color the more setting-independent posts I plan for Contact Light. Thus, I think it makes sense to post some assumptions:

  • FTL: A variant of Hyperdrives. Travel speed is about two days per parsec. Notably, a form of FTL Radio exists but it’s not quite as useful as you might think.
  • Power: D-T fusion with “realistic” fuel requirements. This means skimming and ocean refuelling is not very attractive, and every now and then ships need to refuel at a civilized world that can produce Tritium.
  • Psionics: Exist and are fairly common but not insanely powerful. They are not suppressed or hidden and have influences on culture, politics, and the economy.
  • Aliens: Multiple and in some cases quite humanlike alien species exist.
  • Tech: Standard TL is 13. Higher tech exists but is very rare. Indeed I will try to keep the TL as low as possible.
  • Location: The setting is not related to Earth in any way and may or may not be located in the Milky Way galaxy. There are humans here, yes, but they have never heard of a place called “Earth”.
  • Scale: Roughly one sector.
  • Ships: It’s a Medium ships universe. In Traveller terms, I expect the largest ships to be roughly 100k dTons. Note that we will be able to cram much more stuff into them since they are not floating water tanks.
  • Style: Golden age space opera. Think Larry Niven, H. Beam Piper, Robert A. Heinlein’s juveniles, Isaac Asimov, Edwin Charles Tubb, early George Lucas. Star Flight (video game). Maybe some Firefly. Most of space is a frontier and not a suburb. Utopia and transhumanism haven’t happened. Starships show signs of wear and tear. Kinetic energy makes for great weaponry. Mankind advances, but human problems stay the same.

If you are a Traveller guy you probably see quite easily that the mix is similar to the foundation that Traveller was built on, but with some subtle differences that will have a great effect on the setting.


Hi there. Welcome to Yet Another Blog™.

As some of you are aware (considering the likely distribution as I write this… most?) I am working on a science fiction setting. I am trying to do things in a more or less systematic way so as to end up with a constructed universe that is both manageable for myself and my audience, and elegant in a way that doesn’t break suspension of disbelief too badly.

After some considerations I have found a set of assumptions and parameters that works well for me, and I found that these are a very good match for an old favorite, Marc W. Miller’s Traveller. This is hardly surprising, considering that I grew up on a diet of science fiction not unlike the mix that Miller used as the basis for Traveller.

However, the setting I have in mind is not a Third Imperium Traveller universe, and I can’t – and do not want to – just use Traveller material. Instead, Traveller will serve as the system that will guide the creation of my science fiction setting. Nothing more, and nothing less. I expect that this system will itself get modified and shaped in the process – because that’s what I enjoy.

Limited Scope

This blog will document the process of creating my science-fiction universe. It will focus on technical aspects, on world-building and on process-, system- and game design. Actual content will be posted elsewhere – I will announce it in time.

Why separate blogs?

I normally post my world-building material over at Enderra.  However, I am unhappy with the wild mix of stuff that is in that blog. It lacks focus. Instead of adding to the chaos, I decided to just branch out for this particular purpose. Call it an experiment.