1.0g way too common

[Bombum451]

Observed behaviour

Too many planets have exactly 1.0g of gravity

Expected behaviour

Should be more variation, and not exactly 1.0g

Steps to reproduce

Check out Zarelium (-12, -1, -2). All 5 main planets have exactly 1.0g

My pioneer version (and OS):
v20230203 (e77da96) Windows

[zonkmachine]

Since the surface gravity for our solar system seem to work well, and it's fed real and good data, it should be the automatically generated planets that gets suboptimal values for either mass or radius (or both). The planet mass seem to be generated first (I haven't found where).

The line controlling the surface radius is:

pioneer/src/galaxy/StarSystemGenerator.cpp

Line 703 in 3fed83c

sbody->m_radius = fixed::FromDouble(pow(sbody->GetMassAsFixed().ToDouble(), 0.5));

Simplified: radius = pow(mass, 0.5));
Radius is just a function of mass with no function for varying density built in.

Surface gravity is computed in SystemBody::CalcSurfaceGravity(), the responsible line is:

pioneer/src/galaxy/SystemBody.cpp

Line 597 in 3fed83c

return G * GetMass() / pow(r, 2);

Now we add the radius equation into the gravity equation.
G * mass / pow(r, 2); <- r = pow(mass, 0.5)
G * mass / pow(pow(mass,0.5), 2);
G * 1 !
Since the mass is number of earth masses the result will always be 1.0g for planets between 1 and 200 earth masses.

The surface gravity code above, I trace to commit 5286161
@fluffyfreak

A quick fix could perhaps be to lower the exponent a bit from 0.5 to something like 0.4- 0.45 just to get some variation in there. A more advanced fix would include some randomised mean density. If I'm right above that is.

[zonkmachine]

pioneer/src/galaxy/StarSystemGenerator.cpp

Line 694 in 3fed83c

// http://phl.upr.edu/library/notes/standardmass-radiusrelationforexoplanets

It looks like this file has been moved. I think it's this one:
https://phl.upr.edu/library/labnotes/standard-mass-radius-relation-for-exoplanets

[fluffyfreak]

I have been summoned! 😂

Yeah, that was from when I knew things and was smart, so much time has passed since 2014 (omfg)

[fluffyfreak]

That is some great detective work and yeah I never thought of that before 🤔 I guess that we don't take into account the possible variations in density due to the being Rocky->Ocean->etc type worlds that 1->200 Earth masses can be. Even 1 Earth mass world might have very different radius due to it's composition.

[zonkmachine]

Here's a oneliner that randomizes volume to achieve varying density. I'm also testing adding mean density and escape velocity to the planet infobox. If you accept a simple/temporary fix for this I'm happy to pull something.

diff --git a/src/galaxy/StarSystemGenerator.cpp b/src/galaxy/StarSystemGenerator.cpp
index f97387cd3..ab057ecb3 100644
--- a/src/galaxy/StarSystemGenerator.cpp
+++ b/src/galaxy/StarSystemGenerator.cpp
@@ -703,8 +703,11 @@ void StarSystemRandomGenerator::PickPlanetType(SystemBody *sbody, Random &rand)
                sbody->m_radius = fixed::FromDouble(pow(sbody->GetMassAsFixed().ToDouble(), 0.5));
        } else {
                // Anything bigger than 200 EU masses is a Gas Giant or bigger but the density changes to decrease from here on up...
                sbody->m_radius = fixed::FromDouble(22.6 * (1.0 / pow(sbody->GetMassAsFixed().ToDouble(), double(0.0886))));
        }
+       // randomize radius
+       sbody->m_radius = fixed::FromDouble(sbody->GetRadiusAsFixed().ToDouble() * (1.2 - (0.4 * rand.Double()))); // +/-20% radius
+
        // enforce minimum size of 10km
        sbody->m_radius = std::max(sbody->GetRadiusAsFixed(), fixed(1, 630));
 

[zonkmachine]

The rand function should probably have a normal distribution. overkill

[Web-eWorks]

Closing as fixed by #5592 - future improvements can be tracked via a feature request post in the forums or additional defect reports.

[zonkmachine]

Just dropping this here. The custom systems needs to be checked and fixed separately. In the example below the planet Death has a density that is higher than any known elements. Unfortunately I won't have time to look into this anytime soon.