Bigger nukes do less damage!

Yeah, for some reason atomic bomb damage barely scales with yield, if the bombs are detonated at contact range. I think it does affect the radius of the splash damage, though.

I have a 1.38 Mt design that weighs only 164 kg, but I never use it because the effects are so underwhelming. Cluster-bombs containing several small nukes are far more effective, especially considering the cost of uranium.

The fluency can be ignored in this case. Since the explosion is contained within the ship and the energy of the explosion is the same, the damage done should be equivalent. That's not to say it should be the same, but the fact that the larger nuke does so much less damage. Look at it like a giant pipe bomb: provided the container can withstand detonation, the energy imparted is equal, regardless if I us TNT or the energy equivalent of a different explosive. This is true even for nukes. Since the gunship hull stays generally in tact, energy the imparted in both cases should be similar since the explosion energy is the same in both cases.

This makes quite a bit of sense.

I also think it may have something to do with the way the game computes explosions. I assume it does what all games do and creates multiple ray casts. Larger nukes making more spaced, and more potent for the calculation rays, while the smaller nukes have more condensed, and less powerful rays applying smaller, but more even and, as a result more consistent and large scale of damage when all the pieces interlock.

While I don't necessary disagree with you, some of your conclusions seem dubious to me. First, I don't think that using ray cast for simulating damage has anything to do with energy distribution or conservation. Second, why do you think that the whole energy of your projectile should manifest as 'damage'? For example, both elastic collision and inelastic collisions completely conserve all kinetic energy of the system. They are ideal cases but the point is, only a portion of energy becomes heat and deformation. How exactly that energy is distributed is the "meat" of damage modeling, but it has nothing to do with ray cast or bounding sphere or any other collision model.

Here is my theory of how this all works. I agree that the whole damage allocation is based upon a simplest case: a slug (lets call it a fragment) impacting armor. At the point of impact it it leaves a pattern (actually two patterns, one for heat and one for holes). Everything else is based on that, and under the hood nuclear blasts, fragmentation rounds, lasers and simple slugs all use this effect. This leaves questions on how exactly a warhead with thousands of fragments is simulated; I doubt the game does collision detection for all of them.

I think there is a clue in the CoaDE blog, and it is this:

QSwitched has said that most of these types are modeled along with projectile and parts of armor turning into plasma. So I think the pattern a point impact leaves on armor or module is dependent on its type and initial energy. It determines then will the impact proceed further, and if so, which pattern will it take, depending on whether it has turned into plasma or acquired more material and energy from the current layer. Usually the pattern grows as the fragment goes for layer to layer, but often at some point it just becomes a heat pattern without penetration. It is quite conceivable that a dense thin slug will never shatter or turn into plasma, penetrate both sides of the ship still carrying considerable energy. The energy is lost, as you've said for the purpose of damage allocation, but this is accurate.

I think I read somewhere that modules are a different story. The modules have complex composition which was too hard to accurately model. So a module has three states: undamaged, damaged and completely gone. It becomes damaged when its skin is penetrated (which is just the same as a single layer of armor). I don't know under which condition the module is gone completely. The point is, a fragment cannot penetrate a module and continue onward. If the skin of the module is penetrated, the fragment either damages the module or removes it completely. I don't know if modules have hidden HP or it can absorb a certain amount of energy based on volume or something else. If this is true, then modules can really lead to energy truly vanishing.

Just a note here: when a projectile bounces, it doesn't mean it won't leave a mark on the surface of the armor. The heat will also dissipate from the point of impact proportionally to heat conductance of the material. I am not sure how realistic the specific situation you describe is, but it's not either bounce with no energy loss or all energy converted into heat and deformation.

Likewise, it is a perfectly realistic scenario where a projectile splashes against the armor heating it but without leaving a hole. Whether it will heat it enough to glow is another matter. My intuition tells me it will penetrate the armor or shatter long before it heats it enough to glow in case of physical projectile, but I of course may be mistaken.

Adding nukes to manned spaceships is an easy way to test effects.....

You've accelerated 5 tons to 774.6 km/s or am i misunderstanding something?
This is a 1 gram projectile in the railgun you've linked, right?

I don't think this is the case. I think ray traces can cause area damage.
I'm not sure what happens with the coilgun, but if you look at your second screenshot, you will notice that the single nuke most likely produced exactly six rays in six directions.