Welcome, Guest

Author Topic: Vanishing masses  (Read 2361 times)

Chaos Theory

  • *
  • Posts: 5
Vanishing masses
« on: May 30, 2017, 06:40:27 AM »
Hi there, I am new on this board. My name is Patrick, I am 40 years old and from the Netherlands.

Some month ago or so I bought this game/simulation and took the challenge to make a YouTube channel with it.

But yesterdays video I made took me over the edge. I mean, I placed more than 100 Jupiters, and there were definitely more than 100 Jupiters, maybe even 130 or 140, very close to each other to let gravity do its work. You would expect all those gas balls to fluently flow into each other, right? No rocks to throw with.

After a while there remained two pieces which I finally brought together.

So, collapsing 100-120 Jupiters would give a nice little red star wouldn't it?
No, it did not. According to the statisstics it gave me a Jupiter with a mass of only 0.2 Jupiter mass.

I know there will always be some debris escaping, but 99,8% of all matter escaping?

Now I am not an official astronomer, nor a physicist or any scientist at all. Just an enthusiastic citizen. But can this be right? Can someone tell me what is going on here?

The video can be found here:


« Last Edit: May 30, 2017, 07:03:46 AM by Chaos Theory »


  • Developer
  • *****
  • Posts: 732
    • Universe Sandbox
Re: Vanishing masses
« Reply #1 on: June 01, 2017, 09:47:12 AM »
Thanks for reporting this, Patrick.

Essentially, no, this is not accurate. Dealing with mass loss is a difficult task for us, as it is very computationally expensive to conserve all mass along with all of the other calculations going on. The biggest issue is when there are a lot of fragments, as they are eventually removed (along with their mass) for performance reasons. This also means that results may vary depending on the performance of the simulation, which will be affected by sim factors (number of bodies, etc) as well as hardware.

That said, we definitely want to improve this and address this issue as much as we can. One short-term idea is to simply keep track of and display to the user a "lost mass" value. This would at least give a better sense of how much mass should be in the system.

Longer term, we really want to improve how we balance things like performance and accuracy. Included in this is a higher priority on conserving mass when deciding when to delete fragments, and which ones to delete. Another idea is to transfer any mass to a nearby body when a fragment is deleted, so overall mass is never lost.

There are different ways to approach this, and unfortunately none of them are perfect, but there is a lot of room for improvement.  :)

Hope that helps explain a bit.

Chaos Theory

  • *
  • Posts: 5
Re: Vanishing masses
« Reply #2 on: July 03, 2017, 06:39:26 AM »
Thank you for your reply.

In essence what you are saying is that so many object create even MORE tiny objects that calculating their orbits and properties is computer intensive you actually need a supercomputer to this. This reminds a little bit of Battlefield Bad Company 2 and Red Faction Guerilla in which both environments are destructible but most elements of the properties destroyed vanish.

I see now you are balancing between the choices:
- Do we want to look it as realistic as possible
- Do we want it to be as accurate as possible
and you cannot combine it......(as said, it would require probably a super computer).