It all boils down to how the simulation was written.
If we can make it crash, time travel being possible is very likely.
If the parent universe is running an exploration of possible universes according to consistent laws of physics, upon crash it will probably backtrack the simulation to a previous timestamp, fiddle a little in the phase space, then continue forward.
One way to make it crash is making the simulation crawling to a stop by requiring always more computation to be done by preventing the simulator to take shortcuts in the computing process.
A backtracking mechanism to preemptively avoid crashes is probably deeply rooted in the form of an implicit solver in the time-stepping mechanism to guarantee the consistency of the tangled causal chains.
Which mean it is probably possible to exploit it from within the simulation. By deciding to do heavy computations in the future, depending on what you observe of the past and its implied future according to your previsions, you can pick your past among the physically possible universes. To guide the simulator to the future you desire, you have to create a consistent alternative past toward your possible future.
To not violate the second-law of thermodynamics which has to always increase (monotically) in your universe, you therefore have to reduce the entropy of the past by minimizing it (down to the origin of time, up to your current view point from inside the simulation), aka simplifying history, to give your physical self some room to shape the path from past to future the way you want.
But by the time, you are able to get your time-machine to work, you'll have realized the futility of it and went to the upper plane of existence.
> One way to make it crash is making the simulation crawling to a stop by requiring always more computation to be done by preventing the simulator to take shortcuts in the computing process.
Isn't this the reason for mass-induced time-dilation? When you add more mass (computational complexity) into a region it runs slower, to the point of being removed completely from the sim, hidden behind an event horizon and described by just a couple of parameters.
Could be, but it only works in a non-adversarial context. Biological numerical instabilities are a tougher beast to tame.
When you have to simulate a big star you can do it quickly by approximating it by a sphere of a certain temperature. The bigger the star, the more the law of large numbers apply and the better your approximation get.
When you have to simulate a computer that would be the size of a star, you can't take any shortcut. And the amount of computation that a computer the size of a star would allow you to do according to the limits of physics is huge, and would allow to simulate the full universe many times (see https://en.wikipedia.org/wiki/Limits_of_computation).
One way to avoid a simulation crash would be to introduce time-dilation effect to throttle the quantity of computation locally, but fundamentally it's a very hard problem equivalent to the halting problem.
If the parent universe is running an exploration of possible universes according to consistent laws of physics, upon crash it will probably backtrack the simulation to a previous timestamp, fiddle a little in the phase space, then continue forward.
One way to make it crash is making the simulation crawling to a stop by requiring always more computation to be done by preventing the simulator to take shortcuts in the computing process.
A backtracking mechanism to preemptively avoid crashes is probably deeply rooted in the form of an implicit solver in the time-stepping mechanism to guarantee the consistency of the tangled causal chains.
Which mean it is probably possible to exploit it from within the simulation. By deciding to do heavy computations in the future, depending on what you observe of the past and its implied future according to your previsions, you can pick your past among the physically possible universes. To guide the simulator to the future you desire, you have to create a consistent alternative past toward your possible future.
To not violate the second-law of thermodynamics which has to always increase (monotically) in your universe, you therefore have to reduce the entropy of the past by minimizing it (down to the origin of time, up to your current view point from inside the simulation), aka simplifying history, to give your physical self some room to shape the path from past to future the way you want.
But by the time, you are able to get your time-machine to work, you'll have realized the futility of it and went to the upper plane of existence.