Next to theory and experiments, simulations are one of the most important tools used in research today. Occasionally, scientists develop theories that cannot be tested using today’s practice or technology. Here, a simulation might then be able to point the theoretical physicist where he or she needs to look. Other times, it might happen that there are two different theories that could be suitable for describing reality. If simulations are built based on both theories, their results can sometimes separate the significant from the useless. And sometimes it also happens that there isn’t any theory yet, only data from measurements. If it’s possible to create a simulation that produces the same results as the experiment, then sometimes it’s also possible to derive a theory from the simulation.
Astronomers who want to simulate the cosmos are normally confronted with a choice: either they use their computational power for details or for the largest possible space in their simulation. Both methods have their drawbacks in their informational value: simulations of small numbers of galaxies cannot provide good statistical results, and large-scale simulations lack the details compared with reality. With the TNG50 astronomical simulation, researchers were able for the first time to combine a large-scale cosmological simulation with the high resolution of a detailed simulation, like those that were previously possible only for studies of individual galaxies.