This whole issue speaks to what is the purpose of a simulation. Observations generate the need for a theory to explain them. The theory comes along, and simulations try to check the full ramifications of the theory. When the simulations are difficult, it is not at all uncommon for the simulations to fail to completely link the theory to the observations. Sometimes it means the theory is missing a key element, other times it means the simulations are not reliable. Supernova simulations were carried out for a very long time before they could actually get stars to supernova, and still have problems with it, but no one says that the theory that massive stars go supernova is "wrong". It just means that there must be some aspect of what is happening that is not properly accounted for in the simulations, either because of numerical problems or due to improper simplifying assumptions. It's not unusual.