Going from the parent comment's assertion, the answer would be that if you want to train something to predict physics, the original physics simulation that's generating the data is either:
1. Already doing that, in which case it's already the superior model.
or
2. Not doing that correctly, in which case the synthetic data output is equally poor.
1. Already doing that, in which case it's already the superior model.
or
2. Not doing that correctly, in which case the synthetic data output is equally poor.