Testing and understanding the past is not the same as predicting the future. Our physical world has been extensively tested, which how we learned about it. If

Testing and understanding the past is not the same as predicting the future. Our physical world has been extensively tested, which how we learned about it. If you want to treat gravity stopping in the future as a probability you can, but once but the more you understand the more you would realize that it is astronomically low. This is not the same as saying that from some technical extreme semantic definition that something can be improved without testing results. In a complex system you need to know your results to inform your path forward. Trying to come up with some super narrow, reductive, irrelevant definition on your own is worthless in this context.