The ability to consistently predict behaviour is sometimes considered a significant and powerful aspect of natural science because it indicates that the knowledge we have that allows us to successfully predict the behaviour of something "in the future" is highly reliable and valid.
However, the concept of prediction is one that is not as simple and straightforward as our everyday, common-sense, use of the concept might suggest. Although we will not investigate this idea here, it is possible to argue that the ability to "successfully predict the future behaviour of something" is not necessarily an indicator of reliable and valid knowledge.
Furthermore, natural scientists are careful not to to suggest that it is possible to unconditionally predict future behaviour since the conditions under which a prediction can be made now may not exist at some unknown point in the future.
In basic terms, therefore, the concept of prediction is not necessarily as crucial or fundamental as common-sense might lead us to believe...