Manuel,
"We actually think our opinions matter when it comes down to understanding nature", because they do matter, at least with regards to our understanding. However, I believe, and it appears that you also believe, that our options, beliefs and understanding of nature, has little influence on nature, apart from our earth-bound technology (Which is mostly about how to cause various, desirable effects.) Nature is indeed not about our options, but our opinions about nature cannot be separated from our understanding of nature.
I'm not quite sure what you mean be highly *calculated* guesses. But consider the following rewording of your statement: "all we are doing is making highly accurate guesses without knowing why that is so."
If I could accurately guess which stocks will go up each day, I would soon be the richest man on earth. If I could accurately guess which chemicals would cure cancer, I could add many years of life to millions of people. Even if I have no idea why it works, it may nevertheless, be very beneficial to many people. That is the primary reason people seek to find causes for desirable effects. For the intellectually curious, it would be nice to gain an understanding of why and how each of these causes produce the desired effect. But benefits exists, even in the total absence of any such understanding.
When looked at in this way, there seems to be a clear distinction between technology and science. The first is satisfied with simply obtaining the benefit. The second seeks to understand why and how each newly discovered cause produces the effect. But the distinction is not really so clear. If one could guess the form of a new theory of everything, that was subsequently verified to actually work, no one would care that the theory was obtained via a guess - you would get a Nobel Prize regardless.
Personally, I do not view the role of science to be that of "explaining nature." It cannot do so. Its role is to describe nature. But that includes describing how to bring about (cause) any desired effect, including the ability to predict the future behavior of physical systems, via deductive logic.
Rob McEachern