In everything from physics to biology to politics and economics. It's complicated. More complicated than humanity has sorted out as well as we'd like to think. We sometimes know enough to push for favorable results in our endeavors, with more success than we would if we knew nothing. Even the notoriously chaotic processes which drive the weather, we predict with some degree of reliability. A reliability which is more pragmatic than none at all.
I'm skeptical of the very concept of chaos. I'm inclined to think that it's a word we use when things are unpredictable for reasons we don't understand. In science and the human condition more broadly. We often fail to take into account even stable variables in initial conditions that can produce dramatically different outcomes. Comparisons between similar samples, different outcomes, the result of overlooking important but nuanced phenomenon that occur within developmentally critical time-frames. Eventually, another study or perspective might come along to correct for it, getting a little closer, but missing something else.
This occurs in every rabbit hole I've ever attempted to delve into, ubiquitous and unresolvable, but indicative that there's always more to learn. I've often found myself trying to argue that anything and everything is more complicated than it seems. Reductionism being more a problem of assuming we have all relevant information, when that's never going to be the case, wearing blinders in an attempt to block out the confusion of everything we don't know.
Does the reductionist honestly believe it's all noise, or do they merely treat it as noise, so as not to be impossibly bogged down in their ambitions? I'm guessing the answer to that varies. It's probably complicated.
Seems to me that such behaviour might have developed for being evolutionarily effective. Maybe we jump to conclusions so that we can act on them, neurologically rewarded for doing this, it becomes habitual and self-reinforcing. Doing so produces results, from knowing to bring an umbrella, to being able to design working technology, to not being paralyzed by indecision. It's not understanding in any comprehensive sense, but nothing ever is.