Good research is always useful, where good means something between "reliable and flawlessly explained" and "plausible and very novel". So on one end you could

Good research is always useful, where good means something between "reliable and flawlessly explained" and "plausible and very novel". So on one end you could takes something like Newtons Principa, that synthesised a lot of the known facts and laws at the time into one single coherent framework, and on the other you can have Galileo recognising that he was seeing surface features on the moon and not just patterns on a flat disk and announcing this will some illustrations showing the principle and not what he actually saw. As to the first part of the question, if you talk of "publishing" in academia you usually mean you want your work to be know by other credible researchers in the hope they find it useful. The most reliable way to achieve that without credentials is to befriend someone respected in the field you are interested in before you send them your results, because they are more likely to have patience with any errors or confusion in your presentation that way. There are of course some places that pride themselves on being completely free from filtering, but the concept of "spam folder" along should tell you how good a chance you have of getting any serious attention in those places.