Okay. First off, there are no such thing as stop words and no such thing as a keyword match. Stop words disappeared much more than a decade ago. And for Google, direct keyword matches never happened. Any apparent keyword match is incidental to the semantic processing and weighting of your content. Under no circumstances does Google match keywords. Period. Instead, what appears to be a keyword match is only the highlighting the search query terms in the results. Nothing more. Sites and pages can and often are returned without one of more of the search query terms. One obvious example would be car, automobile, and vehicle. It would be extremely foolish for a search engine to ignore a page that does not contain the term car when it contains the term automobile.
From a semantic standpoint, terms with 2 or less characters have little value in regard to scoring. This is because these terms are not a subject, predicate, object, however, they can be modifiers. In that respect, the value of these terms is understanding the content and not content topic. For example, "Bob bought a car.", Bob is the subject, bought is the predicate, car is the object, however, a is a modifier that indicates a single car. You may say that the plural version of the word car, cars, would do the same. You would be right, but that is not always reliable. Some terms do not have a singular or plural versions. The immediate example would be macaroni. Bob could eat a macaroni or some macaroni. Of course, there are better examples. But you get the point. These terms are not stop words or ignored. They are a very important of the semantic analysis.
In the past, it was believed that so called stop words diluted the value of meta tags such as the title tag, description meta-tag, etc. and that these terms were ignored. Even as that advice was originally written, it was not at all true. There was some level of support for this. SEO are blackbox testers and not necessarily technical people. What appeared to be true was evidenced by the SERPs which is actually a multi-part process. The search query result is heavily modified through the filter processes and the search query terms then highlighted as the last step. So to say that the first SERP entry contained the "best keyword match" as an SEO would, is not necessarily true and often not true. The best query term match may be much lower in the SERPs. What ranks as the first SERP entry may be the most popular as evidenced through CTR (click-through rate), influenced by authority, link value, etc. The list of filters is fairly long and any of them can easily knock a best query result set match out of the top position.
So what is the answer?
Create your site for humans and not for machines! Old advice that we seem to have to echo almost each and every day. Write full sentences and let the full semantic analysis do it's work. Stop trying to feed machines thinking that you are smarter than the whole cadre of multiple PhD's at Google who have spent their entire career understanding search queries and content with a high degree of complexity.