In this post, I’ll write about the very, very, very few times all those grades in your “formative” category are, indeed, likely formative. Otherwise, I make the claim right here and now that they’re almost always summative…
Following the conventional definition of formative assessment, a “formative grade” suggests that it occurs *during* the learning process. Yet, these grades are almost always placed in a grading category that’s calculated into a student’s permanent record. This, effectively, is summative, not formative.
To explain, we need to distinguish between assessing and grading; they’re related, but not the same. So, you might actually give a formative assessment to check student understanding, and maybe even adjust instruction as a result, but once you represent that in the gradebook it becomes a summative grade…usually.
The few times these grades are actually formative is when they serve as placeholders with some kind of pathway to being changed (i.e., removed, or updated once more learning evidence comes in), or…they don’t calculate into the grade at all!
For example, a student with 55, 65, 85, 95, 95, 95, and 100 in a particular grading category clearly has improved their understanding during the learning process. Those low scores show previous struggle, and the high scores show considerable growth, which is all valuable information. If we use that information to determine a grade in a formative way, it would reflect *current* understanding and could change. If, however, we use that information in a summative way, it would lump together old and new, treating each assessment as unchangeable events along the way. This averaging of the older lowest scores with newer highest ones—though a common practice—makes no sense whatsoever.
Instead of that 84 mean average, the final score could be something more like 95 if the teacher is using trend (i.e., mode), or if the student is self-grading and thinks that number best represents their *current* understanding. The final score might even be 100 if the teacher is using a “most-recent-score” system to determine grades for each category. N.B. I don’t recommend this, mostly because it’s pretty high stakes and doesn’t account for off-days towards the end of the grading term, but it’s a common enough grading practice to mention here.
In sum, those example scores above are formative assessment scores because they occur *during* the learning process. What the teacher decides to do with them next determines whether they remain formative grades, or instantly become summative ones, regardless of what the category is called.
Magister P. 
Yep. 100%. The most recent evidence of a certain skillset should always win out. This is why, like you, I’ve completely disaggregated points an percentages from feedback. We go as long as we can without ANY grade showing up anywhere. I tell my students that every day of practice is what it is: PRACTICE. I tell them to always take their learning seriously and ask themselves these 3 key questions at every step: How am I doing? How do I know? What are my next steps? (Thank you, Tom Schimmer Podcast). Since we have to “play school” mid marking period, I observe growth over time, until a critical mass of students has reached a proficient level of understanding of the concepts. This usually occurs just in time to compile a number. This is when I report the feedback to parents in the LMS, still decoupled from a grade. Then I tell students to read my feedback and rate their level of proficiency on the rubric we had agreed on at the beginning of the year so that we can finally arrive at a “number.” I don’t look at their self-reported number yet. I look at their most recent evidence and plot it on the rubric myself. Then I reveal the score they give themselves. If it matches mine, great. If not, we have a built in understanding that the “over-rater does the explaining.” (Thanks again to Tom Schimmer for this one). Then we do the same for the 2nd half of the marking period. The final grade replaces the mid MP grade. Although I wish we could do away completely with grades and go entirely on feedback, I know this will not happen in my career and it is, as far as I am aware, as close as we will get to “meaningful grading.” BUT…I am always open to suggestions.