i have been curious and thought about this topic in some form or another a fair bit, so I beg your patience with my long post.
Karma points on HN are a psychological gamification to encourage posters to write good posts and avoid bad posts with concrete quantized peer feedback.
People like getting points. Currency itself is an ancient precursor to such motivation-driving/measuring point systems. Grades are another from the late 1800s. Experience points are another from gaming in the 1970s. Points give a person a sense of power, progress and control and more subtly may be interpreted by the owner to bolster some piece of their identity (as smart, helpful, knowledgeable, right, etc).
Quantization probably also helps with implicit or explicit goals, being a good chunk of what you need for successful/SMART human goal setting (Specific, Measurable, Achievable, Relevant, Timebound) or AI goal setting, basically efficient feedback loops in general.
Karma points systems on forums are also a form of distributed moderation that helps the forum to scale quality discussions with reduced centralized moderator activity. It was first used widely on an internet forum called Slashdot in the late 1990s. Using karma techniques, Slashdot was able to host relatively meaningful distributed discussions that exceeded the posting volumes of previous online forums such as Usenet, where the largest forums there only handled 3000 comments a day (alt.fan.rush-limbaugh... if I remember my analysis correctly). Usenet had more centralized human moderation and had trouble scaling quality conversations as a result. A good moderator encourages good conversation and cuts off the bad stuff few want to waste time reading. But there are only so many hours in the day to moderate and only so much patience and tolerance for complaints that come with the task. One human or two or three can only give so much feedback. (Until we get LLM/AI moderators trained on all these point histories! Sigh.)
HN, like other sites uses a blend of high-effort centralized human moderator feedback and low-effort peer karma feedback (and high-effort peer replies/posts). The choice of exactly what powers someone with points can get, and whether point totals or history can be seen by whom or are capped or attributable per-topic (stack overflow, kinda) are all forum design decisions that can affect human motivation and/or the potential for abuse.
Some sites show you up votes/downvotes (like YouTube videos, Amazon feedback) so, as you point out, you see the "volatility"/controversy-ness of a post more than you get with a single karma number on a post as on HN.
I don't have firsthand knowledge of HN's motivation for the approach (comments, dang?) but since HN has stated they want the forum to stimulate and reward curiosity, having metrics that help posters more clearly measure their posts' controversy levels may be perceived by HN as potentially doing more harm than providing benefit.
If a design choice is perceived as encouraging or helping trolls measure their success at creating controversial posts, or more precisely measuring the effect of their downgrade brigade, it's not a good thing.
Or it may just be a path dependent arbitrary early design/implementation decision and the risk of change outweighs the benefit of tweaking it.
Don't know, sorry, but hope this gives you food for thought if you don't/can't get an official/authoritative answer from HN and are looking for bystander educated guesses.