The real problem is that every case is different!
Sometimes the action required is obvious, other times it is on a knife edge.
And this is precisely why relatively precise policy can help. In the "obvious" case, the policy will agree with common sense ("obviousness"). In the "knife edge" case, it will make for a relatively ready, repeatable and predictable decision - saving you decision-making time, which, I gather, is one of your priorities, as well as giving posters foreknowledge as to when and why they will be banned if they do this or that - rather than the current situation, which is that people aren't quite sure
what will get them banned, especially in the light of recent temp-bannings by Alex (in particular that of Malf), which have been, to many of us, inexplicably rash and intemperate.
You have to balance how rude someone is to other posters, whether they say anything interesting, whether they just repeat themselves, whether they seem to respond to others, whether they are disruptive, whether they wish to discuss intergenerational sex, how long they have been on the forum, how many, if any people have complained
Now you're talking specifics. Nice! Granted, there will always - as with any behavioural code interpreted by humans - be room for judgement, but, despite that, these separate points that you raise
could be developed into a coherent and reliable policy, and it needn't be terribly verbose. The point of it would be for people to be able to predict
when a moderator will take an action, and
why. If (when) deficiencies are found, they could be corrected.
Those (knowing when and why moderation action would be taken) are not so unreasonable a couple of criteria to be provided, surely, and we needn't begrudge forum members and potential forum members that clarity on fear of turning ourselves into artificially intelligent automatons, surely?