Most of this discussion has, naturally, been about statistics and settings.
However, something I need to cover because it’s very important is that flags are not always a binary issue, “Agree with flag because the post needs to go” versus “This user has flagged wrongly and needs to be downgraded,” with no middle ground. I’ll give examples below:
From my own experience, users who flag the most are often pointing up minor things before they become problems, or requesting a check on someone because they spotted word use matching that of a previously-banned miscreant, or maybe they just want mods to look over a topic and see whether it’s in need of some guidance.
The members whose flags are most helpful because of this kind of sensitivity are not always going to have every flag Agreed (and as I understand it, Agreeing a flag has an effect on the flagged user’s progression to TL3, so Agreeing is not an appropriate default action).
The forum I moderate has been actively engaging our TL3 members with the moderation process ever since we moved to Discourse, we have regular conversations by PM and through topics in the Lounge about matters relating to moderation and how the rules are applied etc., Regulars know they can ask questions, query the precise reasons for any decision, and give feedback & suggestions, and this has been extremely successful and well-received.
Having flags cast for the smallest hint of suspicion of a problem has significantly reduced moderation workload, by catching things early, and as a result the forum is sure to have over a dozen people with what amount to mod-lite powers, and the confidence to use them, online 24/7/365. This has kept our forum stable and cohesive, even as our membership expanded rapidly since the move to Discourse.
It’s also highly effective: Regulars getting involved because they’re confident using flags means that we’ve seen would-be trolls and spammers cut off within a few minutes, and because they see the positive effect of being able to engage as guardians of the forum, Regulars have also stepped up in a big way to guide newer users and show them the ropes, which has had a positive effect on the entire culture of the forum.
TL3 Regulars also know they can take or leave this, it’s not a duty they have to fulfil, but they have tools to hand the moment they spot a problem.
It’s nigh-on the perfect system, and it all hinges upon flags being 1. comprehensible and easy to explain, as they used to be, and 2. members, including TL1 & TL2, being encouraged to feel confident that nothing happens as a result of a single flag, so don’t worry about errors, just act when you see anything that concerns you.
Experienced moderators have false positives pretty often (just like the TSA, doctors, and anyone whose role it is to look for problems): we suspect someone, but then that turns out to be incorrect, and members can expect to have more, not only due to inexperience but also because they can’t check IP addresses or see someone’s previous Deleted Posts, so encouraging people to be confident about using flags does require the assurance nothing bad can happen, to anyone, as a result of a single flag.
And finally, flags becoming more of a “do or die” thing, with one of the two users involved risking loss (X% closer to flag purgatory/one-fifth of a strike against TL3 for a year), presents mods with another level of information to bear in mind, and set of decisions to make, when a flag is raised:
- What do I need to do about the flagged post and/or person who made it, which is part of the job, but now also
- Do I Agree this flag against someone who didn’t break a rule, or diminish the weighting of the person who flagged in good faith?
Untangling some contentious topics or disputes can be time-consuming and complex enough anyway, without that added consideration! 
Factor in that many users of forums do not have English as their first language, and may be reading posts by other users halfway round the world for whom English is also not primary, and that adds another reason flags may be raised, with sincere intent, against a user who’s done nothing wrong.
With those complexities in mind, if the following isn’t possible (or desirable):
… please keep Ignore so that neither the user flagged, nor the user who raised the flag, see any change in weighting or progression towards a higher Trust Level. Ignore would maybe more usefully be termed “No Action,” even?