It's worth noting explicitly here that the zero-point isn't global.
That is, if you are operating at +10 politeness relative to my baseline as a listener, I may be unable to figure out that your "excruciating polite" disapproval is actually disapproval; but the exact same behavior might communicate perfectly clearly to someone else whose baseline is 8 (relative to whom you would be operating at +2) and might be so rude as to inhibit clear communication for Sam down the street, whose baseline is 17 (relative to whom you would be operating at -7).
Recently I've been considering declaring Crocker's Rules. The wiki page and source document don't suggest any particular time limit or training period, and also don't provide any empirical results of testing it, positive or negative. It sounds good in theory, but how does it affect people in the real world?
It seems like an "obviously cool" idea but the risk to one's reputation is worth taking into consideration. If it is clear that the risk is low, and if the value to be gained is clearly very high, we should probably be doing more to encourage it as an explicit norm.
On the other hand, if it is just one of those ideas that sounds better in theory than it is in practice (because the theory does not correctly model reality), or is just yet another signaling game with a net negative value, that is worth knowing as well.
I haven't seen anyone argue against Crocker's Rules or claim it ruined their life, so my estimation is that the risk is low (although there is a small sample size to start with). Also, I have seen at least one statement from lukeprog implying that it has been instrumental in triggering updates during live conversations he has observed, indicating that the value is high (though its causal role is not firmly established in that example).
Does anyone have further data points to add?