Unstable 2.0 doesn't seem like it'd fly at all.
I think he meant OR showed 2.0 stable and yet in real life it was not adequate.
A thing to consider is calibers of stability is determining a length in units of the rocket's width (diameter). Width is one measure of the rocket's size but not the only one. It does affect stability but possibly only 1/10 cal. of margin is caused by width, the rest is caused by length. So a rocket that is relatively fat doesn't need as large a margin as a portion of its width, but a longer one compared to its length. If you somehow added something to the rocket that didn't affect the stability at all, but increased what OR considered to be the diameter, that magnitude of the stability in inches would be the same but in calibers the magnitude would be reduced.
One unrelated reason for using calibers is, if you know the CP, and can balance the rocket for CG, it's easy to eyeball about how many caliber units that is in the field.
There's a feature in OpenRocket that I really like...you can adjust the Angle of Attack (which would be zero with no wind and increases with wind speed) and watch what it does to the calculated CP position. This can alert you to situations where the rocket is stable in zero wind but could go neutral or negative with an unexpected gust.
Yes, this is important. It can help verify a shorter rocket is more stable than you might have thought. On the other hand, with some long rockets you have to compromise between high stability margin at low AoA and becoming unstable at not that much higher AoA.
There are some cases, mainly saucer-style rockets, where base drag is considered a stabilizing force. I am skeptical that it is significant with other rockets. As always, the limit of what should be tried depends slightly on how large and dangerous a rocket ship it is.