As others have said, if it was as easy as you say, Frontier would have done something a long time ago. So, either it isn't as easy as you say, or Frontier doesn't actually want to punish combat logging; take your pick.
(Particularly, I think it's a bit of each. I don't think Frontier would issue bans, or other kinds of strong punishment, without being sure false positives can't exist, which means pure statistical analysis and trusting the client isn't enough for them; at the same time, as the forums show, many players seem to consider combat logging to be an acceptable, and even appropriate, response to griefing and even ganking, which might give them pause.)
Finding the right solution is not easy at all, I never said it was.
Frontier can ill-afford to permanently ban their players anyway, so it'll likely never happen - it's just my preferred option. Blizzard has that luxury, Frontier does not.
Most of the things I've said I suspect are already in place, others are definitely in place. I mention them purely because they aren't difficult to implement. Data collection, error handling and logging is not hard stuff to do. As a developer, I'm automatically assuming that a project on the size and scope of ED would have a plethora of error handling, data collection and logging already implemented. I'm working on a project not even close in size and ours is pretty thorough.
Due to the way, I suspect, commander data is collected and stored, false positives can and do exist, it's inevitable - for example, if you and a friend engage in friendly PvP; the data doesn't account for that. So if you quit the game for something (emergency) whilst in friendly-pvp combat, the game records you as quitting whilst in combat.
Determining if someone CL'd is a matter of pattern recognition, data correlation and historic connection checking. It's tedious. It's boring. It's likely very difficult. I never said it was easy, I never said it was 100% accurate. What I did say, is that they could accurately determine if someone CL'd
within an acceptable margin of error. This doesn't make it foolproof, it doesn't mean they won't make mistakes, it doesn't mean they won't get false-positives.
- - - Updated - - -
... I don't even think a message would be needed, ....
Oh, you'd be surprised. xD