in all respect,I dont see single difference between limiting with OZone 5 at -6b headroom and -12db treshold than 0db headroom and -6 db theshold,the mastering devices arent expecting anything,even the most basic units have treshold function so it doesnt matter if you have 48db headroom or no headroom
The Ozone example was a quick example and perhaps I should've elaborated.
If you have your mix hitting 0dB when it hits the master out (pre mastering plugins) and then that near 0db level is getting sent to Ozone with a mastering chain (with EQ boosts, exciters, compression etc) and then top it off with a limiter/maximizer, your mix may sound "loud," but your squashing a great amount of transients and/or peaks. These transients are key in creating a "full" track that can be played on any size audio system and sound consistant (i.e. every sound can be heard evenly). That's the difference and that's why headroom is important.
However, in the end it's all about what sounds good to you. If you're mastering your own track and it sounds good to you, go with it.
Now if we're talking specifically in regards to creating an optimal mixdown to be sent to a mastering engineer, the fact of the matter is they're going to ask for headroom. Check with any professional mastering engineer and they will always give you a range to aim for in terms of your tracks
Max RMS value. Some may tell you not to have anything peak above -3, others may say keep it between -10 and -6; It's all up to them.
The reason is that they're going to color your track with EQs, compression/limiting, saturation, etc. If you're sending them a signal that is peaking at 0dB already, that means they're going to have reduce the volume of your mix prior to even beginning implementing their mastering processes. So right there you've already robbed your track of some transient data. If you give them a mix with headroom, they can get started with their processes and coloring all the transients that are within your track without the need for any reduction.
Furthermore, giving them a tracks that your DAW has show as being near 0dB means an increased chance of some clashing frequencies/low level distortion in the low and mid range frequencies. When the engineer starts to process the track, these levels will be boosted which will amplify some of the low level defects in your track. Again, this effects the quality of the final product. Will it sound good, of course. Engineers can make anything sound good. Will it sound sonically amazing and rich across any and all audio systems -and your listeners? No.
But again it's all up to the individual. To each their own, and whatever works for a producer and sounds best to them, I say go for it.
Cheers.