Ty Badge. You meant that early aerials are safe in themselves cause you have protective hit-boxes, and late aerials are safer on shield cause shield-stun and you have a more lenient window to follow up right? Serious question, trying to deeply understand this game and make sure I don't have something wrong.
If you do a late aerial after a shine certain oos options, e.g. grab, can hit you that the opponent would be hit out of by an early aerial. In return you can shine directly afterwards to hit them out of most oos options, which would have been unavoidable after an early aerial.
Of course, which oos options you have to respect depends on many factors, so an attack can be safe on shield with a certain spacing or against one character, but unsafe with different spacing or against another character. E.g. Bowser's or G&W's UpB oos can hit you out of most shield pressure unless you space yourself in a way to avoid this (which is much easier against G&W). Frame advantage/disadvantage and the speed of your own attacks thus doesn't tell you the whole story. Often you can do weaker pressure if you manage to cross your opponent up in shield, because then the option of grab oos isn't available to them anymore.
Oh and not to change the subject too drastically... But how do you buffer a double jump?? Or could you explain in a tiny bit more detail what you were saying about Lucas double jump canceled f-air?
You can buffer jumps for up to 3 frames by trying to tap jump and holding up on the control stick. Buffering a DJ is done by jumping with a jump button and immediately after, while you're still in jumpsquat, hitting up on the control stick with tap jump on.
All I wanted to say with the Lucas example is that it can be really easy to get an aerial perfectly low to the ground with certain setups. I should have posted about the mechanics behind this on the Lucas boards. Usually though, you have to be frame perfect if you want to get all the possible advantage after the attack.
This is another distinction of how good shield pressure is. You shouldn't expect to always be frame perfect, even more so if there's no direct feedback whether you were or not.* The more lenient the execution of shield pressure is, the better.
The examples I gave were such lenient pressure techniques: For Lucas DJC FAir execution doesn't matter that much after a certain point - he just doesn't leaves a larger window before the attack, but the window after it (which we're looking at here) is unaffected. Spacies can get away with being a bit sloppy because every aerial timing has merit and mixing them up by mistake prevents opponents from picking up on habits as long as every important timing is hit sometimes. (of course, if you want to read your opponents timing, you have to be on point).
Another attribute that can affect how good an attack is to use as shield pressure ofc is how high the payout of hitting it is and how easy it is to react to hitting out of doing the shield pressure. Another Lucas example (because that's who I experimented with most): Lucas Magnet->wd back on shield is super safe as magnet is +4 on shield (into jump) and already disjointed, and because Lucas has a great wavedash making it hard to catch him. As magnet knocks the opponent backwards if it hits, it's also really easy to react to a hit once your opponent's percentage is high enough and start a combo/hit a killmove out of even an unexpected hit. Airborn magnet->DJC FAir is another option that both pressures the shield and has Magnet combo into FAir automatically on hit at lower percents. Both of these option give up your position on the opponent's shield, though.
Fox's shine->wd->shine is another good example. If the first shine hits you just initiated a waveshine and if it doesn't you at least got into position to pressure the opponent's shield on the backside of their shield. It's very unsafe frame wise, though.
*Frame perfect techniques like multishines/-magnets can be done with very good consistency easier, because you get direct feedback if you fail. Still, most people then don't do them perfectly, i.e. jump cancel later than possible and thus leave larger windows of opportunity for the opponent to escape than frame data would suggest. Drilling techniques without clear feedback to frame perfect consistency is much harder as the only ways to tell wether you messed up are training & experimenting until you're certain you hit the soonest possible timing or to have your opponent punish you for not hitting it.
If you're interested in how shield pressure works, these threads and/or posts might also interest you:
1 (Melee Fox) 2 (3.5 Lucas) 3 (2.5 Lucas) 4 (3.0 Wolf)