I often use the surprise roll to *tell me* who's aware/alert/ready and-or who isn't in situations when an encounter is inevitable. Way quicker and easier than a bunch of perception checks.
Okay, to be honest the third possibility of
might be aware hadn't occurred to me due to the way Gygax lists certain factors as inevitably leading to the negation of surprise, i.e.
making one side aware of the other. Apart from magical detection, he mentions noise and light as giving away the presence of a party, rendering a surprise roll for the monster, or vice versa, unnecessary. The implication seems to be that if the party, or the monster, is not trying to be stealthy, then there is no need to determine surprise, because the parties are assumed to be otherwise aware of each other. In this case the surprise roll is very much like the 5E determination of surprise, using Stealth v. Perception. The difference is that the surprise roll has nothing to do with anyone's skill at sneaking. Thieves, for example, as far as I know, don't have the ability to directly affect a surprise roll, I assume because we're talking about entire sides being surprised. Rangers and elves, however, are less likely to be surprised, and rangers in particular translate that ability to the rest of the group. So, just working through this, it seems like the surprise roll is more like a Perception check that both sides are making against a fixed DC, rather than anyone rolling for Stealth. The DM's decision, then, seems to be between certainty that a side was already aware of another (no roll) and uncertainty whether they were aware or unaware (surprise roll). I can see how that's the same as the roll in fact
determining whether one side is aware of the other in spite of the other side's attempt to be stealthy, but at the same time it suggests to me that awareness is binary. Either you are aware or it is uncertain is pretty much the same as you are aware or you are unaware. If the surprise roll is determining level of awareness out of uncertainty it seems strange that it is determining
how unaware, and that it would be expressed as a higher or lower degree of unreadiness, i.e. a shorter or longer period of surprise. Another thing to consider is the effect of the Dexterity Reaction bonus or penalty.
What if nobody's trying to be stealthy. One group walks around a corner while another group approaches the same corner going the other way. Both groups *could* be surprised, or just one, or neither; and in any situation this needs to be determined before initiative gets rolled.
This is where I had a problem with your blanket statement that all potentially hostile encounters begin with a surprise roll, because this would seem to fall under Gygax's calling out noise as a factor that would negate the possibility of surprise. Of course, this is also the same in 5E; if no one is trying to be stealthy, no one will be surprised.
Point taken.
No, you still have two systems in 5e as you're ignoring the perception checks (and stealth checks, where appropriate) before initiative starts.
I see your point, but in that case part of the function of an initiative roll after surprise has been determined should be seen as the equivalent of the ability of the Dexterity Reaction modifier to mitigate surprise.
Also, in 5e as written you've got the absurd possibility of a person/group/whatever being taken by surprise in the fiction yet still able to act first in the mechanical resolution due to initiative luck, which is flat-out wrong.
I'm not sure how this can happen since surprised creatures are unable to move or act until the end of their turn, at which point although they can take reactions, any other activity must wait until the next round. The similarity of this mechanic to the function of the Dexterity bonus is striking, however. What's unclear from the 1E DMG to my mind is what happens when, for example, a character is surprised for two segments according to the surprise roll, but has a Dexterity Reaction bonus (let's say +3) that offsets the surprise segments. Is that character then able to act during the surprise round, or are her opponents simply unable to attack her? From the text it seems unclear except that obviously her side still loses initiative.
By the way, 1e initiative as written is a hopeless mess; and friends steer friends away from it. 1e surprise mechanics, however, are nice, simple, and worth keeping.
Just out of curiosity, what kind of issues do you see with 1E initiative? Is it side initiative, breaking up multiple attack routines, or something else? It seems pretty straight forward to me, but I'm interested to know why you wouldn't recommend it.