I would say that the first one is according to the normal CR (because lowering the CR just because the party was smart is stupid and no one here claims otherwise I hope) whilst the second one deserves a +1 to it.
I consider them both exactly the same CR. Intelligence is its own reward, and stupidity its own punishment. So if you jump them it's easier, and if they jump you it's harder. Same reward, just the better you play the easier you get it. Doing otherwise just leads to people doing a Leeroy Jenkins intentionally because it's more rewarding than playing smart.
You do have a very good point in that they should just fly over if they can. I usually play lower levels (1-8) so having an entire party capable of flight isn't my normal games. I'm not trying to make a point about the game in it's entirety based on low levels, but the hobgoblin bridge example was simply the first one that came to mind. I'm sure there's a suitable mid-to-high level scenario in the same vein, I'm just not used enough to that levels and the assumed capabilities of a party at said level to provide with such a scenario.
As long as even one person can fly, it works out fine. Consider a level 3 party. Alter Self, to take the form of a flier. Wizard flies across, Benign Transpositions with someone else, and repeat. A party lower level than 3 wouldn't be able to take them all on regardless. Having a wand of the latter isn't that hard at this point and it has plenty of uses outside of this scenario.
As for not a threat if they can only get you halfway to death, I disagree.
If the party can not lock down the entire opposing force with BFC and crippling debuffs in 1-2 rounds (which is often the case when I DM), having enemies that will start killing of your party members in 2-3 rounds is a threat.
If it takes them that long, then they are not threats, plain and simple. You can either take them out with save or loses or by damage, but if it's been three rounds and you're still fighting the same things, expect your dead to join theirs.
If we look at it this way:
You claim that level appropriate encounters kills players in 1-2 rounds, am I correct?
If that is the case (which it might fully well be, it's not my experience but I'm sure there's plenty of situations that is like that) then you have to have damage dealers that can deal high amounts of damage. Because it's the only way to contribute good enough, fast enough. If you don't put down the hurt, you're dead very soon.
Now, if we say that Mister Monster have a Danger level of 10 when it comes to killing people. That means that Sir Swordalot must be at least level, let's say 7 to be able to deal damage high enough to not be a liability.
Yes. Either to land a save or lose of their own, or just kill via damage. And by land a save or lose of their own, I mean on everyone. Most of them hit an area, at least the good ones do and unless you're taking the proper steps to buff your saves the first hits 50-75% of your party and the second hits everyone else, negating them for the combat (at least). Damage is for one person, but you know dead is dead.
If you are a PC, and you're stuck doing damage yes you do have to do a lot of it, as only the last HP means anything.
If we compare this to situations which I would prefer and have more experience with, when it takes more than 1-2 rounds for the opponents to kill someone. Let's say Mister Monster have a Danger level of 7. Sir Swordalot must be, let's say 4 to be contributing.
That doesn't describe actual D&D though.
I do not like games where your death is 1 round away unless you do the perfect action on a regular basis (unless it's a boss fight, then it's ok). Neither do my players. So we avoid such scenarios and that's the way we play the game. It works for us.
One round kills are rare unless it's a strong enemy, you have a weak build, or both. Two are exceptionally common, the leading cause of death even. First takes you to 30% or so, second finishes you off. It's also rare you have to perform the perfect action, unless stuck with a perfection is mediocrity build, then you have no choice. You do have to perform effective actions though. Effective in the sense of can be effective. So if you cast Slow and they save, that's fine, you couldn't have predicted that. If you're playing around moving enemies around on a non Dungeoncrasher, that's not fine, you're wasting actions.
I won't say whether or not you're correct in drawing the line at a certain standard, because I have not enough experience or data to have a definite opinion on it. It's just that in my experience the 1-2 rounds before death usually isn't the standard. If it is because I choose too unoptimized monsters, give my players too optimized characters or you're wrong somewhere I don't know.
If it means that I'm pulling punches as a DM, so be it. It makes the group happy and makes it better for us. I might be. I don't think I am, but it's possible.
It's most likely because you're pulling punches. Just about every enemy out there has a small chance of one rounding level appropriate PCs and a near 100% chance of doing it in 2. You'd have to be doing some very heavy countermeasures to offset this, and then those tend to reduce the chance of being 2 rounded instead of making it so you last longer. That does make a difference, as say... going to a 50% chance that a given round of actions will work on you still means you get two rounded 1 time in 4. Since there are many monsters and one of you, that sort of thing tends to just breed complacency.
The line you draw you see as objective (the math supports it from your POV thus it becomes objective) whilst others here see it as subjective (for a reason or another). I do not have enough D&D-skills to say where the line should be drawn (nor to accept or refute the math you claim), I only know where I draw it myself because it works for both me and my players.
One thing I've found here is that despite this being an optimization board, most here aren't really cut out for that at all. They might think that they are optimizing, because they are making something better while ignoring that the goal is not to make something better, it is to make it good enough. Having 2 dollars is more money than having 1 dollar. You still do not have enough cash to make any significant purchases. Now for this reason they adopt mindsets that have no business being held by any serious player, such as claiming the DM should adjust encounters (which defeats the entire point of optimization, as you could just make whatever terrible character you want and it work just as well as someone that actually tries). The core of that is dismissing objective things as if they are subjective. So yes, others see it as subjective. That isn't a valid argument though.
veekie: 25-50% would mean something like say... Anywhere from 3 to 13, as a level 3 character. the first being 25% vs the standard array on a Wizard and the second being 50% vs the standard array on a Fighter. Well, that's nothing. Lower level enemies hit harder than that. Even stock encounters pull off 14-16 on average quite easily. That's a lot closer to 70% than 25-50%.
