Trying to understand 3.5e cover rules - visual help

[Bobur]

Hi all, I DM 3.5e games for years now, but the cover rules keep confusing me every now and then.

Edit: I was trying to link to an IMG on reddit, didnt work and cant upload here so a link will have to do:
Scroll down a little to see the image.
https://www.reddit.com/r/DungeonsAnd...sual_learning/

So I collected a few scenarios that came up, in the hope of clearing these questions up.

Cover for melee: If any line from you to the target is blocked it recieves cover (+4AC) Creatures can give cover. So in a setting like this with a 3rd creature standing between the dwarf and bugbear, does that create cover?

Melee or reach? Melee and reach/ranged work differently to determin cover. For melee ALL lines have to be clear to avoid cover, where a reach attack is considered a ranged attack for cover. so you pick one corner to shoot from and a target square and those have to be clear.

So its important to understand what it is.
Does a large+ creature, that doesnt have a reach weapon but still can attack 10+ ft away, still counts as a melee attack for cover purposes or as ranged/reach? Common example, our monk grows to large size and can now attack 10ft away.

1/2/3) Similar to 1) - Do the bugbears in 2) and 3) give cover against the ogres attacks?
If 1) doesnt grant cover, then it would be important to determin again if the ogres attacks count as melee or ranged for cover.

4) Corners and large creatures
For attacks as well as when the large creatue attacks it basically just picks a corner to attack from to determin cover. So the dwarf has a clear melee attack onto the ogre and the other way around.
The ranger can attack the ogre too without cover, but since the field isnt 10ft away from the dwarf he will still have to deal with shooting into melee. (-4Atk)

Lets say the ogre in this example has a reach weapon and could attack the ranger. But for some reason doesnt pick the nearest corner to attack from, would that change his range? Like if he picked the bottom right corner instead to avoid potential cover problems.

5) Ranged: Worst case. Shooting into melee AND cover from the dwarf. (-4Atk /+4AC)
Edit: If the ranger picks the upper corner to attack from he can at least avoid cover. ^^

6) But if he steps aside, he can avoid the cover penalty. In this scenario he can attack the back of the orge, which is now 10ft away from the nearest ally and therefore also avoids the shooting into melee penalty. Correct?

7) A wizard tries to avoid the penalty on ranged attacks vs prone targets by moving into melee. Technically the prone Bugbear still recieves the +4AC no matter where the mage is, but it feels strange if he stand right on top of him.

Thanks, I hope this can clear my confused mind up a bit.

[Darg]

1) no, creatures only provide cover for attacks against non-adjacent creatures.

2) reach if beyond 5 ft, normal melee within 5 ft

3) yes, reach weapons use ranged attack rules for cover.

4) no cover bonus to AC. As a bonus, the rule is if any part of the creature is at least 10 ft away from another creature that you don't have the shoot into melee penalty.

If your target (or the part of your target you’re aiming at, if it’s a big target) is at least 10 feet away from the nearest friendly character, you can avoid the -4 penalty, even if the creature you’re aiming at is engaged in melee with a friendly character.

5) no shoot into melee penalty, but yes cover bonus.

6) no penalty and no bonus

7) ranged attacks in melee are still ranged attacks so still recieve the prone penalty. It is strange, but on the other hand the wizard could have chosen a melee touch attack spell and gotten a +4 bonus instead or moved away without provoking an AoO from movement or casting the spell.

[Bobur]

1) no, creatures only provide cover for attacks against non-adjacent creatures.


4) no cover bonus to AC. As a bonus, the rule is if any part of the creature is at least 10 ft away from another creature that you don't have the shoot into melee penalty.



5) no shoot into melee penalty, but yes cover bonus.

Ah, my measurement was off. ^^

But the way I think this works is correct, yes? If it says, "the part of the target has to be 10ft away" you say which square you are aiming at and see if that part has cover.
You cant stand right in front of a dragon and just say "his butt his more than 10ft away from the next melee" but not being able to shoot at that square, correct?

[Darg]

Ah, my measurement was off. ^^

But the way I think this works is correct, yes? If it says, "the part of the target has to be 10ft away" you say which square you are aiming at and see if that part has cover.
You cant stand right in front of a dragon and just say "his butt his more than 10ft away from the next melee" but not being able to shoot at that square, correct?

That depends on the DM. With the rules as they are written you can shoot through squares occupied by a creature that doesn't take up the whole square (gelatinous cube), but suffer the cover rules. You have LoS and LoE. Occupied squares are an abstraction to account for the area a creature uses within combat, not the volume of space a creature occupies (other than exceptions like the cube). In combat you have forward and backward, side to side movements within squares. When you shoot at a back square you are trying to hit the creature who could have moved back to dodge a sword thrust or in the case of the dragon lifted its tail into view to attack with.

[Fizban]

1: All these creatures are adjacent to each other, so their melee attacks ignore cover, unless there's a wall in the square where you put the arrow. If the orc? or dwarf decided to make a ranged attack against each other, then the elf would provide soft cover.

2: The "melee" rule only applies to adjacent creatures, which means in the adjacent square, 5' away. The ogre is not attacking an adjacent foe, therefore its (non-adjacent) melee attacks are subject to normal cover. The ogre can choose its best corner to attack from, but the dwarf can also effectively "choose" its best corner to "defend" from*, so the bugbear provides cover- even if the dwarf were one square north or east of its current position. If the bugbear were one square north or east, the ogre would have a clear line of attack. The dwarf's ranged and reach attacks against the ogre will similarly go through cover from the bugbear. . .

Depending on how you read "target's square." The problem is that large creatures don't take up a square: they take up four squares. . . or rather, four squares on the ground and also the next layer of four squares above that (because they're a 10' cube)- actually, in 3.0 they might not even be a square, since long rectangles were used for long creatures (such as most quadrupeds, including horses). The way the attacking rules actually work, is that you target a single square, and if a creature is in that square (and you pass any miss chance roll), your attack roll is applied to that creature**. So often a medium creature can target an upper/side square of a Large or larger foe without going through cover, even when that Large+ foe would suffer cover attacking them back.

Is this the intent? The FAQs are silent on the issue. It sure seems reasonable to me that smaller creatures have an easier time hitting larger creatures around obstacles- while it might seem fine at first with just Large creatures, it will quickly become ridiculous as Huge, Gargantuan, or Colossal creatures claim that the Human or even Housecat somewhere generally in front of them creates a massive cone of cover that the entire rest of their body sticks out from. While a large creature does not take up a full 10' wide square, it does take up more than a full 5' square, so even when it chooses its best corner to defend from there is still body sticking into the other squares, which can be attacked without going through cover. This is also supported by the "shooting into melee" rule, which clearly states that you target specific portions of larger creatures and can deliberately choose them to avoid that penalty.

*This is why it says you need a clear line to all corners of the target square, because creatures move around within their squares and can thus be in whichever corner is best for them. It's much easier to think of when you consider from the target's perspective, that they only need to find one corner of their square that will grant them cover against their foe's best corner.

**If there are multiple creatures in that square due to grappling, a ranged attack hits one of them at random.

3: Yes, and there are more bugbears than necessary- even just the one in the square directly east of the dwarf provides cover to the dwarf at all angles, as does the one to their northeast. Only if you remove the bugbears in the squares both east and northeast of the dwarf (so only the one two squares east remains), does the ogre have a clear line of attack- with a ranged weapon from its rear squares. If I've got the large+reach weapon range right (going by the template in the back of the DMG), it looks like it does need to attack from that one specific square with its reach weapon to be in range(why people go making things more complicated, large creatures don't need reach weapons. . . ), so for the reach weapon any of those bugbears provides cover.

You've also drawn the ogre's attacks from the least advantageous corner- it doesn't matter when all those bugbears are there obviously blocking all lines of attack, but if there were fewer, they might have a clear line- and that line would be from anywhere except the furthest forward point you've chosen. The attacker is always trying to get around cover from the side. It turns out that for the reach weapon it's still not good enough even when drawn from one of the "side" corners of the square it's attacking from (directly in the center of its overall 10' side), but by putting lines on the picture you immediately lock perception into thinking those are the correct lines, when the question you're asking most directly hinges on whether or not they even are the correct lines.

Ideally we'd have a whiteboard and you would draw the positions and we would draw the lines (and then recheck three times to make sure we aren't talking out of our rears).

4: I'm not sure if it's ever explicitly stated anywhere, but if the corner you're picking to attack from in order to avoid cover is out of your attack range, I would certainly rule that you can't avoid cover for that attack. As for the ranger, here we have the Large/square problem I just mentioned: you've clearly drawn lines and suggest that the ranger can attack the ogre without cover by targeting the square of their choice, but there is a possible reading where the ogre can collapse its body down to a single point to take cover behind the dwarf and even that bit of wall there- a very silly reading, but constantly repeating "square" for everything can confuse things eventually.

5: There's literally no reason for the ranger to draw their attack line from the southeast corner of their square, since being only a single square themselves there is no difference in range and it would cause them to go through cover. Drawing the line on a picture creates an immediately misleading example, because they shouldn't have even bothered thinking about that line.

Edit: it seems there's a secondary question about shooting straight through the ogre's front squares at its rear to avoid penalties? Creatures only provide soft cover and do not provide it to themselves, but it's a basic common sense ruling that you can't just shoot "through" their front to an arbitrary square in the back. That said, you can also make a common sense judgement that it's not that silly because creatures that big aren't actually moving their bodies very much, relatively speaking, so the shooting into melee penalty not applying most of the time thanks to "shooting them in the butt from the front" isn't a problem. But you don't even need to make a ruling- you can just shoot up, because anything that's 15' tall fighting a 5' tall creature on the ground, always has a square on its front that is 10' away from the "melee" (this also deals with many cover problems). And while the ogre is only 10' tall in size, you keep using examples where the dwarf is also 10' away, so there will be one square on that front facing which is far enough away to avoid the penalty.

6: The problem here is what "X feet away" means, because sometimes it sounds like X feet between the squares, but that's not actually how it works. Adjacent is adjacent, 10' reach is 10' reach, but there's only a 5' square between two targets fighting at 10' reach. Adjacent is thus 5' away. In the picture provided, the dwarf is already far enough away that they're not threatening the ogre (unless they have a reach weapon), but the ogre's reach allows it to threaten the dwarf, so we must presume that the squares from which it does threaten the dwarf are "in melee." Shooting at the rear squares avoids the penalty, yes.

Edit: or as mentioned, the high square on the ogre's front which is furthest from the dwarf also avoids the penalty.

7: Sure, it's perfectly reasonable to rule that if you're adjacent to a prone character they don't count as prone for ranged attacks. They still get their AoOs, albeit at a -4 penalty for prone, so it's not the safest idea.

[Bobur]

@Fizban:

"but the dwarf can also effectively "choose" its best corner to "defend" from*"

I am not sure that I recognize that.
Where is that rule written? That a creature picks a square to defend from and what does it do in practice?

It sounds like the target can just undo the choice of a ranged attacker by forcing a bad angle unto him for cover, but I dont remember reading something like that nor have I heard of it.

-------


*This is why it says you need a clear line to all corners of the target square, because creatures move around within their squares and can thus be in whichever corner is best for them. It's much easier to think of when you consider from the target's perspective, that they only need to find one corner of their square that will grant them cover against their foe's best corner.

Should I think of it more a a cube then and you need a clear line to all corners? Adding the 3rd dimension makes it tricky because it blurs the line between thinking of it in a 2d way and which corners to check vs a 3d way and checking sides of a cube.


God, I really hate 3.5 cover rules.

[Darg]

"but the dwarf can also effectively "choose" its best corner to "defend" from*"

I am not sure that I recognize that.
Where is that rule written? That a creature picks a square to defend from and what does it do in practice?

The rule is that the creatures can pick any corner to attack from, not defend from. Easy thing to get mixed up, but rules wise is different.

Should I think of it more a a cube then and you need a clear line to all corners? Adding the 3rd dimension makes it tricky because it blurs the line between thinking of it in a 2d way and which corners to check vs a 3d way and checking sides of a cube.


God, I really hate 3.5 cover rules.

3.5 has flight rules, but nothing about the game being 3d. Which is kind of dysfuctional.

But it's probably best just to assume everything is a cube the size of a creature's space. The rules just work better that way. Similar to a video game hitbox.

[Fizban]

@Fizban:

"but the dwarf can also effectively "choose" its best corner to "defend" from*"

I am not sure that I recognize that.
Where is that rule written? That a creature picks a square to defend from and what does it do in practice?

It sounds like the target can just undo the choice of a ranged attacker by forcing a bad angle unto him for cover, but I dont remember reading something like that nor have I heard of it.

Edit: Ah, on a second read the part that you actually got confused on was the defender picking a "square" to defend from. That's not how it works: the attacker picks the square they target, and for cover purposes, the defender effectively "chooses" the corner of that square (because if any corner of that square has cover, the attack counts as going through cover, so the defender effectively "chooses" the best corner). The fact that creatures move around in their spaces and squares is why you can "lean" to make your attack from the corner of your choice, and why you must check for cover on all corners of the targeted square: because the defender could "lean" into the most defended corner (and hey look, it's so much easier to understand if you just write it that way in the first place). The attacker always gets to choose which square they target for their attack.

An important distinction to make is the difference between "square" and "space". I may not have even been perfectly clear everywhere above, but square should only ever refer to 5' square, while space should be used when one is referring to a creature's whole space. No matter your size, you always attack from a square that is within your space, to a square that is within the foe's space- but many of the rules are written from a perspective of standard Medium PCs where the two terms are often interchangeable and so they'll just say one or the other rather than working through the full thing.
(end edit)

The part you quoted is where I explained it.

Directing the attacker to pick a corner to attack from, and then evaluate it against every corner of the target's square, and if any of those lines intersects it counts as cover, is the same thing as if the defender was actively picking the best one out of those possible corners to "defend" from. The attacker never picks a corner of the defender's square to attack, they must evaluate all possible corners and "take" the one that gives cover to the defender if there is one. The attacker can pick the square they wish to target, which for large creatures means the specific 5' square within that target's greater space, and the corner of their own square to attack from, but that's it.

So yes the defender can "undo" an attackers choice, in the sense the ranged attacker never got to choose the corner of the square they wanted to attack in the first place. They pick the square to target and the corner of their own square from which they attack, and then if there is any possible line that gives cover, the defender gets cover, whether the attacker likes it or not.

I've said it back and forth several times now in case it will click when read slightly differently, but the point remains that it's far easier to think of it as the defender picking the corner of the specific square in their space that is being attacked (which for large creatures can be any square within their space that the attacker has chosen to target), because that way you're only thinking about one line. The line which starts at the corner of the attacker's square which is best for the attacker, and ends at the corner of the defender's square which is best for the defender.

Because that's how it actually works in practice, rather than getting confused thinking about lines that don't matter. If the attacker has one square, and they're attacking a specific square, there is only one actual line that matters, and that line either does or does not go through cover. The only way they get a "benefit" from picking a different corner to start their attack line, is if they were picking the wrong one in the first place and just now realized where they should have been starting from. And the defender gets the best possible result out of every possible corner in the square that is being attacked, regardless of what corner of the attacker's square it's starting from. There is only one line that matters: the attacker's best, to the defender's best. The only choice a Medium attacker actually makes is which square they're targeting, the "corner choices" are all automatic.

In short, they've written the cover rules to be carried out entirely by the attacker, which in some ways was a poor decision. It avoids questions of awareness or flat-footedness, but means that the attacker has to think about more possible lines of attack and makes it seem like they have more of a choice than they do, when phrasing it as "to the defender's best corner of the targeted square" would have been much simpler.

Should I think of it more a a cube then and you need a clear line to all corners? Adding the 3rd dimension makes it tricky because it blurs the line between thinking of it in a 2d way and which corners to check vs a 3d way and checking sides of a cube.

The game is always in 3d, except when it doesn't matter, the same way normal physics always applies, except when it doesn't. If shooting up gives you a clear shot, it matters enough. (If you want to make things confusing then start thinking about what squares a creature's "vitals" are in for purposes of sneak attack, oh look suddenly TWF blender rogues don't work on tall creatures). This is also very useful for "adjusting" the areas of your spells: you can point a Lightning Bolt at the ground (or into the air) so that it doesn't go past a certain square, or target your Fireball a certain distance above the ground so that it only dips low enough to hit a smaller area on the ground.

The funny part is that the flying rules are actually more 2d, because they're written as ways to make elevation changes during 2d movement for creatures that have to maintain a minimum fly speed in order to stay airborne, rather than as point to point movement. Most things you can think in terms of direct range, but the different grades of flight and their up/down efficiency and turning radius means for weaker flight you have to rigidly work from a 2d+elevation perspective.

God, I really hate 3.5 cover rules.

It shouldn't be that hard, that's why I'm giving you the practical phrasing: "Attacker chooses a 5' square to attack from, and a 5' square to aim their attack at. Draw a line from attacker's best corner to defender's best corner, and if it goes through or intersects a border with a square that grants cover, the defender has cover."

Low cover and adjacent melee attacks are the things that get weird, since low walls only provide cover if you're within 30' of the cover and you need to check distance of both the attacker and the defender to said cover, unless you're in melee in which case it counts the same as a wall. Meanwhile adjacent melee flips it around and says that you effectively don't get to choose the corner of your square to attack from, because if any line goes through or intersects cover it counts, but only if that cover is a wall.

So use the practical phrasing: "Adjacent melee attacks never gain soft cover from creatures, but are more susceptible to walls: instead of picking the best corner of their square to attack from, if the attacker's square is also adjacent to a wall that is waist high or higher, they must choose the worst corner of their square to attack from."

Basically there's two almost completely different cover systems going on: adjacent melee cover from walls, and everything else. But it's written as a mish-mash rather than as two distinct systems, directing huge swathes of "melee" attacks to the "ranged" cover rules, instead of just writing the two separately the way they actually work.

And most of the complexity comes in when you start having multiple creatures of varying sizes and then giving them reach weapons. You can simply not do that and avoid a lot of confusion: don't have PCs larger than Medium or the occasional Large (long) creature, don't have reach weapons on big creatures.


The benefit of wrapping your head around the cover rules, is that they're kindof massively critical to the basic tactics of the game. So much of char-op's whining about big/reach/ranged creatures just "walking around" stop working when you apply cover correctly, even moreso when the PCs make intentional use of it. Big creatures are easy to hit around cover and without triggering the shoot-into-melee penalty, but have a much harder time getting around even just single squares of cover that Medium PCs are using (as long as the cover is tall enough). That's a massive part of how you deal with their oversized attack bonuses and damage: put the tank in front, and stay behind them.

[Bobur]

@Fizban:

Thanks, I think I get what you meant.
It confused be because it sounded like a rule that I havent read yet, where it is just a different way to phrase it. ^^

Sadly, avoiding big creatures and reach isnt an option for me, since becoming large is our monks no1 go-to tactic. But with these cover rules, maybe my minions can hold out against him slightly better. ^^

Edit: I added a different image, with new examples. Only the last one is a bit uncertain for me.
The wizard (hat) can range attack the large creature near the corner easily.
If the large creature uses a ranged attack too, the fighter (shield) would create cover for the wiz since one corner (the one further back) is blocked by him. Correct?
So far thats the only example where it is easier for the smaller ones to get cover. In many cases, if one has cover, the other usually does too.
- What confuses me here is that the best angle for the defender is still a straight line. Which easily connects to his front corner, but he shares the edge with the fighter, so does the straight line connect to the corner further back or does it get blocked? My assumption is that it gets blocked.


So:
When I think in 3d.

I now imagine a fighter with a reach weapon, in the example of no 2). So the dwarf has a reach weapon, there are bugbears in front of him and a large + creature behind them.
Applying your way: I would start with the attacker (dwarf) trying to use his weapon from a top square and attack a top square of the ogre. Thinking that you lift your weapon above their heads to poke at the big guy.
That technically works.
But now the ogre can use his best corner of that square which is one of the lower corners, forcing cover (the bugbears) onto the attack. You can imagine it as the ogre lowering his head behind his allies.

Turning that scenario around: The ogre chooses his top square to attack from in the hope of reaching over the bugbears. Looking at it in a 3d way, the dwarf now chooses his best angle, which would be one of his lower corners (0ft hight) so he basically takes cover behind the bugbear in front of him.

[Fizban]

Nice drawings

The first two drawings on the left with arrows from left to right should start from the bottom left corner of the single square, since that corner is further to the side and will give the best angle (it doesn't matter in this example, but one could be drawn in which the front corner you're using wouldn't work, but the back corner would, and since they're all corners of the same single square there's no difference in range). But in both cases the defender would "choose" the very center or center left points of their space (the top right or top left corners of the targeted square), and thus have cover. Which I think is what you're trying to show in the second one with the red arrow- I'd have put an X over the whole drawing with the invalid green line example, and then use just the red line to show cover is active in the second example.

The middle two drawings have the large attacker using their best starting corner, and the red line where the Medium on the left has cover is the one that would be used, yes.

On the third pair of drawings:
-The top right drawing's (reach) spear guy would want to use his top left corner, to attack the Large foe's top left square, but once again it doesn't help because the Large defender will "choose" either the lower left or lower right corners of that square (the center left or center points of their overall space, either of which makes the line intersect with the side of a cover space), which allows them to gain cover from the person between the spear guy and the large guy.
-The lower right drawing's red line is correct: the Large creature is attacking from their best corner, and the defending spear guy has their best corner.

(You're still drawing lots of lines, effectively illustrating the way the rule is written to check all attack lines from the attacker's corner to all of the defender's "possible" corners, but indeed this shows why that phrasing is confusing, because you end up with a bunch of "green" lines that suggest the attack should work, when even one red line means that none of the green lines exist. If you're having trouble immediately intuiting the best to best starting with the attacker, maybe try starting with the defender instead, since you seem to be finding the defensive points easily enough. Remember, the system is not trying to make an attack work- the system is trying to make cover work, and it's only if there is no possible way for cover to work, that the attacker gets to avoid cover. That's why it's easier to think of it as defender's best, and why writing the rules so that the attacker is meant to check for cover was a bad idea.)

So yes, when there's a person directly between you and the monster, both you and the monster gain cover. And with Large creatures adjacent to Medium creatures it's hard to avoid that. But with Huge creatures, and when the Large creatures/Medium with reach are not adjacent (using that reach), and with the Medium creatures standing both behind and with some distance from their Medium meashields, it gets easier to make clear shots for them while the big creatures are stuck going through the meatshield as cover. Particularly when the bigger creature has to make their attack from a specific space along their side due to the exact range of their reach, while the ranged Mediums can shoot at different parts. Naturally, if you're attacking the same square that the target is attacking from, cover for one is going to be cover for the other, exactly the same as if it were two single-square Medium creatures. But bigger creatures have more than one square to target.

Ah, there's the hat in the very last set of pictures. You have the wizard attacking from the correct corner, but it's hard to tell what's going on near the Large target because there's no grid under the rest and there's a gap between the wall and the target . Assuming the intent is that there is cover in the square directly left of the top left square of the target's space, and that the wizard must specifically target the lower left square of that space due to range, then I think they should still be okay as the defender uses the top left corner of the targeted square (the center left of their space, the upper green line which lacks an arrowhead). This is because the cover check looks for the line passing through or intersecting sides, but not points, and if I'm interpreting the map correctly the cover meets it at a point there.

For the second part, you have the Large creature attacking from the wrong corner of its space (unless it is limited by range ie: reach), but yes, it might seem a little weird at first, but the wizard can claim cover there. If you reorient and think from the Large creature's (let's say Ogre) point of view: looking straight ahead, there's a meatshield slightly left of center, and a wizard slightly left of that, but if the wizard leans a little away and to the Ogre's right ("choosing" their top left corner) and the meatshield a little to the ogre's left, they're lined up enough that the meatshield is in the way. And the meatshield doesn't have to be an ally doing that on purpose, because for game purposes they're always moving around in their space so there's always some amount of time where the wizard can lean over for cover: enough that they can get a +4 bonus.

This is because the two are off-center from the ogre: if they were centered straight across, the ogre would use their outside front corner and the wizard their inside back corner, the line of which would not intersect the side of the fighter's square. And this might seem weird because putting them off center doesn't change anything, the Ogre should just turn to the left, right? But the simplicity of the grid can be misleading there: by putting the two Mediums adjacent to each other and off-center from the straight grid line from the Ogre, if the ogre just "turns" and we reoriented the grid based on that, then the two Mediums would no longer be "flat" next to each other: the wizard on the outside edge is actually standing slightly farther away from the ogre and slightly behind the meatshield (this is intuitively obvious when you exit cover calculations and just think about the distance and angle of lines from center of creature space to center of creature space: yes the wizard is obviously slightly further away, the line is longer). It might seem very slight , but when you account for moving around in a combat space it blurs into more overlap. And "reorienting" the grid this way would actually require you to select squares for the fighter/wiz on the new grid, which would clarify how far back the one wants to stand from the other (which their players would be picking, presumably in such a way as to preserve the cover they want).

Now consider how often you end up with a pair Medium creatures specifically lined up centered side by side facing straight across from the Large ones (well, actually probably pretty often if you've got straight 10' wide hallways). And how it's impossible for two Medium creatures to be centered straight across from a Huge creature. And how if one of the Medium creatures just stands one square further back they get a whole bunch more "coverage". Thus, it's not hard to get your squishy behind a meatshield to claim some cover, though depending on the size of the target and the range of the squishy they might still suffer cover themselves.

With forced 3d:
Yup, you've got it right. With Large and Medium creatures that close together, and even for ranged attacks at varying distances when everyone is all standing on the same level ground, forcing 3d doesn't really change anything, people between you and the target are going to provide cover. It's when you have Huge+ creatures or differences in ground elevation (or of course flying) that 3d starts to matter for getting past potential cover.