Pumping AC

[Boggartbae]

So in another thread people were discussing whether or not it’s worthwhile to think of an AC bonus as a flat change in percentage points (i.e each +1=5% fewer damage points) or whether it was better to think of it in terms of the percentage change in damage that you were already taking

So if an enemy does an average of 20 damage when they hit you, but they need to roll a 19 or higher, they do 2 DPR on average, and getting +1 AC would halve that number to 1 DPR, saving you 1 damage per round.

But if you think of it as an additional 5% off of the 20 average damage, then it still works out to saving you 1 damage per round.

So if each additional point of AC saves you the same amount of damage, regardless of if your enemies need to roll higher than a 3 or higher than an 18, why would you ever think of AC bonuses in terms of percentage rather than percentage points? Is it something about criticals? What am I missing here? Please explain :smallconfused:

[Menzath]

I think part of the problem of thinking of AC like this is, that AC is not very inclusive to being damaged by anything but mundane means.
And alot of the scarier damage is from saving throws, or special abilities that have either a save or Target touch AC.
Now AC is good, but only to a degree.

[Nifft]

In my experience, the best way to avoid damage is to end an encounter faster.

Defense is useful -- it can help keep you alive long enough to win -- but winning requires offense, and focusing resources on offense first seems like the best way to minimize damage to yourself.

So, I'd recommend getting as much defense as you can get for cheap, and spendin the majority of your resources on offense.

[Boggartbae]

Ok I understand that AC doesn’t protect against everything, and that dead enemies can’t deal damage, but disregarding all that, the question I’m trying to ask is Whether or not armour class scales exponentially as it gets higher (until they need a 20 to hit you and additional AC bonuses don’t do anythin) or whether each point of AC gives you the same 5% increase in effective hp versus mundane attacks.

[Nifft]

Ok I understand that AC doesn’t protect against everything, and that dead enemies can’t deal damage, but disregarding all that, the question I’m trying to ask is Whether or not armour class scales exponentially as it gets higher (until they need a 20 to hit you and additional AC bonuses don’t do anythin) or whether each point of AC gives you the same 5% increase in effective hp versus mundane attacks.

The decrease in damage can be represented exponentially, but IMHO that's misleading.

Here's how I'd look at it:


Expected_damage = hit_damage * chance_to_hit

Given that all else remains the same, each point of AC reduces expected_damage by 5% of hit_damage, until chance_to_hit equals 5%.

BUT ALSO, 10% is 200% of 5%, so some people may claim the change in damage is exponential. This is technically true, but IMHO it's rather misleading.


However, in many cases all else does NOT remain the same:
- Tripping your attacker imposes additional costs on some future attacks.
- Power Attack exists, so your attacker may choose to reduce a self-imposed attack penalty -- at this point, you're probably trading 1 point of AC for 2 points of damage, not 5% of hit_damage.
- Dropping two of the three attackers reduces your chance to be hit significantly.
- A miss chance from blur or displacement can reduce your chance of being hit below even 5%. With greater invisibility and a 5 ft. step, you might not even be in the square which the opponent targets.

[Zanos]

You're thinking in terms of absolute damage rather than relative turns active, which is what matters.

Consider a character with 200 hp against an enemy that does 50 aDPR unadjusted for AC. Let's say it has +10 to hit.

Relevant AC range:



AC
Chance to Hit
Adjusted Damage
Rounds Alive


22
.95
47.5
4.21


23
.9
45
4.44


24
.85
42.5
4.71


25
.8
40
5


26
.75
37.5
5.33


27
.7
35
5.71


28
.65
32.5
6.15


29
.6
30
6.67


30
.55
27.5
7.27


31
.5
25
8


32
.45
22.5
8.89


33
.4
20
10


34
.35
17.5
11.42


35
.3
15
13.33


36
.25
12.5
16


37
.2
10
20


38
.15
7.5
26.67


39
.1
5
40


40
.05
2.5
80



Also, have a graph:
https://www.wolframalpha.com/input/?i=plot+200+%2F+(50+(1+-+0.05x))+0+%3C+x+%3C+20,+0+%3C+y+%3C+100


As you can see, the value in rounds alive of an AC point increases the further you are from a creature's attack bonus. The first point only adds .23 rounds to expected survival time, where the last point adds 40 rounds to expected survival time and literally doubles survival time over the previous point.

[Fizban]

The attack bonuses of enemies and AC of PCs vary wildly even before accounting all the situational and magical possibilities. Most importantly for understanding AC, there's also a soft cap+expectation of how much AC the frontline has, after which you have to invest more significant resources to climb higher.

The game expects frontline to be in full plate, heavy shield, amulet of natural armor, and ring of protection, all properly balanced in their magical bonuses. Yes, you're expected to use a shield if you don't want to get hit, no, this should not be surprising, especially at later levels when that shield is supposed to be +4 or 5. With the full kit and a bit of dex, against MM1 foes of CR equal to your level, you'll usually be somewhere in the 50/50 range. If you ditch the shield then AC sucks because you just dropped from 50/50 to 70/30 or worse (from even a modest +2 shield). If you make a point of boosting your AC with feats, prestige classes, and/or buff synergy, you can reach the point where enemy accuracy hits that tiny 15, 10, or 5% where dps isn't really a question anymore, just smooth sailing with the occasional lucky hit in. Thus, an AC build can block hits from the creatures you're supposed to fight multiple times per day, stretching hp reserves across the day, and against bosses with +8 or 10 higher will still have a miss chance where glass canons won't (a miss chance that stacks with miss chance)

This is right there in core: Dwarven defender gives +4 AC over 10 levels (plus the +1 from the Dodge prerequisite), Combat Expertise lets you flip a switch for +5, and Defensive Stance lets you flip a different switch for +4. If you're actually kitted for standard AC, +8 or more tilts you into nigh-untouchable, that's it- and a lot of that is actually dodge bonuses backed up by uncanny dodge. The problem is that people turn up their noses at shields, (putting them 2-7 points in the hole already) and basically every class or feat that boosts AC aside from Abjurant Champion.

Against more powerful monsters in other books you might need to bring in more. Every AC boosting PrC is essentially inferior to the Dwarven Defender, it's supposed to be the best and they actually respected that so everything else is watered down variants, but some of them are available to non-dwarves at least. Tome of Battle gives you Pearl of Black Doubt to cripple multiple attacks and counters to ensure the first misses, or just Stonefoot Stance for another flat +2 against large, both available long before you can enter the defense boosting PrCs- which incidentally start ramping up at about the same time monster attack bonus, WBL, and magic all do, in the 10th-13th band.

Right in the band where people like to focus because magic is hot, and complain about how melee can't survive, is where the line is drawn between defensive and glass canon melee characters. And they're surprised the constant procession of glass canons can't take a hit, even with their Animated Shields- which cost extra, pushing back max shield enhancement, and they're short a couple points from no PrC, and they won't use Combat Expertise because they're too busy Power Attacking, and hey look that's like 8 points of AC you could have had against scary bad stuff but no, straw glass canon is glass.

Edit: interesting table idea, but the AC numbers are all off as +10 vs AC 22 is 0.35 chance of hit, not 0.95. I think you meant to use +20 to hit. Numbers aren't all that representitive either, +20 is around 10th level high attack bonus, but only very high op characters will have 200 hp at 10th. A minimum AC for 10th is say, 25, which can be boosted to 30 with Combat Expertise. Depending on monster attacks range from lots of +15s to single +25s. If 25-30 seems a little low that's cause it is- only +2 armor+shield and a +1 ring or amulet, no buffs, no PrCs, just expertise to turn on/off (most people seem to recommend level+20). And it's still enough to blunt that incoming damage.

[Thurbane]

Touch AC is always something worth improving, when you can afford to.

A lot of the nastiest spells/effects in the game are delivered by a (ranged) touch attack.

[Calthropstu]

So in another thread people were discussing whether or not it’s worthwhile to think of an AC bonus as a flat change in percentage points (i.e each +1=5% fewer damage points) or whether it was better to think of it in terms of the percentage change in damage that you were already taking

So if an enemy does an average of 20 damage when they hit you, but they need to roll a 19 or higher, they do 2 DPR on average, and getting +1 AC would halve that number to 1 DPR, saving you 1 damage per round.

But if you think of it as an additional 5% off of the 20 average damage, then it still works out to saving you 1 damage per round.

So if each additional point of AC saves you the same amount of damage, regardless of if your enemies need to roll higher than a 3 or higher than an 18, why would you ever think of AC bonuses in terms of percentage rather than percentage points? Is it something about criticals? What am I missing here? Please explain :smallconfused:

Because of several reasons.
First, there's a cap. If you get to the point where they need a nat 1 to miss or a nat 20 to hit, changing the ac any further is pointless.
Take something with a +40 to hit. My ac is 22. Raising my ac to 60 would require a huge investment while raising it to 42 or less would be absolutely pointless. So raising AC at that point gives zero return. Likewise, if my AC is in the 50s and I am fighting things that have +30 or less to hit, raising AC is also giving zero return.
At that point, investing in hp or other attack reduction and survivability factors becomes more prudent.
Second, that's not exactly how it works out. Bad/good rolling can change that in a heartbeat. I've seen a guy go an entire session rolling nothing but a 15+.

[Geddy2112]

In most circumstances, AC becomes a bit of a losing man's game at higher levels. Things can hit characters either automatically, or characters are built around it. A general rule of thumb is Level+10 is bare minimum, 15+level is good for most characters, and 20+level is tanking. At really high levels AC usually won't protect against a creature's primary attacks, but secondary and itterative attacks can usually be blocked. At really high levels you are better off with miss chances and immunity.

At a certain point, you can have redundant AC. To the point even a tarrasque with true strike needs a natural 20 to hit.

It is possible to get AC really high, but you generally won't be able to do much else. In pathfinder, you can have a level 20 monk with over 100 AC and a lot of other defenses, but that is silly and far more than needed.

[Nifft]

You're thinking in terms of absolute damage rather than relative turns active, which is what matters.

The other thing that matters is MINIMUM turns active.

In the worst case, can you survive a very bad round and therefore use your next turn to take counter-measures?

Hit points are good for this, and stuff like Immediate action crit-negators can be amazing.


Finally, you want to minimize the chance that you'll lose turns from conditions other than death. Immunity to Stun, Daze, Trip, Paralysis, Sleep, Nausea, Involuntary Teleportation ("InTel"), being buried in mud, being buried in sand, being buried in liquid magma, Petrification, Domination, Fear... the list goes on.

[Calthropstu]

The attack bonuses of enemies and AC of PCs vary wildly even before accounting all the situational and magical possibilities. Most importantly for understanding AC, there's also a soft cap+expectation of how much AC the frontline has, after which you have to invest more significant resources to climb higher.

The game expects frontline to be in full plate, heavy shield, amulet of natural armor, and ring of protection, all properly balanced in their magical bonuses. Yes, you're expected to use a shield if you don't want to get hit, no, this should not be surprising, especially at later levels when that shield is supposed to be +4 or 5. With the full kit and a bit of dex, against MM1 foes of CR equal to your level, you'll usually be somewhere in the 50/50 range. If you ditch the shield then AC sucks because you just dropped from 50/50 to 70/30 or worse (from even a modest +2 shield). If you make a point of boosting your AC with feats, prestige classes, and/or buff synergy, you can reach the point where enemy accuracy hits that tiny 15, 10, or 5% where dps isn't really a question anymore, just smooth sailing with the occasional lucky hit in. Thus, an AC build can block hits from the creatures you're supposed to fight multiple times per day, stretching hp reserves across the day, and against bosses with +8 or 10 higher will still have a miss chance where glass canons won't (a miss chance that stacks with miss chance)

This is right there in core: Dwarven defender gives +4 AC over 10 levels (plus the +1 from the Dodge prerequisite), Combat Expertise lets you flip a switch for +5, and Defensive Stance lets you flip a different switch for +4. If you're actually kitted for standard AC, +8 or more tilts you into nigh-untouchable, that's it- and a lot of that is actually dodge bonuses backed up by uncanny dodge. The problem is that people turn up their noses at shields, (putting them 2-7 points in the hole already) and basically every class or feat that boosts AC aside from Abjurant Champion.

Against more powerful monsters in other books you might need to bring in more. Every AC boosting PrC is essentially inferior to the Dwarven Defender, it's supposed to be the best and they actually respected that so everything else is watered down variants, but some of them are available to non-dwarves at least. Tome of Battle gives you Pearl of Black Doubt to cripple multiple attacks and counters to ensure the first misses, or just Stonefoot Stance for another flat +2 against large, both available long before you can enter the defense boosting PrCs- which incidentally start ramping up at about the same time monster attack bonus, WBL, and magic all do, in the 10th-13th band.

Right in the band where people like to focus because magic is hot, and complain about how melee can't survive, is where the line is drawn between defensive and glass canon melee characters. And they're surprised the constant procession of glass canons can't take a hit, even with their Animated Shields- which cost extra, pushing back max shield enhancement, and they're short a couple points from no PrC, and they won't use Combat Expertise because they're too busy Power Attacking, and hey look that's like 8 points of AC you could have had against scary bad stuff but no, straw glass canon is glass.

Edit: interesting table idea, but the AC numbers are all off as +10 vs AC 22 is 0.35 chance of hit, not 0.95. I think you meant to use +20 to hit. Numbers aren't all that representitive either, +20 is around 10th level high attack bonus, but only very high op characters will have 200 hp at 10th. A minimum AC for 10th is say, 25, which can be boosted to 30 with Combat Expertise. Depending on monster attacks range from lots of +15s to single +25s. If 25-30 seems a little low that's cause it is- only +2 armor+shield and a +1 ring or amulet, no buffs, no PrCs, just expertise to turn on/off (most people seem to recommend level+20). And it's still enough to blunt that incoming damage.

Psh, screw all that noise. My glass cannon has an ac of 8. Eat my wall of hp.

[Boggartbae]

Ok thank you everyone. Thinking about it in terms of rounds alive, and looking at the math involved in calculating that (as well as the graph) has demonstrated the point really well.

It is really frustrating that this makes AC rather binary in effectiveness, and you need to heavily invest with things like PrC’s or shields before you turn the curve of increasing your expected rounds alive. It explains why my characters always feel squishy even when I invest a little in defense.

I’m a really paranoid player, so I generally try to play an undead or a construct for the immunities, and then pump my dex so that my touch AC is high.

It’s nice to know that AC really does scale exponentially though. Not understanding this was seriously keeping me up at night.

Edit: I’ve been playing with the graph and it looks like HP matters rather little when the surviveability from AC is concerned. Even halving the HP of the character to 100 still yields and almost identical result of a spike in turns alive after lowering enemy hit rates to below 80%. Obviously HP matters for other reasons, such as falling and lava swimming, and is the primary stat keeping you alive in combat for when the enemies hit you more than 80% of the time, but this has still been a very eye-opening conversation.

[Calthropstu]

Ok thank you everyone. Thinking about it in terms of rounds alive, and looking at the math involved in calculating that (as well as the graph) has demonstrated the point really well.

It is really frustrating that this makes AC rather binary in effectiveness, and you need to heavily invest with things like PrC’s or shields before you turn the curve of increasing your expected rounds alive. It explains why my characters always feel squishy even when I invest a little in defense.

I’m a really paranoid player, so I generally try to play an undead or a construct for the immunities, and then pump my dex so that my touch AC is high.

It’s nice to know that AC really does scale exponentially though. Not understanding this was seriously keeping me up at night.

Edit: I’ve been playing with the graph and it looks like HP matters rather little when the surviveability from AC is concerned. Even halving the HP of the character to 100 still yields and almost identical result of a spike in turns alive after lowering enemy hit rates to below 80%. Obviously HP matters for other reasons, such as falling and lava swimming, and is the primary stat keeping you alive in combat for when the enemies hit you more than 80% of the time, but this has still been a very eye-opening conversation.

Actually, you can pump your hp for similar results. But none of this considers tactics. The ac 8 glass canon I mentioned only has to last 1 round of hand to hand... ever. Tactics can turn that chart upside down and inside out.
An archer, for example, could take infinite shots at a melee build without ever getting hit. Setting up traps and leading your opponents into them... ac is actually your least important defense.

[Fizban]

I say again, it should not be surprising that characters who refuse to use shields have bad AC. The investment required to have a normal AC is only "heavy" if you demand walking around shield-less and can't be bothered to invest *any* build resources. Why would you expect be good at something you've put zero effort into?

The answer, I expect, is "because spellcasters are," but someone who refuses to take and use metamagic is going to be bad at blasting, the same way someone who refuses to learn BFC spells is going to be bad at control, a wizard who refuses to buy and learn extra spells won't be any better than a sorcerer, a cleric who refuses to prepare anything but Divine Power won't be able to protect the party, and on and on.

As for one-rounding everything- that's not tactics. It's actually about the closest thing to a definition as there is for "overpowered." Yes, if your game is all about builds that can one-shot anything they fight, defenses matter significantly less, this should not be surprising either. Such a game requires significantly more tactical maneuvering, the same way tactics become more important when you move from swords and crank-crossbows to miniguns and rocket launchers (assuming the enemies are similarly one-shot capable at all times). Because you chose to make it that way.

But I really, really doubt I'm actually going to convince the people who dig in their heels on this, so I'll leave it at that.

[Elkad]

Actually, you can pump your hp for similar results. But none of this considers tactics. The ac 8 glass canon I mentioned only has to last 1 round of hand to hand... ever. Tactics can turn that chart upside down and inside out.
An archer, for example, could take infinite shots at a melee build without ever getting hit. Setting up traps and leading your opponents into them... ac is actually your least important defense.

Is that a L20 build? He can disable every enemy (in range) in a single round?

Not saying it isn't possible, but it sure seems unlikely. An otherwise easy encounter with a swarm of low level creatures could do a hell of a lot of damage. 100 CR4 guys with Longbows at +7 for 1d8+2 are going to do 650 a round to you. That's a CR17 encounter.

With a completely unexceptional AC27, you take 32 damage the first round instead. If you take out 10 creatures a round, that's ~150 damage overall, easily survivable.

[Calthropstu]

Is that a L20 build? He can disable every enemy (in range) in a single round?

Not saying it isn't possible, but it sure seems unlikely. An otherwise easy encounter with a swarm of low level creatures could do a hell of a lot of damage. 100 CR4 guys with Longbows at +7 for 1d8+2 are going to do 650 a round to you. That's a CR17 encounter.

With a completely unexceptional AC27, you take 32 damage the first round instead. If you take out 10 creatures a round, that's ~150 damage overall, easily survivable.

No, he escapes hand to hand without fail. Whether it be flight, dimensional travel, invisibility, or simply running away with his movement speed around 100. With a concentration check bonus of around 30 at 12th lvl in pathfinder, his ability to escape is fairly absolute. Once out of hand to hand, he resumes bombardment and/or summoning from relative safety.
He generally goes into combat with summons active to act as a wall of hitpoints to protect him from melee even further, and has a ring of counterspells protecting him from dimensional anchor.
Tactical withdrawals are a strong winning strategy. Another way to win with ac meaning little involves illusions. Create a false you, send it into an area full of traps and watch your would be attackers fall to their doom.
One time I had a wizard cast 2 spells against a party, causing them to waste over 80% of their resources trying to destroy an "invisible" opponent. Imagine a 200 foot tall open tower, with criscrossing wires all over with bells attached.
The wizard was hiding behind hing a silent image spell on a small balcony 100 feet up. He used a wand of mage hand to manipulate the bells.
After the party ran out of blasting spells and abilities and the sorcerer's flight spell wore out, he revealed himself.

Like I said, tactics can render attack roles meaningless. A hide in plain sight rogue can do similar using hit run and disappear. A ranger can fire a bow from so far away, striking back is impossible. An invisible, flying alchemist can drop bombs from above to devastating effect. A pounce/spring attack fighter can severely change up that little table by restricting the attacks he recieves to 1 per round. He can then add reach to further increase his damage output altering the table further. An uber charger designed to instagib most opponents also renders ac moot... since his target never gets an attack at all.

Anyways, as I was saying AC is literally your least important defense.

[emeraldstreak]

shields

Shields may be annoying, but both 3.5 and PF provide some alternatives even for low level non-casters. The spell 'Shield' is personal range and cannot be made into potions, but Force Screen can be made into psionic tattooes at double the price. There's a Dragonmark in Eberron that gives Shield as SLA, and the optional feat Hidden Talent allows Force Screen manifesting. Pathfinder has a feat named Shield of Swings that is a flexibility choice between AC and damage for wielding two-handed weapons. And these are options for the 'shield' component of AC alone, there're plenty more for armor, natural armor, dodge, deflection, etc.

[Boggartbae]

I don’t think that armour class is the least relevant defense measure. Having superior tactics is great, but your dm can use them too, and all they have to do is switch some numbers around and give your monsters improved initiative, and you could wind up getting grabbed by a bonehoard or similar monster that can effectively kill you in one hit. In those situations AC is all that’s standing between you and death, so ignoring it in favour of tactics and positioning will get you killed eventually, since your dm can vary encounters to make certain tactics useless, and force you into a pure numbers fight.

Besides, it never hurts to be overly paranoid, and it’s not like having high AC limits your tactics. In fact, it actually raises the options available to you and your party, and with a clever head, could lead to saving their lives.

[Deophaun]

Yes, you're expected to use a shield if you don't want to get hit, no, this should not be surprising, especially at later levels when that shield is supposed to be +4 or 5.
Except the characters with the highest AC are often forbidden from using shields as they stop applying their Charisma or Wisdom modifier.

The problem is that people turn up their noses at shields, (putting them 2-7 points in the hole already) and basically every class or feat that boosts AC aside from Abjurant Champion.
Chameleon seems rather popular.

[Fizban]

Except the characters with the highest AC are often forbidden from using shields as they stop applying their Charisma or Wisdom modifier.
I am intrigued but mostly confused- is this a Monk reference, or possibly a Pathfinder reference? Something that stacks the same stat to AC multiple times? Vow of Poverty? Shapeshifing into teeny hig-dex forms? Probably sor/monk/abjurant champion or similar. Regardless, a highly optimized AC build that happens to ban shields as part of a gambit to gain more AC hardly disproves the trend of two-handed builds that eschew even the slightest investment in AC. Refusing to use a shield when you don't have a replacement that won't stack with a shield anyway, means you're going to have bad AC.

Chameleon seems rather popular.
Again, confused, because Chameleon doesn't boost AC? Only indirectly via spellcasting or setting your variable feat to an AC booster.

[Deophaun]

I am intrigued but mostly confused- is this a Monk reference, or possibly a Pathfinder reference? Something that stacks the same stat to AC multiple times? Vow of Poverty? Shapeshifing into teeny hig-dex forms? Probably sor/monk/abjurant champion or similar. Regardless, a highly optimized AC build that happens to ban shields as part of a gambit to gain more AC hardly disproves the trend of two-handed builds that eschew even the slightest investment in AC.
The comment was not meant to do such.

As for your question, I've found it trivially easy to boost through dips in either monk, swordsage, battle dancer, or just picking up a monk's belt+ascetic mage. Abjurant Champion builds tend to be more Int-focused in my experience for some reason (I used it on a dragonwrought kobold that used alter self to turn into a shadow dragon wyrmling, so yes to your teeny form, only this one had +7 NAC). The question is "do you care about AC?" Armor, ironically, exists for those who don't, hence why I used that particular quote. That may be an expectation of the designers, but if you don't want to be hit and you're holding a shield, you're doing it wrong.

Armor-reliant builds are also often counterproductive for defense and offense due to speed penalties (also, swim penalties: in how many gaming groups has a knight in shining armor drowned?), which as you pointed out movement is more valuable for survival. There are ways around that, but those also often require permissive interpretations of RAW and magic item creation rulings.

Again, confused, because Chameleon doesn't boost AC? Only indirectly via spellcasting or setting your variable feat to an AC booster.
It's the indirect boosts that matter the most. The variable STAT boost to Dex, Cha, or Wis on the appropriate chassis is almost as good as Dwarven Defender's boost for AC alone, and if you're picking up divine casting you get armor of darkness for a +8 deflection bonus, with significant gold savings as you no longer have an overpriced ring to worry about. Throw on owl's insight on a Wis-SAD gish and that's another 5 AC; more if you use CL boosters.

Boosting AC is all about the indirects. Not only does it boost your number, but being indirect boosts they are performing double duty elsewhere so you aren't really sacrificing much if anything to get those numbers. Take the alter self into the shadow dragon wyrmling above. I'm not merely getting +7 NAC and +1 size (small kobold to tiny dragon), but a base 60 ground and 120 fly speed, attainable at level 4.

[tiercel]

One of the points I see made here is that "when a monster has +40 to hit then AC just isn't worth it" -- er, Power Attack mitigation? (Also, mitigating iterative attacks, against opponents who actually use them, much less a full-attacking Power Attack user.)

The point is that an opponent with a really high attack bonus isn't going to be just rolling that bonus - they'll be throwing a lot more attacks and/or throwing a lot more damage because of their high bonus, and cutting into either one is worthwhile. Generally, AC seems to be underrated by a fair number of people, especially given how common weapon/natural weapon attacks are.

It is fair, however, to note that AC has to be balanced with a lot of other considerations (touch AC, saves, mobility, resistances/immunities, etc) even when it comes to defense, much less the serious downside of neglecting offense for defense. I'm not arguing the whole "ending the battle through battlefield control/damage is the best defense" approach is wrong, per se, but dumping AC seems to me like just the other extreme to pumping it to the point where it causes serious losses in offense or other defenses.

[Zanos]

The other thing that matters is MINIMUM turns active.

In the worst case, can you survive a very bad round and therefore use your next turn to take counter-measures?

Hit points are good for this, and stuff like Immediate action crit-negators can be amazing.

Finally, you want to minimize the chance that you'll lose turns from conditions other than death. Immunity to Stun, Daze, Trip, Paralysis, Sleep, Nausea, Involuntary Teleportation ("InTel"), being buried in mud, being buried in sand, being buried in liquid magma, Petrification, Domination, Fear... the list goes on.
Topic was about AC, so I just answered why AC is more useful the more you have, to a cap.

Hit Points scale linearly, but apply to more attacks.

I agree that other defenses can be more important than AC.

[Nifft]

Topic was about AC, so I just answered why AC is more useful the more you have, to a cap.

Hit Points scale linearly, but apply to more attacks.

I agree that other defenses can be more important than AC.

Agree.

I'm adding to your point, not correcting it.

It's my hope that the additional comments help create a foundation for the OP to have a more complete analysis going forward.

[Boggartbae]

Topic was about AC, so I just answered why AC is more useful the more you have, to a cap.

Hit Points scale linearly, but apply to more attacks.

I agree that other defenses can be more important than AC.

I'll be honest I'm always chuffed if my threads start conversations, but you did answer my question perfectly, so I need to say thank you for that. That graph is what really did it, so thanks for taking the time to make that for me.

And yeah as I said before I know that AC isn't the be all end all for defense, but It's still nice to understand the math behind optimizing.

[Techwarrior]

Something that gets ignored quite often when discussing AC (not touch) is that even if it doesn't necessarily stop you from getting hit by monsters at your level, boosting AC can have benefits. The most applicable version of this is Power Attack. In a game where your AC is actually coming into play, sometimes just having enough AC that the dragon doesn't dump 10 points of BAB into Power Attack is saving you damage without meaning that you're avoiding getting hit.

[TalonOfAnathrax]

AC is fantastic when it comes to dealing with groups of power-level-wise creatures.
Even a semi-optimised AC allows a PC to essentially ignore threats like orc tribes (for example a rogue I played with was pumping AC and could basically ignore whole orc tribes around level 11 or so).

[Crake]

Because of several reasons.
First, there's a cap. If you get to the point where they need a nat 1 to miss or a nat 20 to hit, changing the ac any further is pointless.
Take something with a +40 to hit. My ac is 22. Raising my ac to 60 would require a huge investment while raising it to 42 or less would be absolutely pointless. So raising AC at that point gives zero return. Likewise, if my AC is in the 50s and I am fighting things that have +30 or less to hit, raising AC is also giving zero return.
At that point, investing in hp or other attack reduction and survivability factors becomes more prudent.
Second, that's not exactly how it works out. Bad/good rolling can change that in a heartbeat. I've seen a guy go an entire session rolling nothing but a 15+.


Psh, screw all that noise. My glass cannon has an ac of 8. Eat my wall of hp.

Say hello to my good old friend: power attack.

Power attack is the single most important reason to keep your AC at least somewhat relevant. You might have a wall of hp, but the lower your AC, the harder enemies can hit you with power attack, and if an enemy is sporting a +40 attack bonus, you can bet yourself most of it is from his bab. Sure your AC of 42 would mean he still only misses on a natural 1, but now he can't take -20 to hit for +40 bonus damage on his swing, or, if he has leap attack and a valorous weapon, then you'd be saving yourself something within the realm of 120 damage per hit.

The only time AC becomes irrelevant is when it's gone too high; lower AC still poses problems for you for the simple reason of "power attack".

[Eldariel]

Say hello to my good old friend: power attack.

Power attack is the single most important reason to keep your AC at least somewhat relevant. You might have a wall of hp, but the lower your AC, the harder enemies can hit you with power attack, and if an enemy is sporting a +40 attack bonus, you can bet yourself most of it is from his bab. Sure your AC of 42 would mean he still only misses on a natural 1, but now he can't take -20 to hit for +40 bonus damage on his swing, or, if he has leap attack and a valorous weapon, then you'd be saving yourself something within the realm of 120 damage per hit.

The only time AC becomes irrelevant is when it's gone too high; lower AC still poses problems for you for the simple reason of "power attack".

However, there are auxiliary angles to defense. Blinking/Mirror Images/etc. work vs. a great number of things and give you good enough resistances not to care about PA much of the time. And of course, you can always just take Elusive Target to negate Power Attack entirely. And combine it with Starmantle Cloak (or the Starmantle spell) and ensure you can make DC15 Reflex saves (the Evasion combo would probably bring about banhammer so avoid that). Add DR and the common means of doing damage are almost completely ineffective against you (ToB martial adepts and charge multipliers are still relevant though).

And Leap Attackers generally have Shock Trooper too so there your AC is quite irrelevant unless it works vs. their full attack bonus. That said, bags of HP are generally always applicable; Shared Vigor + Share Pain with your Psi Crystal can get you pretty high (and even just False Life shared, though stuff like Spectral Touch and Vampiric Touch can add on top of it, combined with Shield Other can do similar stuff with magic - though Vigor makes for bigger numbers).


In general, it's a matter of opportunity cost. Pumping your AC from 8 to 40 costs more than the average calculated benefits are taking in all the variables. Many enemies might not even attack your AC and of all that do, well, the amount of damage saved by pumping your AC to make Power Attacking less profitable for them is far less likely to help you than other ways to invest the same amount of resources into other defenses or offenses.

[Fizban]

Boosting AC is all about the indirects. Not only does it boost your number, but being indirect boosts they are performing double duty elsewhere so you aren't really sacrificing much if anything to get those numbers. Take the alter self into the shadow dragon wyrmling above. I'm not merely getting +7 NAC and +1 size (small kobold to tiny dragon), but a base 60 ground and 120 fly speed, attainable at level 4.
I feel like you're kinda missing my point. A gish build abusing the best of everything they can find finds it trival to boost AC, the same way they'll find it trival to do anything, because they're a high-op spellcaster. If the DM is allowing that then no amount of statistical AC analysis is going to matter, because the game has left behind any standards for which a broad analysis applies.

I take it as a given that when people are talking about AC, they're actually talking about non-spellcasting AC. Apparently my breakdown of core AC expectations for melee characters didn't get that assumption across. I also respond to the default game, not the high-op metagame. All of your examples are quite useful though, as it's still uncommon for high-op builds to mention AC, even though as you point out it can be easily done.


However, there are auxiliary angles to defense. Blinking/Mirror Images/etc. work vs. a great number of things and give you good enough resistances not to care about PA much of the time.
Miss chance is actually pretty overrated. They don't stack, and there's essentially a soft cap of 25% as taking Blind-Fight and closing your eyes gives you an effective 25% as long as you know what square they're in. That's also before the question of weather those illusions work against Blindsight, and they're beat by True Seeing regardless, which is available in cheap items and on many high level monsters. Against things that don't have access to any of those, you could probably stay out of reach or get the same result from AC without serious investment, and AC would stack with the miss chance.

Blink is the exception to the stacking, but it also fouls your attacks (and is expensive in item form). Unless you use Greater Blink, which is a 5th level spell (should be more like 7th). They're both personal only, so once again you're a spellcaster, see above. Blink is also extradimensional travel, and supposedly blocking that is required to maintain high-op high-level play, so it's not going to work in the dungeon.


Add DR
The DR available from prestige classes is also criminally underrated. Typically 5/- at 15th, read a different way this is effectively +5hp per hit that lands on you. People love Stoneskin, but this is always on with no limit, cost, buff time, or dispelling, and works against adamantine weapons. Well actually they love Heart of Earth, but again, don't see why one would want to compare the effectiveness of spellcasting defenses against non-spellcasting enemies, most people seem to take spellcasting=win as a given. Comparing the defenses of a gish to the defenses of a non-gish is the same as comparing the offense of a gish to a non-gish.

[Thurbane]

I generally find the true worth of AC lies somewhere in between the inflated value that WotC staff assigned to it, and the almost useless status a lot of high-op gamers assign to it.

It will, of course, vary by table: campaign type, enemies, party composition etc.

[Deophaun]

I feel like you're kinda missing my point. A gish build abusing the best of everything they can find finds it trival to boost AC, the same way they'll find it trival to do anything, because they're a high-op spellcaster.
There's the problem: you think that's high-op. Hitting an AC between 50-60 at level 20 is mid-op in my book, and if that's too high for some, I can always find means of converting AC into something else (like the fierce weapon enhancement or Robilar's Gambit) to bring my defenses down to whatever is assumed.

I take it as a given that when people are talking about AC, they're actually talking about non-spellcasting AC.
While I find the difference between spellcasting and non-spellcasting to simply be about what works in an AMF and what doesn't. Some classes just have better spells (wizard) than others (fighter).

I generally find the true worth of AC lies somewhere in between the inflated value that WotC staff assigned to it, and the almost useless status a lot of high-op gamers assign to it.
See? Right here: the fact that I care about AC at all excludes this from high-op.

[Crake]

However, there are auxiliary angles to defense. Blinking/Mirror Images/etc. work vs. a great number of things and give you good enough resistances not to care about PA much of the time. And of course, you can always just take Elusive Target to negate Power Attack entirely. And combine it with Starmantle Cloak (or the Starmantle spell) and ensure you can make DC15 Reflex saves (the Evasion combo would probably bring about banhammer so avoid that). Add DR and the common means of doing damage are almost completely ineffective against you (ToB martial adepts and charge multipliers are still relevant though).

And Leap Attackers generally have Shock Trooper too so there your AC is quite irrelevant unless it works vs. their full attack bonus. That said, bags of HP are generally always applicable; Shared Vigor + Share Pain with your Psi Crystal can get you pretty high (and even just False Life shared, though stuff like Spectral Touch and Vampiric Touch can add on top of it, combined with Shield Other can do similar stuff with magic - though Vigor makes for bigger numbers).


In general, it's a matter of opportunity cost. Pumping your AC from 8 to 40 costs more than the average calculated benefits are taking in all the variables. Many enemies might not even attack your AC and of all that do, well, the amount of damage saved by pumping your AC to make Power Attacking less profitable for them is far less likely to help you than other ways to invest the same amount of resources into other defenses or offenses.

You make all kinds of great points. None of which really negate the fact that having lower AC means the enemy can power attack for more though. Shock trooper may be a thing, but if the enemy has to cut their own AC to hit you with a full power attack (because they don't have too, if your AC is low enough), then you're essentially enabling your party to deal more damage to the enemy, since it's AC is now terrible, and they can now all full power attack.

Having other defenses is really irrelevant to the issue, because the discussion isn't about defenses, it's about the times when that one massive power attack manages to a) hit the right mirror image, b) roll well on the concealment miss chance and c) you fail the reflex save on starmantle cloak. The only thing that came even close to arguing the point was elusive target, but with requirements involving dodge and mobility, I don't see it getting much play, especially when it only functions against your designated dodge target.

Also, generally AC is pretty easy to boost. For anyone other than classes that literally cannot wear armor due to their class features, 0 ACP, 0ASF sectioned plate is fairly trivial to get, and costs about 16-21k for +9AC, and you don't even need proficiency to use it, because it's got 0ACP. For a bit of extra money, you could even get it to count as light armor, meaning no speed penalty (the halfweight armor property, +3 ability), and have be super light, like 12.5lb, same weight as a mithril chain shirt, or just get a pair of boots of striding and springing for +10ft movement speed.

A mithril buckler is likewise super cheap and able to be cheaply enhanced for another bunch of easy AC, so just with those two, a decent dex score (the above fullplate allows up to +4), a barkskin and/or shield of faith from the casters will net you another 8-10 AC, and if you can alter self/polymorph into a high natural armor form like a troglodyte, that's another bunch from natural armor. All that is fairly easily achievable by the mid game, for a total of something like 40AC without even really putting much of a dent in your money pocket.


Blink is also extradimensional travel, and supposedly blocking that is required to maintain high-op high-level play, so it's not going to work in the dungeon.

I also just wanted to add on this: Blink also requires an ethereal plane, so it doesn't even function on extraplanar adventures, only on the material/shadow plane.

[Calthropstu]

In order to attack my glass cannon ac 8 you have to:
Get past the other members of my party.
Get past my summons.
Get past my speed.
Get past my dimensional travel.
Overcome my mirror image.
Get past my invisibility.
Get past my flight.
And, if by some miracle you succeed at all that, get past my auto rez.

I see people touting true sight, but that is only effect to 120 feet. With a movement speed of 100, That is NOT hard to avoid.

[Elkad]

In order to attack my glass cannon ac 8 you have to:
Get past the other members of my party.
Get past my summons.
Get past my speed.
Get past my dimensional travel.
Overcome my mirror image.
Get past my invisibility.
Get past my flight.
And, if by some miracle you succeed at all that, get past my auto rez.

I see people touting true sight, but that is only effect to 120 feet. With a movement speed of 100, That is NOT hard to avoid.


And those 100 L4 archers will do most of that.
Ignore the summons (or other party members - in which case I get 400 archers instead) in favor of the correct target.
Longbows don't care about your speed (other than maybe a -2 to hit if you can make a range break)
If you are present to summon (or nuke - you did call yourself a glass cannon after all), you likely aren't extradimensional.
Mirror images soak up 8 arrows if you are visible, otherwise none.
Flight doesn't matter to longbows either.

So the only thing left is pinpointing your square for firing while you are invisible. If they can do that, you get hit with 50 arrows. Or 75 if they have blindfight.

Auto-rez in place solves nothing unless you can win initiative again and escape. Otherwise you just die again.

DR5/- would work well of course in this example.

[Calthropstu]

And those 100 L4 archers will do most of that.
Ignore the summons (or other party members - in which case I get 400 archers instead) in favor of the correct target.
Longbows don't care about your speed (other than maybe a -2 to hit if you can make a range break)
If you are present to summon (or nuke - you did call yourself a glass cannon after all), you likely aren't extradimensional.
Mirror images soak up 8 arrows if you are visible, otherwise none.
Flight doesn't matter to longbows either.

So the only thing left is pinpointing your square for firing while you are invisible. If they can do that, you get hit with 50 arrows. Or 75 if they have blindfight.

Auto-rez in place solves nothing unless you can win initiative again and escape. Otherwise you just die again.

DR5/- would work well of course in this example.

Blind fight wouldn't help at all, too far away. Your 400 archers wouldnt be able to hide, you'd be seen from far far away... which means you'd never get a single arrow off at me. Hooray for large earth elementals.1d3+1 per round means I could cast from relative safety, send wave after wave of dr 5 creatures at them... which they would reatreat after losing about 1/3 of their forces. Or I could simply become a huge fire elemental and laugh at their feeble attempts to hurt me with dr 10.
So yeah... not really much of a threat. I calculated out how many lvl 1 creatures he could kill in a knock down drag out fight and estimated it to be around 8k before he ran out of spells.

[Deophaun]

I see people touting true sight, but that is only effect to 120 feet. With a movement speed of 100, That is NOT hard to avoid.
If you're fighting in an open field, true. I find such engagements to be the minority, though, so I would never rely on being 120' away from my opponent.

As for getting past other party members, that only counts if those other party members are also glass cannons: relying on the AC of other party members is still relying on AC. It's like saying buffs are overrated because the cleric casts them for you.