AC: is it worthless?

[TypoNinja]

In my experience AC only helps for the first little while. After that you start losing ground unless you are willing to dump pretty much every penny into it.

The reason is scaling. At low levels you can pick up AC points cheap. But once you've filled all the item slots with AC boosters, and now have to upgrade existing items instead of buying new sources, the costs multiply fast.

Attack power on the other hand pretty much goes up for free. Melee type monsters get +1 BAB per hitdie, and 'brute' types usually end up with fairly high strength to go with it too. The result is your average beat stick gets better and better at hitting a given target as the levels go up, and the cost of keeping AC competitive just keeps climbing.

[Spuddles]

In my experience AC only helps for the first little while. After that you start losing ground unless you are willing to dump pretty much every penny into it.

The reason is scaling. At low levels you can pick up AC points cheap. But once you've filled all the item slots with AC boosters, and now have to upgrade existing items instead of buying new sources, the costs multiply fast.

Attack power on the other hand pretty much goes up for free. Melee type monsters get +1 BAB per hitdie, and 'brute' types usually end up with fairly high strength to go with it too. The result is your average beat stick gets better and better at hitting a given target as the levels go up, and the cost of keeping AC competitive just keeps climbing.

WBL scales nonlinearly too, though. And due to the high granularity of the system, getting hit 10% to 20% less often translates more to dying less often than it would suggest if you treated that as just 10% or 20% more HP.

I find it's usually worthwhile to shoot for decent AC on non-casters.

[ILM]

I disagree. For most characters, there's a significant difference in the ease of getting "Poor", "Decent", and "Good" AC as I defined it earlier.

I have to side with you on that one. Even if AC is all or nothing, it is no more so than miss chance, which is literally "get sword-to-faced or be suddenly elsewhere". With miss chance, you're flipping a coin; at least your AC protects you against the mooks who can only hit you on a 20.

Sure, individual high CR creatures may or may not be hard to survive with AC alone, but remember that four CR 6 creatures also work out to an EL 10 encounter (if my math is right). Eight CR 4 guys are EL 10 too. And while your 30 AC may only protect you 50% of the time against one CR 10 dude with +20 to his attacks, that same 30 AC will make you 95% invincible to the CR 4 mooks with +8 to hit. That's also something to bear in mind.

Essentially, 50% miss chance works out to setting your AC at whatever the enemy's attack bonus is plus 10, and then rolling to hit that (after rolling to hit your regular AC). It has benefits and disadvantages.

[LordBlades]

WBL scales nonlinearly too, though. And due to the high granularity of the system, getting hit 10% to 20% less often translates more to dying less often than it would suggest if you treated that as just 10% or 20% more HP.

I find it's usually worthwhile to shoot for decent AC on non-casters.

AC cost scales quadratically, WBL scales nonlinearly, but less than quadratically, which means for every extra point of AC you need to pay an increasing % of your WBL as you progress in levels.

[LordBlades]

Double post due to damn internet issues

[Spuddles]

AC cost scales quadratically, WBL scales nonlinearly, but less than quadratically, which means for every extra point of AC you need to pay an increasing % of your WBL as you progress in levels.

AB from monsters, at least SRD monsters, hits a high peak around the mid levels, then the rate of increase slows down. You're also only hitting quadratic costs past +2 bonuses, of which you can garner many.

[sonofzeal]

AC cost scales quadratically, WBL scales nonlinearly, but less than quadratically, which means for every extra point of AC you need to pay an increasing % of your WBL as you progress in levels.

If you're investing in a variety of different sources of AC, it usually works out to something reasonable.


Actually, I'm working on a plot of Mr McHumanFighter's AC vs monster attack bonuses, and already I'm getting some odd patterns. AC starts in the Decent/Good range, but decreases down to the Poor/Decent range around level 10... and then starts catching back up as more advanced sources of AC like Defending Armor Spikes become available and wealth starts scaling more quickly, getting back up to Good somewhere in the teens.

I'll post the chart when it's finished.

[ILM]

AB from monsters, at least SRD monsters, hits a high peak around the mid levels, then the rate of increase slows down. You're also only hitting quadratic costs past +2 bonuses, of which you can garner many.

Uh no, it's actually pretty linear all the way to 20 (and beyond, I just didn't plot those). This is 575 SRD monsters from CR 1/8 to 20. If I had traced the averages by CR, you'd get an almost straight line except for a small dip around 7 and 13.

[LordBlades]

If you're investing in a variety of different sources of AC, it usually works out to something reasonable.


Actually, I'm working on a plot of Mr McHumanFighter's AC vs monster attack bonuses, and already I'm getting some odd patterns. AC starts in the Decent/Good range, but decreases down to the Poor/Decent range around level 10... and then starts catching back up as more advanced sources of AC like Defending Armor Spikes become available and wealth starts scaling more quickly, getting back up to Good somewhere in the teens.

I'll post the chart when it's finished.

Thing is the higher you go in level, the more you start to encounter monsters that have SLAs (or other stuff) that do not target AC. Did a quick filter on CR12+ monsters in the SRD, and quite few of them don't have a credible threat that doesn't target AC.

The trend you're experiencing with the graph is pretty much consistent with my practical experience(for non casters): low-level AC can be useful, mid level AC not so much (because you can't keep up apparently) and it stays not very useful as you go up in level because, although you can keep up, the AC targeting attacks become less frequent, and you need to worry more and more about other defenses.

[TypoNinja]

WBL scales nonlinearly too, though. And due to the high granularity of the system, getting hit 10% to 20% less often translates more to dying less often than it would suggest if you treated that as just 10% or 20% more HP.

I find it's usually worthwhile to shoot for decent AC on non-casters.

I feel like If I spend enough cash on AC to be competitive I don't have enough cash left over to be useful at anything else. Though admittedly that's a subjective judgement rather than something I could stat out at this point. anyway, the upshot is, I tend to only spend on AC when the character is built around it, otherwise AC boosting is incidental.

Me likey Glass Cannon :P

[Spuddles]

I feel like If I spend enough cash on AC to be competitive I don't have enough cash left over to be useful at anything else. Though admittedly that's a subjective judgement rather than something I could stat out at this point. anyway, the upshot is, I tend to only spend on AC when the character is built around it, otherwise AC boosting is incidental.

Me likey Glass Cannon :P

To get truly good AC, imo, you need to devote class levels to it. Which is unfortunate.

[Necroticplague]

AC isn't worthless when it's actually used, the problem is that it has a binary nature (to make it worthwhile, you either invest large chunks into it, or you dont bother), that it is easy to get around (Oh, wow, you have an AC in the high 70s. My DFA,sorcerer,beguiler,warmage,swordsage,DN,dragon shaman,shadowcaster,and truenamer dont really care). The best way to avoid damage is to make sure the enemy is dead before they can attack you. Plus, AC is relative to the AB of what you're fighting, while other defenses are usually flat chances regardless of opponents (Oh, wow, an AB of +70? My cloak of minor blur still blocks 40%, while an optimized AC might not do anything).

[Spuddles]

AC isn't worthless when it's actually used, the problem is that it has a binary nature (to make it worthwhile, you either invest large chunks into it, or you dont bother), that it is easy to get around (Oh, wow, you have an AC in the high 70s. My DFA,sorcerer,beguiler,warmage,swordsage,DN,dragon shaman,shadowcaster,and truenamer dont really care). The best way to avoid damage is to make sure the enemy is dead before they can attack you. Plus, AC is relative to the AB of what you're fighting, while other defenses are usually flat chances regardless of opponents (Oh, wow, an AB of +70? My cloak of minor blur still blocks 40%, while an optimized AC might not do anything).

These points have already been addressed in the thread, and they're largely irrelevant or misleading. It may be true in pvp battles, but that's even worse, because everyone should have an item of true seeing, which makes miss chance worse than worthless.

Uh no, it's actually pretty linear all the way to 20 (and beyond, I just didn't plot those). This is 575 SRD monsters from CR 1/8 to 20. If I had traced the averages by CR, you'd get an almost straight line except for a small dip around 7 and 13.

Sweet graph

There's a substantial hump at CR9, where the upper bound of attack bonuses go up to 30 from about 20. Then it mostly goes back to the same rate. I suspect you've got some inflation at post CR10, due to dragon inclusions, and it actually decreases. Sonofzeal's first post crudely outlines what non-dragon ABs are doing in his OP, and from that was what I was discussing. I should have been more clear.

I suspect that hump at mid levels are giants.

[Gwendol]

Nice plot. What's the goodness of the linear fit?

Edit: Yeah, looking at the graph there is an almost flat range from about level 10 - 15.

[ILM]

There's a substantial hump at CR9, where the upper bound of attack bonuses go up to 30 from about 20. Then it mostly goes back to the same rate. I suspect you've got some inflation at post CR10, due to dragon inclusions, and it actually decreases. Sonofzeal's first post crudely outlines what non-dragon ABs are doing in his OP, and from that was what I was discussing. I should have been more clear.

Not really though. There's 2 monsters at CR 9 and 2 at CR 10 that break the expected zone. The CR 9 monsters are the Legendary Bear (a Bear to which you've applied the (epic) Legendary template, bringing him to 20 HD) and the 8th level Astral Construct. Pretty specific, but they're in the SRD. The CR 10 monsters breaking out there are the Legendary Tiger and 9th level Astral Construct. Finally, the two top dots at CR 11 are giants, you're right: specifically, the Hill Giant Dire Wereboar in Dire Boar form and Hybrid form, respectively.

Aside from those crazies, the upper AB is virtually a straight line all the way to 20.

Dragons may be tougher than other opponents, but then again it's Dungeons and Dragons, not Dungeons and Bulettes. I think they're worth including. Still, if you'd rather see a chart without them, here it is.



484 monsters. I removed all True Dragons. The curve does flatten a bit but then again there aren't all that many monsters left, are there? All of 11 monsters CR 17 and up, that's not really a sample anymore. Sure, you can take Dragons away to look at things, but you're removing 2/3rd of CR 15+ encounters (literally: it goes from 66 monsters to 22).

[Malachei]

This is a nice graph, ILM, and a good analysis. Thank you!

Do you have all monsters in a single source, such as Excel? I know a site that lists the monsters, but analysis is pretty limited. If you did this yourself, would you be willing to share this?

[sonofzeal]

Here's my preliminary chart. Once it's finished, I hope to be able to plot a line for a few hypothetical characters and see where they fall relative to the range. I think I'm going to add in a CR 3 figure, and there's always the possibility of filling in some of the gaps later on. I also hope to be able to express the variance on these figures somehow (CR 17 was especially swingy).

Kudos to ILM for their work, but it seems to include all enemies instead of just melee threats, and the nature of the graph hides repeated entries so it's hard to get an accurate sense of where the numbers really are. It's still a good baseline though, and does present slightly different data than I'm going for, so both are useful for separate things.




For those curious, preliminary results indicate average Melee Threat AB rises by about one and a third per level. I'll try to give a more precise figure later on.

[Axier]

All I know is that by level 10, using Ardent dips, good equipment, and a few feats, I can get a flatline AC of about 30-40 for ALL of my ACs, using something that gives me basic uncanny dodge; and the psionic feat that adds heavy armor AC to touch. Sure, anything with incorporial touch hits me unless I have ghost touch armor, but not much will hit me. Train up concentration and take a dip into warblade to take care of most of the saves, with a ring of evasion, its pretty solid. Sure, I cant do much else, but I make a great team distraction.

[Rejusu]

Firstly to everyone making graphs you're triggering a major pet peeve of mine: not labelling your axis's! I mean yes it's not hard to work out what each axis is supposed to be, but it's no excuse for not giving them nice handy labels.

Secondly, we can determine the chances of melee enemies hitting your AC until the cows come home. But does it really answer the question of how worthwhile AC is? I mean it's probably too complex to really work out with numbers, but there's a lot of other factors that need to be considered (and several have already been raised) like:
- What about enemies that don't target AC?
- Is AC more worthwhile than simply turning the enemy into a fine red mist before he even has a chance to hit you?

As an example in the game I'm currently playing I'm running a Half-giant psychic warrior. Now my AC is 17 (6 from banded mail cause I couldn't afford full plate and 1 Dex) which at ECL3 is in the "decent" range. However I can churn out enough damage to kill a CR3 creature in a few rounds. What's more thanks to my high trip modifier (+12 against medium creatures) I can give most melee threats a -4 to their attack rolls by putting them on the floor.

Last encounter we did was against about 9 enemies (which were frankly very over statted for the party level) and though it was tough we managed to wipe them out. Worth noting too that several of the enemies were archers and that by putting them on the floor I rendered them completely unable to even attack, let alone get past my AC.

[ILM]

Fine then, no dragons and bruisers only? Let's assume those are the guys getting better than 3/4 BAB/HD because I don't have the time to go through it line by line, and it should give a proper idea:



278 monsters left. Red line plots the average for each CR. As you can see, it doesn't mean much past CR 15 when there are like 3 monsters to pick from. I'm still seeing a near-linear increase until 15; beyond that, there's not enough data anymore.

If you include the Dragons, though, like I like to do because I'm stubborn, here's what you get:



For the math-inclined here, the average can be approximated to AB = 1.9*CR with a R² of 0.99. Linear.


Still, like Rejusu reminds there (and you can blame my laziness for the axes), however AC plots out, it doesn't help at all against the non-bruisers who have alternate modes of attack: 205 monsters out of 575 in the SRD, or 35% of everything you're likely to encounter.

[Amphetryon]

Still, like Rejusu reminds there (and you can blame my laziness for the axes), however AC plots out, it doesn't help at all against the non-bruisers who have alternate modes of attack: 205 monsters out of 575 in the SRD, or 35% of everything you're likely to encounter.

ILM, I don't suppose you could apply your graph-fu to plotting out where those 35% of non-bruisers lie in the CR scale, could you? It would provide evidence for or against the claims (which I tend to agree with) that most of them reside above level 7.

[sonofzeal]

Now with axis labelled!

Mr McHumanFighter here had a starting dex of 12, no wizard or cleric buddy for GMWing, uses a Greatsword, and is restricted to SRD-only content. He spends something less than half his wealth on AC - I aimed for a third, but mostly just eyeballed since it's past midnight here after a full work week and my ability to calculate thirds of six digit numbers in my head diminishes at that point.

He's following the rule of thumb of 50% WBL on defence, 25% on offence, 25% on utility. I heard it on these boards not long ago, and have been using something similar myself for a while. The gap between what he's spending on AC and 50% of his actual wealth can go into Cloak of Resistance and banking towards that Cloak of Displacement.



As you can see, he struggles through the mid levels but zooms ahead towards the end.

If anyone wants details, let me know.

[sonofzeal]

ILM, I don't suppose you could apply your graph-fu to plotting out where those 35% of non-bruisers lie in the CR scale, could you? It would provide evidence for or against the claims (which I tend to agree with) that most of them reside above level 7.

Having just gone through most of the SRD monsters (or so it feels)... it seems like lvl 8 is a breakpoint. At lvl 8, there's suddenly a relatively large contingent of monsters that don't target AC. That ratio actually seems to stabilize somewhat going up - most top level monsters are capable in melee along with whatever magical or supernatural power they bring to the table. But yes, the dividing line seems to be around lvl 7.

That said, the number of solidly non-melee monsters is still largely in the minority. Any random monster out of the book, and chances are it's got some way of potentially kicking tail in melee.

[ILM]

ILM, I don't suppose you could apply your graph-fu to plotting out where those 35% of non-bruisers lie in the CR scale, could you? It would provide evidence for or against the claims (which I tend to agree with) that most of them reside above level 7.

I don't see a particular pattern, but this tests the limits of my 3/4 BAB approximation. First of all there's a lot of low HD/CR creatures simply because 1 BAB when you have 2 HD is still below 3/4 (even though a Hyena, for instance, which falls in this category, is a perfectly valid bruiser), and there are a few creatures that fall into this categody simply by virtue of crazy amounts of HD, such as Greater Elementals (15 BAB, 21 HD, CR 9). I was really just trying to get a rough approximation and of the 205 monster entries this method removed, only 19 are CR 10 and above. I think if you want something more specific you'll need to defer to sonofzeal's work, since he seems to have gone into the details of each and every entry. Then again, it's not like there's always a clear line between bruiser and caster: is a Pit Fiend a melee threat or a primary caster?

edit: for the record though, I share your feeling. I just get the idea that even creatures that aren't supposed to be mainly meleers got decent BAB from WotC.

[sonofzeal]

I think if you want something more specific you'll need to defer to sonofzeal's work, since he seems to have gone into the details of each and every entry. Then again, it's not like there's always a clear line between bruiser and caster: is a Pit Fiend a melee threat or a primary caster?

Thanks! The OP now has a pretty extensive list, btw.

As for Pit Fiend... well, it's obviously a caster-sort. I don't think many people would look at its stat block and figure that its first tactic is going to be biting faces off. That said, WotC gave it six natural attacks, Improved Grab, Constrict, and a nasty poison. Melee violence is obviously something it's expected to do reasonably well at, compared to its CR. For my purposes, that's all that matters.

Aboleth Mage is a much more dubious case. It's got natural attacks that do nasty things, but it also has 10th level Wizard casting... but 5th level spells when the PCs have 9ths? IDK. It's a hard one to place. I ended up including it, but assigned it a lower weighting, kind of averaging between including it and not.

[ILM]

Thanks! The OP now has a pretty extensive list, btw.

Yes but you see, I think that's essentially where our divergence lies: between all the stuff you excluded, you have 11 data points from CR 15 to 20. There are 44 dragons in that range and a total of 66 monsters by my count. You're excluding 80% of your sample.

At high levels, I think dragons aren't the exceptions, they are the curve.

Oh, and @Malachei: if memory serves I found some kind of dump of the SRD in csv format somewhere. I just extracted the data I needed and formatted it for easy sorting and stuff. If you PM me an email I can send it to you on Monday.

[GoodbyeSoberDay]

Blindsight and Mindsight both spring to mind.

Mindsight doesn't counter miss chances except for Invisibility and the like. I suppose pinpointing a creature allows one to ignore mirror images? Still, there's Blinking.

I'm not saying miss chances are useless, I'm just saying it might be countered. Also, there are items with true seeing, such as the Scout Headband, usually considered a must have items for melee characters.

Miss chance counters are usually only available at much higher levels.

what is this? I can't find it in the SpC, it sounds useful (if situational)

Races of Eberron, removes illusions in X area for a little while, 8th level spell. That's what kills it.

[Malachei]

Oh, and @Malachei: if memory serves I found some kind of dump of the SRD in csv format somewhere. I just extracted the data I needed and formatted it for easy sorting and stuff. If you PM me an email I can send it to you on Monday.

Thanks. PM underway.

[Spuddles]

ILM, I find your data and analysis thorougly convincing. Monster AB is extremely linear. Also, you rock. I ****ing love graphs.

[Piggy Knowles]

Mindsight doesn't counter miss chances except for Invisibility and the like.

Mindsight is wonderful for a lot of things, but ignoring miss chances isn't one of them. An invisible enemy detected via Mindsight will still have total concealment:

This works much like blindsense--the creature knows what square each thinking being is in, but it does not see the being, and the being still has total concealment unless the creature can see it by some other means.