Inside of a ring or out, ain’t nothing wrong with going down. It’s staying down that’s wrong.
~ Muhammed Ali
In my last post, Bench-Pressing: Character Creation by the Numbers, I outlined some benchmarks for Armor Class (AC). I stand by the benchmarks I outlined, but there are intrinsic problems with using AC as a defensive benchmark. Some characters will sacrifice their AC for more HP or DR. Some will use spells like displacement to levy miss chances against the enemy. Others might use lay on hands to heal themselves and extend their combat mileage. In any case, there are any number of defensive tactics in Pathfinder beyond AC that will effect how long your character survives attacks.
This article will outline a process by which you can evaluate a character’s “survivability” in terms of enemy attacks. AC will be a component of survivability, rather than the whole. I will not examine an exhaustive list of every defensive measure in the game system, but the process I outline should be simple enough to replicate for any given character evaluation, regardless of what defensive measures they employ.
The benchmarks, as I see them, should be measured in how many full-attacks from a typical CR appropriate monster you can take before dying. A Blue rated character can take 4+ full attacks, a Green can take 3, an Orange can take 2, and anything below that is Red.
Blue ratings are extremely hardy, those who find themselves constantly in the thick of battle (melee range, usually) and is the most likely person to be attacked. 4 full attacks means even if they go last, they (and the rest of the party) will have 3 rounds of actions to contend with the threat before retreating or going down. I have rarely seen combats last more than 4 rounds against a single enemy. Basically, a Blue character can survive even the worst possible circumstances against a typical enemy’s attacks.
Green characters are quite well suited to melee as well, though they can’t simply laugh off a few rounds of hits. They are good at absorbing attacks, but if they get isolated from the party/escape routes for too long, they could actually be in danger.
Orange characters generally aren’t trying to weather melee combat, but if something sneaks up behind them and gets a good round in, they should still be on their feet with time to take defensive measures or retreat. This is a perfectly fine benchmark for ranged characters of any kind to aspire towards.
Red should be avoided by anyone, however. A Red character can easily be ‘ganked’, or killed before they even get an action. Surprise rounds, bad positioning, or even just losing initiative are events that occur every so often. Having a minimum of an orange survivability benchmark will drastically reduce the likelihood that any of these events results in an outright character death.
While Average Monster Statistics can give us some useful information, the chart unfortunately lacks the number of attacks an average monster has. For this reason, I’ll simply use one of the more commonly referenced monsters for benchmarking/optimization exercises: the Fire Giant. As such, we will assume our evaluated character to be level 10.
The Fire Giant has three attacks. When Power attacking, his full attack is +18/+13/+8 for 34.5 average damage. We will basically benchmark his EDV against ourselves, giving us an Expected Damage Taken or EDT.
We can now imagine the worst-case scenario in terms of HP: a villager with AC 10, and no spells or special abilities to mitigate the damage. This villager is going to take 108.2 EDT/turn. If he wanted to be Blue rated, he’d need 433 HP. That would be…difficult to achieve.
Now let’s suit the villager up in +4 Full-Plate with a +3 Heavy Steel Shield. His AC is now 28, but he still has no other damage mitigating abilities. Still, the EDT/turn diminished to 35.8 per round, meaning he’d take 107 HP or so in three full attacks. If we had 16 CON, we’d need 92 HP to stay green. A d10 class will be comfortably green in survivability, a d8 can get there by putting their favored class bonuses into HP.
Using a survivability metric rather than raw AC means we can even incorporate more esoteric defenses. The displacement spell, blinding an enemy, a failed save against debilitating portent, or having complete concealment from another source cuts enemy damage in half. Let’s take our villager’s fancy shield away (leaving him with 23 AC) and cast displacement on him.
At AC 23, the Giant’s EDV against our villager jumps up to 62.6 meaning he’d need (188 – CON Score + 1) to get a green rating. That’s unlikely to work out, but with displacement on top he needs only (94 – CON + 1), as half of the Giant’s successful attacks will now fail outright from the miss chance. Even a d8 class with 14 CON can achieve a green rating in these conditions by putting their favored class bonuses into HP.
Let’s look at DR for a moment. We’ll say our villager is secretly a 10th level earth kineticist, rocking DR 6/adamantine. We’ll take away the displacement, but keep the +4 Full-Plate. DR only helps us when we get hit, so we need to know how many hits we take in an average turn. The giant’s power attack bonuses are +18/+13/+8, meaning we get hit 75% of the first attack, 50% of the second, and 25% of the last. So, we take 1.5 hits per turn on average, meaning we reduce damage by 9 per turn. Over the course of 3 turns, that’s 27 HP, meaning he’ll need (161-CON + 1) HP to remain Green rated. So, DR is a nice little bump to survivability, but compared to miss chances or AC, it isn’t doing much. This shouldn’t surprise us, considering that by this level monsters will not uncommonly possess DR 15 or so. The system simply doesn’t want players to have meaningful DR in their possession.
I’ll wrap up with my favorite defensive measure in the game: mirror image. At 10th level, we’ll have an average of 5.5 images. We’ll say our caster has 14 Dex, shield, and mage armor active for a total of 20 AC. On the first attack, the giant can only miss on a natural 1, meaning his damage output before images is (34.5 x 0.95), or 32.775. This then gets divided by (1 + avg. # of images) to find the EDV against the caster, which is 5.04. The next attack hits 70% of the time, for an average of 24.15 damage pre-images. We divide by 5.5 now (as one of our images broke on the first attack), to get an average damage of 4.4. There is actually a 5% chance that this attack didn’t break an image, meaning our next attack will go against 3.55 images. So, (34.5 x 0.45)/4.55 gives us 3.412 average damage for the third attack. Our total Expected Damage Taken is a mere 12.85, despite having a pretty pathetic AC.
The main issue with mirror image, displacement, and other illusion spells is that true seeing becomes nearly ubiquitous as a constant spell-like or supernatural ability in high level play. I honestly think that is a major flaw with the Pathfinder system, as a huge variety of spells (including the entire illusion school) becomes completely useless after level 12 or so.
In any case, I hope this guide provides you with objective methods for determining how well your character can weather the slings and arrows of outrageous attack rolls.