The D&D 5e rules for barkskin:

[…] the target's AC can’t be less than 16, regardless of what kind of armor it is wearing.

In 3.x, barkskin provided "a +2 enhancement bonus to the creature’s existing natural armor bonus." In 5e, the wording has changed to the above. This wording has been uniquely contentious (see Background below). I did not participate in the D&D Next playtest. My understanding is that barkskin worked more like it did in 3.x, in that it provided a (natural armor?) bonus to armor class. I would like to know if anyone, through participation in the playtest, or through analysis of the changes that bounded accuracy has brought to combat rolls, can shed light on why the new wording was adopted.

The question is asked in support of creating a homebrew version of barkskin that is more like 3.x, but can work within the context of 5e and bounded accuracy. Background discussion follows.


The interpretation of the effect of the barkskin has been a uniquely contentious. Two long threads have looked at it on ENworld (1 and 2), and How does barkskin work? has also been addressed on this site.

In summary, the "anti-stacking" crowd has argued that RAW is straightforward, and that nothing, including shields, cover, or magical bonuses such as Bracers of Defense stacks with barkskin. The "stacking" crowd has (IMO) been somewhat weaker on justifying their side with RAW, but has tried to appeal to common sense in asking how getting partially behind a wall can reduce the protective effect of the spell.

I don't really want to revisit that argument. If Jeremy Crawford can tweet that shields stack with barkskin, and then reverse himself in a Sage Advice ruling, it is obvious to me that the spell is at least poorly worded, and possibly poorly designed.

At this point, I want to just rewrite the spell for homebrew. My main concern with just adopting it as-is from 3.x is how it interacts with bounded accuracy in 5e. So I would like to know about the history of its development for 5e.

  • 3
    \$\begingroup\$ Answerers please remember D&D Next playtest materials were and are under NDA and information covered by that NDA should not be shared on this site. \$\endgroup\$
    – mxyzplk
    Jan 29, 2016 at 1:40

2 Answers 2


I would say these are the relevant factors:

  • 5th Edition D&D is a different game than previous editions. It's not an update or revision, so there was no "change" made to Barkskin. Barkskin is a spell that makes the target's skin tough and resilient, making them resistant to harm. For 5th Edition, the choice was made to mechanically describe this effect in terms of changing the target's armor class.
  • Based on the desired power level of Barkskin and the appropriate range of protection that should be available for characters with access to it, the designers decided that setting a minimum Armor Class for the target was an appropriate mechanical effect.

Consider the power level of a "stacking" Barkskin as compared to other non-armor AC enhancements (various Unarmored Defense class features, Mage Armor, etc.). An AC of 16 that's improved by a shield is what a barbarian with very good attributes gets. It's better than what a comparable monk gets. AC bonuses that accumulate are restricted to prevent them from over-accumulating (a character can only wield a single shield, and any magic items that grant AC bonuses require attunement).

I agree that it's an unusual effect, but I think that the description is clear, especially if you read it with the mindset that seems to apply to 5e spells -- that they do what they say, and nothing else. "The target's AC can't be less than 16" is unambiguous. It doesn't provide a new formula for calculating AC. It doesn't say anything about a new "base AC", or anything like that. It just says that after you calculate AC for the target, you take the higher of what you calculated and 16.

The earlier question you reference describes the spell effects as "vague" and "ambiguous" because:

... you have to ask what AC is it referring to? Total AC, Base AC or just armor AC?

But "total AC", "base AC", and "armor AC" aren't 5e concepts; the spell description is only poorly worded if you are making assumptions about 5e based on other game systems.

As always, it's your game. If you want, you can have Barkskin have the effect of "Until the spell ends, the target's AC is 16.", and treat it like any other AC-setting calculation. Or you could make it "Until the spell ends, the target's AC is calculated as if they were wearing chainmail."


I realize this question was asked some time ago, but the existing answer did not respond to it. The question asked was what advice might be given for constructing a homebrew version of Barkskin that retains the flavor of 3.x but is consistent with 5e rules.

As I am currently dealing with the problematic nature of the barkskin spell in my own campaign, I took to the web for help and have found that there is not much help to be had. While there are lots of responses in the vein of "just use it as it's written", they are not very satisfying to the folks that can't help but ask "why is it written that way?"

Likewise, interpreting (or defending) a rule does not do much to address the fact that it is a poorly written rule. I say it is poorly written because it is internally inconsistent with all the other rules dealing with AC, and because it has in actuality created significant amounts of confusion and frustration among players of the game.

So to answer the original question, here are some suggestions for how to adjust Barkskin to make it work more like everything else in the game already does.

TL;DR is at the end of this post

I think there are 3 threshold questions to be asked:

  1. Should the spell involve natural armor rules?
  2. How does the protection granted work with other adjustments to AC (stacking)?
  3. Should the spell be scalable? (In other words, should it's power increase if cast with a higher level spell slot?)

1. Natural Armor

3.x barkskin affected the target's natural armor, which was one component of total AC, along with armor, shield, deflection, dodge, and enhancements. In it's simplest form, it gave +2 enhancement bonus to natural armor, in the way that one might have +2 armor or shield. In most cases that just meant that it gave +2 natural armor bonus. This bonus was then added to armor, shield, etc to determine total AC.

Of course, this isn't how AC works in 5e. Natural armor is still in the rules, but is is clearly designed as an alternative to (artificial) armor. So, many creatures "have" natural armor, whereas the PCs "wear" armor. They are meant to be mutually exclusive. Because of that, it renders the distinction largely meaningless in game terms (they function in exactly the same way), and it really serves only to show the difference between thick/magical skin on creatures and (usually) metal clothing worn by people.

Given this, it is up to the homebrewer to decide if they want to classify barkskin as natural armor or not, but it is clear that it CANNOT stack with other armor being worn. In a way, I think this is what the designers were going for with the 5e wording of the spell, they just didn't explain it very well.

So "natural armor" or not is up to you, it's all just flavor text. In game terms it replaces worn armor (if any), but only if it is a higher AC than that armor.

2. Stacking - how does it work with other AC modifiers? In his response on the D&D site Jeremy Crawford doubled down on the spell as written, saying that under the effect of the spell your AC is the higher of whatever your AC is normally OR the number 16. That 16 number cannot be modified in any way. If something adjusts your normal AC to make it higher than 16, you get that instead, but the 16 itself will never change.

While I agree that this is the proper execution of the rule as written, it doesn't change the fact that to many people that rule is unsatisfying. I believe the dissatisfaction comes from two things: it is unsupported by any story-driven rationale, and it is arbitrary. Nothing else in the game works this way with regard to armor class. It is an outlier, and a somewhat non-sensical one when you start playing out scenarios:

  • (AC 15) chain shirt and shield
  • (AC 16) barkskin
  • (AC 16) drop the shield because it doesn't matter
  • (AC 17) pick up a +2 shield, because NOW the shield matters (for reasons)

So how do we craft a rule that makes sense? The answer is fairly obvious - just make it work like normal armor. Call it natural armor if you like, but make it fit within the existing rules. In essence, the mage armor spell works like +1 studded leather armor. So make barkskin a slightly better version. Two suggestions:

  • Barkskin provides AC 14+dex modifier (max 2) - essentially a breastplate
  • Barkskin provides AC 16 (no dex mod) - essentially chain mail

OPINION: I would argue for the breastplate option (although in story terms barkskin is in no way analogous to a breastplate). I think medium armor effects make more sense than heavy armor. If you want heavy armor effects, modify the stoneskin spell in the same way (since normal weapon resistance isn't often that helpful by the time 4th level spells are available anyway).

Once you make this adjustment, everything else works itself out and it fits within the rules, just the way mage armor does. It is sturdier than mage armor, but it has a shorter duration (8 hours to 1 hour - I would remove concentration because that destroys any cost/benefit analysis for using it). This solution also resolves all your issues with alternative AC determinations (like Unarmored Defense). If the target if affected by Barkskin, they are considered "armored", and abilities that require a character to be "unarmored" cease to operate.

3. Scalability I think the last thing to consider is whether to make the spell scalable the way many damage spells are. Again, I see two interesting options:

  • An additional +1 to AC for every two spell slots higher (15 at 4th, 16 at 6th, 17 at 8th, all with +2 dex modifier max)
  • Removal of dexterity limitation at a higher level, say 5th or 6th (14 + unlimited dex, making it effectively +2 studded leather.

I think either of these works and allows the spell to have some utility as PCs reach higher levels. I think the first one is better because it keeps the flavor of medium armor. If someone wants to use an 8th level spell slot to give themselves a 19 AC for an hour, I don't see any reason why they shouldn't be able to.


So, here's my suggestion for a barkskin spell that reconciles better with existing 5e rules.

Barkskin- 2nd-level transmutation

  • Casting time: 1 action
  • Range: Touch
  • Components: V,S,M (a handful of oak bark)
  • Duration: 1 hour

You touch a willing creature. Until the spell ends or the caster uses an action to dismiss it, the target's skin has a rough, bark-like appearance, and effectively gains the qualities of medium armor (AC 14 + dex modifier [max 2], no disadvantage to stealth). If the target does not have medium armor proficiency, they are considered proficient only for the purposes of this spell. If the target is wearing other armor, only the higher of the two AC calculations is used (the effects do not stack). The target of this spell is considered armored and may not use abilities that require them to be unarmored. At Higher Levels. If this spell is cast using a higher level spell slot, the base AC increases by 1 for every two levels higher (15 AC for 4th level, 16 AC for 6th level, 17 AC for 8th level).


Not the answer you're looking for? Browse other questions tagged .