I just saw this question (this comment especially) and started wondering, what happens with a shield guardian loop?

More specifically: what happens when a number of shield guardians wear each other's amulets thus creating a loop in which damage would travel?

Where does the last one damage due to rounding go as it is transferred through the loop. Does it float in the link until one of the shield guardians dies?

| improve this question | | | | |
  • \$\begingroup\$ This has (or is likely to) come up in your game? \$\endgroup\$ – Dale M Mar 9 '18 at 1:28
  • 1
    \$\begingroup\$ Though this is not likely to happen I think this is worth looking into since it is a mechanic that could possibly be exploited. I think this question is more preventing this becoming a problem than wanting a solution to a current problem. \$\endgroup\$ – rpgstar Mar 9 '18 at 1:39
  • \$\begingroup\$ I'm unsure what is being asked here. Are we talking about two shield guardians, each of which is wearing each other's amulet? \$\endgroup\$ – Gandalfmeansme Mar 9 '18 at 1:42
  • \$\begingroup\$ @Gandalfmeansme that is what SevenSidedDie refered to as a "shield guardian loop" in the comments of the last (current) answer in the linked question. \$\endgroup\$ – rpgstar Mar 9 '18 at 2:06

The last point of damage will continue to float between the guardians until the chain is broken

When one of the guardians takes damage, then half of that damage remains on the guardian, and the other half gets passed onto the next guardian. That guardian will then take half and pass half on. So forth and so on, until there is only a single point of damage left. Because the damage being passed on is rounded up, that half of 1 becomes 1, and that one point is continuously passed between the guardians.

The "floating" point (pardon the pun) has no effect.

The creature being protected by a guardian never feels the effects of the damage that is redirected to the guardian. If you are at 8 hp and are dealt 10 damage you do not start making saving throws before the guardian takes 5 of that damage away from you. Since each of the guardians is being protected in the same manner, none of them ever feel the effects of the point.

When the loop is broken, the point will immediately be permanently applied to whomever is now last on the chain.

Each instance of damage will create it's own looping point.

Multiple looping points of damage would not be in sync with each other, and therefore they would not be able to split and leave half behind. Even if you wanted to argue that that is what should happen, there's no way to determine which guardian they should end up on.

In theory you could use this to "store up" an arbitrary amount of floating damage, which would all be immediately inflicted on the last guardian of the chain once the chain is broken. But unless you can figure out how to make that last link something you want to kill, I see no practical way to abuse this loophole.

| improve this answer | | | | |
  • \$\begingroup\$ But such damage transfer is "instant" as far as RAW goes. It doesn't transfer forever, it transfers an infinite number of times. Or rather, it doesn't transfer forever in the game world, but if you do this then RAW you've effectively frozen game-time relative to real-time, since the damage never resolves and you can never go to the next action. \$\endgroup\$ – stellatedHexahedron Mar 12 '18 at 18:10
  • \$\begingroup\$ @stellatedhexahedron where does it say that the tranfer is instant? \$\endgroup\$ – Arcanist Lupus Mar 12 '18 at 19:13
  • \$\begingroup\$ "half of any damage the wearer takes (rounded up) is transferred to the guardian". Not "transferred to the guardian after a certain amount of time", just "transferred to the guardian". I suppose you could argue that technically, it doesn't say the effect is applied immediately, but if you do so you've opened the door to arguments such as "Technically, the Gorgon's gaze just says I turn to stone, not that it happens immediately!" \$\endgroup\$ – stellatedHexahedron Mar 13 '18 at 14:53
  • \$\begingroup\$ I don't think that the game distinguishes between "instantaneous" and "to fast to measure" \$\endgroup\$ – Arcanist Lupus Mar 13 '18 at 17:48

This is definitely open to the specific GM making the ruling. I would start with "All damage must be immediately applied somewhere" and go from there. The idea of missing damage or "storing" damage until someone takes off an amulet and they spontaneously explode seems ridiculous to me. The possibilities, in order of my preference, would be:

  1. A guardian can only share a particular instance of damage once. This means that once the guardian that took damage takes part of it again (once around the loop), they stop dividing it.
  2. The last point of damage does not get redirected (or gets redirected around the loop exactly once)
  3. Shield Guardians will not protect Shield Guardians (This may be the most RAI answer)
| improve this answer | | | | |
  • 1
    \$\begingroup\$ @SevenSidedDie I removed the last 3, since they were either for humor or stuff I don't think makes sense, and not my preferred. I think the remaining 3 are all valid choices, and definitely disagree with the currently upvoted answer. Unfortunately, nobody has given me a reason for downvoting so shrug \$\endgroup\$ – Kyle W Mar 12 '18 at 15:43

None of the golems suffer the last point of floating damage

There is a simple and elegant solution to your question that is already supported by one of the mechanics in the game: just letting it silde.

This is what happens in the case of damage resistance (which is very similar to how our golems work): when the damage dealt is odd, you take half damage rounded down - this rounding down leading to a "floating" point of damage that just gets lost into the ether. This interpretation is further supported by how the Warding Bond spell works, where in the case of odd damage, both creatures take half damage (rounded down), and one point of damage is dissipated:

This spell wards a willing creature you touch and creates a mystic connection between you and the target until the spell ends. While the target is within 60 feet of you, it gains a +1 bonus to AC and saving throws, and it has resistance to all damage. Also, each time it takes damage, you take the same amount of damage. The spell ends if you drop to 0 hit points or if you and the target become separated by more than 60 feet. It also ends if the spell is cast again on either of the connected creatures. You can also dismiss the spell as an action.

While the golem trait does not work exactly like this, as it states that the transferred damage is rounded up for the absorbing golem, in the case of a loop I would argue that this is the most sensible interpretation and the closest we can get to a RAI resolution, without unnecessarily complicating the game with extra bookkeeping for multiple instances of floating damage and staying true to the spirit of the rules.

| improve this answer | | | | |
  • \$\begingroup\$ I like this answer, but the Bound trait does not function exactly like Warding Bond since the transferred damage in the former case is rounded up.. This could be used to support that it is a possible ruling; i just take trouble with the statement: "this is exactly what happens..." \$\endgroup\$ – David Coffron Mar 9 '18 at 9:07
  • 1
    \$\begingroup\$ @DavidCoffron You're right, but I think it's the closest we can get to a RAI resolution to this, without unnecessarily complicating the game and staying true to the spirit of the rules. And it would be how I would handle the situation. :) \$\endgroup\$ – Olorin Mar 9 '18 at 9:11
  • \$\begingroup\$ @DavidCoffron I relaxed (to be read: removed) the "exactly" clause and included an argument for resolving the ruling like this \$\endgroup\$ – Olorin Mar 9 '18 at 9:18
  • 1
    \$\begingroup\$ I think the RAI solution is that shield guardians aren't intended to wear each other's amulets. There's nothing explicitly against it, but the lore makes it clear that they're intended to be bound to wizards, and since shield guardians can't speak they can't command another shield guardian whose amulet they are wearing, which limits the functionality anyway. So my conjecture is that there's no point figuring out a RAI solution to the floating hit point when I suspect RAI would prevent this from working from the start. \$\endgroup\$ – Bloodcinder Mar 9 '18 at 13:21

All damage is continuously transferred .

Since the shield guardian's damage is cut to 1/2 each (with respect to the two involved in the loop at the time), only 1 of the shield guardians leaves damaged before the next shield guardian's damage transferrence takes effect.

This is because in 5e damage is generally rounded down. The shield guardian who takes the 1 damage has half of the damage removed rounding down to 0, while the transferred shield guardian's 1/2 is rounded up to 1. This continues indefinitely in a loop so it is unclear where the damage ends up.

Bound: ... half of any damage the wearer takes (rounded up) is transferred to the guardian.

So what happens

The result is a divergent series that doesn't have an obvious value since each shield guardian either takes 0 or 1 at any given time. It is possible various summation techniques may work such as Cesaro summation which gives a value of 1/n where n is the number of shield guardians. I would rule that this is the damage left over after the rest is transferred to the other shield guardians which should be rounded down to 0.

| improve this answer | | | | |
  • \$\begingroup\$ Let us continue this discussion in chat \$\endgroup\$ – David Coffron Mar 9 '18 at 8:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.