D&D (2024) Illusion Magic in 2024

[Kobold Stew]

While there is more information in PHB 2024 about Illusions thanin 2014, there still isn’t enough, and problems persist in play. What follows is my trying to make sense of the rules so that illusions remain effective but are not game-breaking. Thoughts welcome. I am trying to get playable rules that work.

tl/dr: Illusion magic rules are poorly written but are only meaningful if they are not trivially overcome. Attacking an illusion to make it fade from sight is trivial and by itself should not count as "physical interaction". The final section presents the guidelines I think are implied by the PHB, and are (I feel) balanced for players and NPCs.

Preliminaries.
Illusion magic is cool, and it should be fun for both the player and the DM. The nature of illusion magic in combat (when things are most regulated; a consistent interpretation for combat will solve other applications), isn’t to cause damage (though some illusions do that), but to change the enemy’s behaviour; it is a form of battlefield control, with a lot of narrative freedom for the caster. Against a single opponent, illusions might affect a chosen target; in a crowd, many might be distracted or confused. However, Illusion rules have to be reciprocal, the same for a PC and an NPC, and that’s where the issues start. Illusions aren’t a “win button”, but they also can’t be trivial. (Everything that follows in PHB 2024 only – no legacy features or interpretations.)

Rolling to Disbelieve.
Central to an understanding of illusion magic is the way they are countered (what in AD&D we called rolling to disbelieve). The wording of Minor Illusion (cantrip) and Silent Image (level 1) are sufficiently similar to establish ground rules. For visual illusions with both spells:

1. “Physical interaction with the image reveals it to be an illusion, since things can pass through it.”
2. A creature can take a study action to examine an image. This can “determine that it is an illusion with a successful Intelligence (Investigation) check against your spell save DC.”

Additionally, “If a creature discerns the illusion for what it is, the creature can see through the image.” In both spells, this appears immediately after point 2. It is at least possible, then, that this is a corollary of case 2 – i.e. the discernment and the ability to see through the illusion is a result of making a successful Study action. In this case, to see through an image (a) requires an action, and (b) a successful Intelligence check. This is a high threshold: most monsters are not proficient in Investigation, and the DC of many casters will be quite high. An investigation check can (and often will) fail, and it will take an action.

If case 1, physical interaction, similarly dismisses the illusion, the threshold must also meet a high threshold. Otherwise, case 2 is meaningless: physical interaction would always be preferable.

What does physical interaction do?
Physical interaction “reveals” the illusion to be an illusion. This is either the same result as a successful Investigation check, or a lesser result; and the most natural lesser result would be coming to the belief that something was an illusion, but not seeing through it. Natural language would suggest it’s the same effect. What do the rules say?

The PHB tells us (p. 24):

Interacting with Things. You can interact with one object or feature of the environment for free, during either your move or action. For example, you could open a door during your move as you stride toward a foe.

If you want to interact with a second object, you need to take the Utilize action. Some magic items and other special objects always require an action to use, as stated in their descriptions.

Interacting is tied to the Utilize action. This suggests to me that a character can have a “free” physical interaction with something within reach once per turn if they believe it is an illusion.

Do other activities count as physical interaction? There is no guidance, but I think the presumption is no. Many play that if you fire an arrow at a suspected illusion you reveal it at range. If so, the bar becomes trivially low, and this is always a better strategy.

If Interaction does create the same result as an Investigation check, then the offsets have to be vaguely commensurate. To my eye, they are.

Investigating:

• INT-dependent chance for failure.
• Can be done at range (though one can imagine a DM imposing penalties, such as disadvantage at more than 30’ or in dim light).
• No risk to self (of falling, burning yourself, whatever).
• The Keen Mind feat lets you Study as a Bonus action.

Interacting:

• no chance of failure; auto-success at melee range
• one “free” attempt per round; otherwise costs an action.
• Risk of injury to self
• Possibly requires a free hand (DM call).
• A Rogue Thief can Utilize as a Bonus Action.

Interacting will be preferred most of the time, especially since it can be “free”, but at least there are natural circumstances in which either of these might be used. The separation of the description of the cause (physical interaction) and effect (seeing through the image) is suspicious but may just be bad writing.

Attacking an illusion
Attacking an illusion, with the intent of doing damage, is mechanically different. Normally, if your eyes are closed but you know which square a target is in (when playing on a grid), you attack with disadvantage (Cover is different, and illusions do not provide cover). There is the possibility of doing damage, however to the target. If additionally this also let you see through an illusion, then this is always optimal, and the only reason to Investigate or Interact is to use the one free interaction: the Study Action and the Utilize Action are never preferred.

Attacking:

• takes an action
• Uses optimized attack stat and eventually magic bonuses
• No risk of failure to cause illusion to fade
• Possibility of damage to anyone inside
• Can be done at range
• No risk to self
• Costs a piece of ammunition.

If an attacker knows (or accurately suspects) where the target is, they can attack with disadvantage. By asking for a roll with Disadvantage, the DM signals something is amiss to the player, which may make them Study or Utilize, but it does not in itself cause the illusion to fade.

Consider, now, a gnome rogue casting Minor illusion of a 5’ crate with holes in it (or slats, whatever) as something to Hide within. Yes, you can Hide in it, and the enemy might not see you, but they have reason to know where you are, and can attack that location (at disadvantage, which more or less corresponds to ¾ cover anyways). Depending on the nature of the illusion, you might not need to Hide: an illusory door in a doorframe may arouse no suspicion at all (DM’s call).

Anyone witnessing such an attack, though, may suspect something is an illusion, and either continue to fight through it (at Disadvantage), or spend an action (or a free interaction) to try to disable it. But, importantly, each person seeing the illusion must see through the illusion on their own (it’s like those Magic Eye autostereograms).

Suspecting an illusion
You might believe something is an illusion even though you can see it. This could happen for a number of reasons:

• Someone tells you it is an illusion and you believe them (it is not enough for someone else to have successfully seen through the illusion);
• You suspect from the description from the DM that something is amiss (e.g. context doesn’t match setting);
• You have just seen another illusion being cast (the “fool me once” principle);
• You have taken an attack and the attack has passed through the illusion.

Suspecting something is an illusion, or witnessing someone interact with it, doesn’t remove the illusion for you, but may prompt you to Study it (a Perception check may reasonably be required if doing something else, like fighting).

Illusion rules for the PHB 2024.
The clarifications I want are pretty minimal, and with them none of the above would be needed. Of the following, 1 and 2 are in the PHB, but not stated as clearly as I would like; 3, 4, and 5 are, I believe, implicit, but should be spelled out.

1. Two things cause an illusion to dissipate/become faint for you:

• Study action investigation check (usually within 30’ in bright light).
• Move within 5’, take free interaction (or Utilize action).

2. You can’t attack an illusion, but if you know where an enemy is within an illusion, you can attack it with Disadvantage.

3. A creature who is aware of the illusion, but does not dissipate it, is still impacted by the illusion.

4. An illusion does not provide cover, but it does block line of sight, giving something you can Hide behind or within.

5. Only those who have already seen through an illusion can offer Help to others.

An illusion will typically slow down an enemy, and make them change their behaviour. A spell is still constrained by concentration, area of effect, and perhaps range.

Finally, the Illusionist’s Improved Illusion ability has three effects: many illusion spells will have greater range; additionally, “You can create both a sound and an image with a single casting of Minor Illusion, and you can cast it as a Bonus Action.” The first of these makes an illusion more credible, and enemy viewers are less likely to want to Investigate. Its biggest impact will be out of combat, though. The second, though, means that Minor Illusion will be cast regularly in combat. And that’s why more robust Illusion rules are needed.

 

[Charlaquin]

Solid analysis of the illusion rules in both books, I agree with your interpretations. I do have a bit of a nitpick though:

Illusion rules have to be reciprocal, the same for a PC and an NPC, and that’s where the issues start.

D&D is an inherently asymmetrical game, so the rules for how NPCs interact with illusions don’t necessarily need to be the same as those for how PCs do, and in fact I think they probably shouldn’t be. Because when a PC encounters an illusion, the player doesn’t immediately know if it’s an illusion. But when an NPC encounters an illusion, the DM always knows that it’s an illusion. The DM can’t fairly make the decision of when/if the NPC should think to try to inspect the illusion, as their certainty that it is an illusion will inevitably bias that decision in one way or another. There really should be some mechanical way to determine when/if an NPC decides to inspect an illusion.

Personally, here’s how I would do it: when a PC encounters an illusion, simply describe to the player what the PC perceives, and leave it to the player to deduce its illusory nature from context. If the player has their character interact with the illusion in some way that would reveal its illusory nature, just skip the Investigation roll. They’ve already figured it out, the “roll to disbelieve” is just redundant.

On the other hand, when an NPC encounters an illusion, make an Investigation check right away (no action required), not to dispel the illusion, but to determine in an unbiased way if the NPC thinks it might be an illusion. If they fail, just have them treat it as if it were real, until and unless they witness something that proves otherwise. If they succeed, they suspect it of being an illusion, but still need to take an action to interact with it in some way to be certain. Again, no need to make a check when they do so, you’ve already made it.

 

[UngeheuerLich]

Solid analysis of the illusion rules in both books, I agree with your interpretations. I do have a bit of a nitpick though:

D&D is an inherently asymmetrical game, so the rules for how NPCs interact with illusions don’t necessarily need to be the same as those for how PCs do, and in fact I think they probably shouldn’t be. Because when a PC encounters an illusion, the player doesn’t immediately know if it’s an illusion. But when an NPC encounters an illusion, the DM always knows that it’s an illusion. The DM can’t fairly make the decision of when/if the NPC should think to try to inspect the illusion, as their certainty that it is an illusion will inevitably bias that decision in one way or another. There really should be some mechanical way to determine when/if an NPC decides to inspect an illusion.

Personally, here’s how I would do it: when a PC encounters an illusion, simply describe to the player what the PC perceives, and leave it to the player to deduce its illusory nature from context. If the player has their character interact with the illusion in some way that would reveal its illusory nature, just skip the Investigation roll. They’ve already figured it out, the “roll to disbelieve” is just redundant.

On the other hand, when an NPC encounters an illusion, make an Investigation check right away (no action required), not to dispel the illusion, but to determine in an unbiased way if the NPC thinks it might be an illusion. If they fail, just have them treat it as if it were real, until and unless they witness something that proves otherwise. If they succeed, they suspect it of being an illusion, but still need to take an action to interact with it in some way to be certain. Again, no need to make a check when they do so, you’ve already made it.

I agree. I would just describe it as the arrow passing through. At high range, tmit might not be obvious if it does so beacause of an illusary nature or some other effect.
An illusiory creature could as well be a ghost. So I might allow a free study check.

 

[Shiroiken]

As someone about to DM a campaign with an illusionist PC, I've been thinking a lot about how to manage illusions. My general view has always been that in a world of magic, most creatures aren't going to assume anything is an illusion without cause. This makes illusions strong, since this typically reduces interaction with the illusion, often requiring an action. For example, if the illusion is that of a cage, it's typically an action to break out of it, just like it would be for a real cage.

I waffle about requiring the Study action to see through the illusion though. Minor Illusion could be used to create a 5 ft cube of stone that someone could sit in and grant disadvantage to all attacks against them for 1 minute with no concentration. That's really powerful for a cantrip.

 

[jgsugden]

You make a few assumptions that underlie your assertions relating to this limited subset of illusions. Here are some challenges to those assumptions:

You speak as if attacking an illusion is always an option, but you can't always attack an illusion. First, you state that being able to identify an illusion with an attack renders the ability to identify with investigation (without physical interaction is "meaningless"). There are many situations, particularly outside combat, where attacking the illusion is not an option. As such, the ability to discern without the illusion being physically contacted is relevant in a significant number of situations. For example:

• Where the illusion is of something of value you are not allowed to touch and you do not know if it is an illusion or not.
• Where you can see the illusion, but it is behind a transparent barrier.
• Where weapons are not permitted to be used.

Example: The PCs are hired to protect an art gallery, but one of the paintings was stolen before they even began to work. They would not be allowed to attack or even touch the paintings for fear of damage to them. If they suspect the trick, they'd need to touch them (breaking rules of the job) or use investigation, generally.

One might argue that these situations would seem rare and would only occur with contrivance. That is mostly true - and is a feature, not a bug. Someone is making an illusion. They are attempting to deceive. They want to set up the situation so that the illusion is harder to detect. These meaningful restrictions on illusions limit their use intentionally. They create a story as part of the social and exploration pillars of the game.

You mistakenly say there is not guidance on interaction. Second, you assert there is no guidance on what constitutes physical interaction. There is, and you quote it. “Physical interaction with the image reveals it to be an illusion, since things can pass through it.” Because things can be observed passing through the illusion (usually through sight, but potentially with other senses), we can tell it is an illusion. If we observe something passing through the object we have revealed it through physical interaction.

You assume there are only two ways to discern these illusions. The rules specifically identify the investigation and physical interaction route, but the spell language does not limit to discernment to these two options. Discerning's most common definition means perception - but it is not limited inherently to sight. If your PC knew what a troglodyte was, and found a sleeping one in a tunnel, but noted no smell of troglodyte and then said, "I'm pretty sure this is an illusion - troglodytes smell and this one doesn't", the DM could make the call that they discerned through the lack of odor that the troglodyte is an illusion.

Note that the spells do not specify that a caster can see through the illusions inherently. If they want to see through their illusion of a wall, they need to discern it is an illusion. Do you require a wizard to put a finger in the wall illusion to be able to see through it? Or do you allow them to discern it automatically because they created the illusion and know what to look for to 'discern' it is an illusion?

Now for an assumption of mine: An investigation check can be made to discern the illusion, but what is an investigation check doing? It is used to figure something out. You're using it go to from a state of not knowing to a state of knowing through observation of a notable clue and interpretation of it. In this case, you're looking for flaws in the illusion that give it away, right? This is not explicitly spelled out in great detail (likely due to the restriction on how much space they can devote to any given spell in the books), but it makes sense.

So, what if you're not looking to find a flaw, but instead have been told there is a flaw by someone that has seen it already? You no longer need to investigate to find the flaw - so long as you can see the flaw that has been pointed out, you can discern the illusion. If my friend Bob saw an arrow hit the illusion before I arrived, Bob can see through it, but I can't. Bob can tell me an arrow passed through the illusion, and I might thus suspect that it is an illusion, but I can't discern it unless I interact with it or investigate it to confirm. However, if Bob investigated it and noted a defect that they can communicate to me, it is up to the DM on whether they can convey that information to me in a way I can see. Would that require a perception check, instead? That is up to the DM. They need to figure out what it takes. I don't roll a perception check to find a coin on a table or see an arrow shot at a target, usually ... so the DM needs to determine if the flaw is something that requires a perception check.

Other 'Solutions': If you stick with a more common and traditional interpretation of the rules for this subset of illusions, there are ways to beef them up that do not require you to change how they work for everyone, but allow a specific spellcaster (or spellcasters) to benefit from more leniency. In my setting, for example, there are guilds for wizards of each subclass. Membership allows you certain privileges (copying spells, discounts, lodging while traveling, easily identify guild members, etc...) and other subclass specific benefits. In the case of Illusionists, you get a common ring that allows you to buff illusions a bit. One of them is that you can upcast an illusion spell by one (additional?) level and strike the sentence "If a creature discerns the illusion for what it is, the creature can see through the image" from an illusion spell if it exists in the description. This gives the illusion of a wall some of the same benefits as a fog spell in terms of blocking sight.

 

[kigmatzomat]

I think there is one aspect missing: animated illusions.

• Minor image makes a static illusion.
• Silent image and Major Image can be animated, using the caster's action.

So assume an illusion of a humanoid statue. A sword or arrow attack at a Minor Image passes through and should justify some kind of "heeeyyyy..." reaction from the attacker.

A sword or arrow attack against a Major Image statue being actively animated, well, the statue can dodge and counter attack "wildly" so that it "misses". There's no implausibility here, so no rational for any disbelief. Only after some number of "Wait, neither of us even touched each other?" Sequences, or the "why doesn't the stone statue's footsteps make sound?" does it become implausible.

But what about the choice of illusion itself? Replace the statue with the illusion of a shadowy humanoid. The Minor Illusion might hold up for a round or two of combat as swords and arrows passing through it are part of shadow's scary factor. But when it doesn't attack, that's the "hey....." moment. Though a second Minor Illusion of faint, creepy whispering might extend it the realism for a round or two.

The animated Shadow illusion can "fight", and then swords and arrows missing or passing through while its attacks fail to land are more plausible. Not infinitely more plausible, but more plausible.

The animated illusions don't cause damage but they consume caster actions, ostensibly to consume opponent actions and/or inflict penalties, like "flanking" by distracting enemies.

More powerful illusion spells, like Major Image, include other sensory effects so an illusory ghost can include a "chilling" aura, an illusory fire elemental radiates "heat", or a stone statue's footsteps make sound. These should be harder to recognize. Which makes sense, higher level spells should be more effective.

I don't think it should be spell level based per se, but there should be increased benefit from a more powerful spell.

 

[kigmatzomat]

I also forgot to mention low-level illusions mimicking higher level spells. Silent Image does a great Wall of Fog. And if your allies know it's an illusion, they can quickly see through it. Or even better, make the illusion before combat starts so they can disbelieve then "animate" it forward to engulf enemies or the party during combat within 10 minute.

It's a lot more to disbelieve attacks passing through fog.

Or illusions of spells that create a glowy totem/spirit/glyph so that enemies just don't go near it. Major image does a decent Prismatic Wall, Wall of Light, or Wall of Fire.

 

[Kobold Stew]

Thanks for the responses!

Solid analysis of the illusion rules in both books, I agree with your interpretations. I do have a bit of a nitpick though:

D&D is an inherently asymmetrical game, so the rules for how NPCs interact with illusions don’t necessarily need to be the same as those for how PCs do, and in fact I think they probably shouldn’t be.

Fair enough, and a roll to see how an NPC reacts is totally fair.

They’ve already figured it out, the “roll to disbelieve” is just redundant.

On the other hand, when an NPC encounters an illusion, make an Investigation check right away (no action required), not to dispel the illusion, but to determine in an unbiased way if the NPC thinks it might be an illusion.

Something like this totally works, in that both sides then have a reasonable shot at a "free" investigation check. That's the degree of symmetry that I think is reasonable (even if the information is assymetric, as you note.
At the same time, what I see in the 2024 rules is that in most cases it still requires a non-attack action to make the image fade, even if you believe something is an illusion. What you describe (and the free study check suggested by @UngeheuerLich ) is closer to the Passive Investigation from 2014, which (from what I can tell) nobody liked or used.

I waffle about requiring the Study action to see through the illusion though. Minor Illusion could be used to create a 5 ft cube of stone that someone could sit in and grant disadvantage to all attacks against them for 1 minute with no concentration. That's really powerful for a cantrip.

It is powerful, yes. But it also changes the nature of the fight, which is an inherent good (in that it then rewards tactical responses). There are a lot of ways to push and pull characters, and this gives a reason for them to come into play.

If you don't require the Study action, though, then that whole mechanical feature will never get used. There is always a better course of action.

 

[Kobold Stew]

Thanks for the detailed response!

You make a few assumptions that underlie your assertions relating to this limited subset of illusions. Here are some challenges to those assumptions:

You speak as if attacking an illusion is always an option, but you can't always attack an illusion.

Everything you say for this point is true, but not relevant to solving the low-level problems that the rules present. I love these ideas, and yes out of combat illusions can do lots of fun things.

You assume there are only two ways to discern these illusions. The rules specifically identify the investigation and physical interaction route, but the spell language does not limit to discernment to these two options. Discerning's most common definition means perception - but it is not limited inherently to sight. If your PC knew what a troglodyte was, and found a sleeping one in a tunnel, but noted no smell of troglodyte and then said, "I'm pretty sure this is an illusion - troglodytes smell and this one doesn't", the DM could make the call that they discerned through the lack of odor that the troglodyte is an illusion.

Totally fair, and so we come to the basic tension that (in my view) the rules do not adequately resolve. The rough skill system distinguishes checks for Wisdom (Perception) from Intelligence (Study), and says, specifically, that Study is the way you normally use your senses to see through an illusion. Perception (the skill) might lead you to suspect an illusion, but the image remains there for you until you unfocus your eyes, like the Magic Eye sailboat.

Note that the spells do not specify that a caster can see through the illusions inherently. If they want to see through their illusion of a wall, they need to discern it is an illusion. Do you require a wizard to put a finger in the wall illusion to be able to see through it? Or do you allow them to discern it automatically because they created the illusion and know what to look for to 'discern' it is an illusion?

Ha! I had a mini paragraph on this (do you and your party automatically see through illusions), and my sense is no, they do not. You still need to unfocus your eyes successfully.

Now for an assumption of mine: An investigation check can be made to discern the illusion, but what is an investigation check doing? It is used to figure something out. You're using it go to from a state of not knowing to a state of knowing through observation of a notable clue and interpretation of it. In this case, you're looking for flaws in the illusion that give it away, right? This is not explicitly spelled out in great detail (likely due to the restriction on how much space they can devote to any given spell in the books), but it makes sense.

A single example would do wonders, and I wish one had been included. I don't buy the space argument -- this is the identity of a core subclass, and the rules aren't sufficient.

I don't know what the Investigation check does, and yes it might be looking for inconsistencies. But the Magic Eye explanation works too. (The Feats section gives reasons to believe that it isn't looking for inconsistencies, though: Keen Mind gives the ability to Study as a bonus action (what should be a basic feat choice for Illusionists); "keen observation" is part of the Observant feat, and it gives a bonus to Search. That to me suggests that the rules are trying to distinguish precisely these things.)

Other 'Solutions':

All of this is cool, but at a higher level than what the basic rules suggest is needed.

 

[Kobold Stew]

I think there is one aspect missing: animated illusions.<snip awesome discussion>

More powerful illusion spells, like Major Image, include other sensory effects so an illusory ghost can include a "chilling" aura, an illusory fire elemental radiates "heat", or a stone statue's footsteps make sound. These should be harder to recognize. Which makes sense, higher level spells should be more effective.

I don't think it should be spell level based per se, but there should be increased benefit from a more powerful spell.

I agree with all of this (and with the point about imitating spells), and each of these aspects makes it harder to suspect the initial illusion, and so more likely to affect how an opponent responds. Thnak you!