The 2024 Player's Handbook Wants the Player to Know All the DCs For Some Reason

I had a chance to look over the new PHB and wanted to record some of my thoughts about what I found there. This isn't meant to be an exhaustive changelog with minor revisions to individual classes, subclasses, races, feats, spells, and the like - others have already done that. It also isn't a "review", per se - others have done that, too. Rather, I wanted to call attention to a few topics in particular that I thought were noteworthy changes - some good, some bad, some...interesting - that I haven't seen reported on elsewhere.

It's worth noting that some of these items might be elaborated on further in the DMG. These are just my impressions based on what's available so far.

I'm going through these items in order as they appear in the book, so there's no real rhyme or reason as to how this post is organized.

I love this piece by Noor Rahman in the rules glossary. Click to embiggen.

Skills

I want to call attention to the fact that skill proficiencies are still, by default, tied to ability scores. That is, the player makes a Charisma (Performance) check. I mention this because this little blurb is still included on page 14:

Skills with Different Abilities

Each skill proficiency is associated with an ability check. For example, the Intimidation skill is associated with Charisma. In some situations, the DM might allow you to apply your skill proficiency to a different ability check. For example, if a character tries to intimidate someone through a show of physical strength, the DM might ask for a Strength (Intimidation) check rather than a Charisma (Intimidation) check. That character would make a Strength check and add their Proficiency Bonus if they have Intimidation proficiency.

I understand why skills are tied to ability scores by default - players and DMs can both benefit from having some idea of what normally goes with what - but I personally really wish they had done away with the default pairings entirely. 

It is enough to know that a character is proficient in Intimidation - the DM gets to decide if a situation calls for a character's proficiency to be added to a Strength check, a Charisma check, or some other ability check entirely, and leaving it open might entice players to come up with creative ways to advocate for the use of their skills in unusual situations. Maybe that is just too "advanced" for the PHB, and I certainly wouldn't want players arguing that they can use Strength to make a Perception check, because that doesn't make any sense (not that that's stopping the 2024 barbarian).

The "Skills with Different Abilities" rule is one of my favorites in 5e, and I wish it was used more, but players see "Persuasion (Charisma)" on their character sheet and assume that NPCs can be persuaded only by the character's innate ability to influence (Charisma), and not with logical or intuitive arguments (Intelligence and Wisdom, respectively).

Social Interaction

This is one of my favorite topics to discuss. The social interaction "pillar" is often criticized for being too threadbare. Does the 2024 PHB improve it?

The 2014 PHB took a rules-light approach to social interaction, explaining that players could use a "descriptive" or "active" approach - i.e., describing what their character does or says while interacting with an NPC versus speaking in-character and saying exactly what the character says. The DM then uses their knowledge of the NPC's personality to adjudicate the results, sometimes calling for an ability check. The DMG, in turn, has information on NPC's attitudes (friendly, indifferent, or hostile), associated social interaction DCs, and how to actually adjudicate the interaction.

The 2024 PHB now provides an example of social interaction, including examples of both descriptive and active roleplaying. The friendly, indifferent, and hostile attitudes return, and are now included in the PHB's rules glossary. Player characters have advantage on ability checks to influence friendly creatures and disadvantage on ability checks to influence hostile creatures. The DM can now call on player characters to take the Influence action to urge an NPC to do something.

I like that the PHB explicitly states "The DM will typically ask you to take the Influence action", which will hopefully avoid players simply yelling "I take the Influence action!" instead of actually roleplaying the interaction. Hopefully. 

The rules for the Influence action also state that requests that align with the NPC's desires automatically succeed, and those which don't automatically fail. An ability check is only needed if the NPC is "hesitant". Here's the kicker, though:

The GM chooses the check, which has a default DC equal to 15 or the monster's Intelligence score, whichever is higher. On a successful check, the monster does as urged. On a failed check, you must wait 24 hours (or a duration set by the DM) before urging it in the same way again.

First, I can see why the DM chooses which ability check the player makes, but I also think it should be framed more as a discussion. The player should be able to say "I say 'blah blah blah'." If the DM then says "Okay, make a Persuasion check", the player should be able to say "Wait, no. I was trying to threaten them" (or whatever). It also, again, might preclude the player from trying to convince an NPC using a logical or intuitive argument (if the DM habitually falls back on the default Charisma check). Like with any other part of the gameplay loop, the player and DM need to be on the same page about what the player is trying to do and how they're trying to do it. 

Maybe I'm reading too much into it, but the codifying of social interaction in this way feels so rigid. It doesn't seem like it's teaching new players how to navigate these situations in a way which ensures that the DM is hearing them and adjudicating their actions appropriately.

Second, what is with those DCs? 15 or the NPC's Intelligence score - no matter the NPC, no matter their personality, no matter the player's approach? The friendly and hostile attitudes already provide advantage or disadvantage on ability checks, so the DCs don't necessarily need to be modified by attitude as they were in 2014, but I'm otherwise a bit baffled by this. 

Less intelligent NPCs are universally easier to Influence, regardless of whether or not the attempt to Influence them has anything to do with how smart they are, and intelligent NPCs are universally harder to Influence, even if a well-reasoned argument might reasonably appeal more to a more intelligent NPC. A more intelligent NPC might be able to see through lies more easily, but if they are in a situation where they are in danger of being harmed or killed, shouldn't an intelligent NPC be able to recognize that, and perhaps be more susceptible to Intimidation?

Shouldn't some NPCs be more susceptible to one approach versus another? A cowardly politician who goes to the theater and regularly engages in intellectual discourse in the king's court would be difficult to deceive, persuade, or impress with a slam poetry performance, but would be easier to intimidate. A king who values emotion over facts and logic would be more susceptible to an impassioned plea than a cold accounting of the facts at hand. Shouldn't player attempts at Deception be opposed by the NPC's Insight? Shouldn't player attempts to Intimidate the NPC be opposed by the NPC's morale?

Admittedly, I'm realizing that my own social interaction procedures use flat DCs based on the NPC's disposition, which don't account for different approaches when attempting to interact with them, but this is as simple as noting the NPC's personality traits, ideals, bonds, flaws, and goals, and modifying the DCs up and down accordingly if necessary. 

The fact that the 2024 PHB's rules don't even allude to these differences between NPCs is what's driving me nuts. It suggests to players that the most optimal form of social interaction is to only ever use whichever skill they have the highest bonus in. That might be the approach most players take anyway, but if players know that different approaches work better or worse with different NPCs, they'll more likely make some attempt to understand what resonates with an NPC before deciding who in the party should speak to them and how. 

As printed in the 2024 PHB, the game is telling the players that all approaches work equally well with a given NPC, so they should just use whichever skill is best for them. It does offer this tidbit:

When interacting with an NPC, pay attention to the DM's portrayal of the NPC's personality. You might be able to learn an NPC's goals and then use that information to influence the NPC.

But this only really applies if the players are able to identify the NPC's goals such that their attempts to Influence the NPC are automatically successful - if they have to roll, it doesn't make any difference.

There's also something to be said about the limitation that a character can only urge an NPC in a given way once per 24 hours. Does that mean a character can only attempt to Influence the NPC with Persuasion once during that time, or will different approaches to persuading the NPC be treated as different urgings? It doesn't read that way. At least the rules are explicit here that the DM can choose a different duration, although I don't know how much good it does.

Hiding

After plenty of complaints of ambiguity about how hiding worked in the 2014 version of the rules, the 2024 PHB offers some clarity and...simplification. From page 368:

With the Hide action, you try to conceal yourself. To do so, you must succeed on a DC 15 Dexterity (Stealth) check while you're Heavily Obscured or behind Three-Quarters Cover or Total Cover, and you must be out of any enemy's line of sight; if you can see a creature, you can discern whether it can see you.

On a successful check, you have the Invisible condition. Make note of your check's total, which is the DC for a creature to find you with a Wisdom (Perception) check.\

The condition ends on you immediately after any of the following occurs: you make a sound louder than a whisper, an enemy finds you, you make an attack roll, or you cast a spell with a Verbal component.

First, this clarifies when a character can hide. The 2014 PHB, by comparison, says only "The DM decides when circumstances are appropriate for hiding." This is probably helpful to some people, although as a DM, I've never had a problem making a common-sense ruling as to when a creature can hide. I have heard that there are some DMs who are super stingy about when they allow characters to attempt to hide, so making the conditions player-facing is probably meant to alleviate that particular pain point.

It is weird that successfully hiding gives the creature the Invisible condition, which is not to be confused with being actually invisible in the reality of the fictional world. The capital I Invisible condition doesn't make the creature lower case i invisible, yet in this case, a creature that uses the Search action has to make a Perception check which beats the total of the other creature's Hide action in order to find them. The searcher does not see the hider just by looking at them, even if they are standing out in the open, so...they are invisible, kind of? 

Does Truesight, which allows a creature to see objects and creatures with the Invisible condition, then negate hiding? It shouldn't, because in the fiction, Truesight is the ability to see through illusions, shapechanging, and into the Ethereal Plane - being hidden by non-magical means is none of those things. It's literally giving me a headache to write this out.

Aside from that, the biggest problem I see with this is similar to the main gripe I had with the Influence action. What is with that flat DC? It used to be that Stealth checks were opposed by the monster or NPC's Passive Perception, which made plenty of sense - it represented the creature's ability to passively notice another creature attempting to hide without actively looking for it. Now, so long as the other conditions for hiding are met (and I'm accounting for special senses like Blindsight and Tremorsense here), a creature has the same chance of successfully hiding regardless of the situation or what/who it's hiding from.

A 1st-level rogue, assuming a starting Dexterity of 16 (+3) and expertise (+4) in Stealth (total of +7), has a 65% chance of hiding from anyone and everything. At 7th-level, once the rogue gets Reliable Talent, which gives them a minimum of 10 on the die when they roll any skill check with proficiency, that same rogue can successfully hide 100% of the time.

That's fantastic for the class fantasy of the rogue, and monster and NPC Passive Perception scores in the 2014 rules were such that this was often the case anyway, but it's also extremely simplistic and limiting. Again, aside from monsters and NPCs with the aforementioned special senses, all monsters and NPCs are equally easy to Hide from if the conditions are met.

It should be easier to Hide from a drunken guard than from a paranoid beholder with eyes more or less literally in the back of its head. The beholder can more easily find a hidden creature, I suppose, but it should also be harder to Hide from it in the first place.

Stabilized Player Characters Versus Unconscious Creatures

I'm not going to go off on this one, but I found it weird. In both the 2014 and 2024 rules, when a player character drops to 0 hit points but isn't killed and is then stabilized, they regain 1 hit point and become conscious after 1d4 hours.

In both the 2014 and 2024 rules, a player could choose, when reducing a creature to 0 hit points with a melee attack, to knock the creature out instead of killing it. Presumably, in the 2014 rules, unconscious monsters and NPCs worked the same as unconscious player characters, regaining 1 hit point after 1d4 hours.

In the 2024 rules, an unconscious monster or NPC is instead reduced to 1 hit point (because at 0 hit points they are dead, unless the DM makes an exception). The monster or NPC then starts a short rest, meaning they always regain consciousness after 1 hour, they get to roll all of their hit dice and heal for that amount, and they get back any abilities that recharge on a short rest.

It has more or less always been true that player characters and monsters and NPCs follow different rules, but in a few instances it really bothers me, and this is another such instance. Not a huge deal, but a weird quirk of the rules I haven't seen mentioned elsewhere.

Languages

I've written about languages before. One of my complaints was that it felt too easy to learn exotic languages, and that they weren't sufficiently differentiated from standard languages. The 2024 rules change languages just a bit. It's weird.

First, in desperation to remove any connection between a character's race/species and their culture, player characters no longer gain languages based on their race. Instead, all characters speak Common and two other standard languages.

The funny thing is, it's not clear how a character learns a rare language (exotic languages are now rare languages). The PHB says: "Some features let a character learn a rare language." For example, druids learn Druidic, rogues learn Thieves' Cant, and rangers learn two languages from their Deft Explorer feature. I don't know that there's any other way to learn these languages.

So, according to the rules, an aasimar can't know Celestial, an elf can't know Sylvan, a drow can't know Undercommon, and a tiefling can't know Abyssal or Infernal. A reasonable DM can obviously make an exception, but I do find it funny.

A piece by Chris Seaman shows how adventurers' equipment evolves with the tiers of play.

Starting Equipment at Higher Levels

Finally, a change I kind of sort of like. The 2024 PHB gives some recommendations for what kind of starting equipment newly made higher-level characters should have - some extra gold and magic items.

If I have a concern about this, it's that if a character with a bunch of magic items and treasure dies, the rest of the party can take their stuff, then a new character enters the game with more magic items and treasure. The more characters die, the more magic items and treasure are added to the party. It doesn't make sense to me that if a character with 20,000 gold and 10 magic items dies, the party salvages their stuff, and then another character with 20,000 gold and 10 magic items joins the party. The party as a whole is now 20,000 gold and 10 magic items richer because they messed up and someone died. 

The way I've handled this in my games is to treat magic items and treasure as belonging to the player, not the character. If a character dies with a bunch of magic items and treasure, that player's new character enters the game with equivalent magic items and treasure. If, however, the other characters loot the dead character and either use or sell their stuff, that stuff no longer belongs to the player whose character died, and their new character enters with fewer items and treasure.

It's a bit meta and not entirely fair to the player with the dead character, but they should have thought about that before they died. And they should convince the other players to give them some of their stuff back so that their new character isn't underpowered (making sure the new character isn't underpowered is in the other players' best interest as well).

Backgrounds

As far as I can tell, the 2024 PHB no longer even suggests that backgrounds are customizable, aside from this:

Each background includes a brief narrative of what your character's past might have been like. Alter the details of this narrative however you like.

The 2014 PHB referred to its backgrounds as "sample backgrounds", and went into detail as to how to customize them or work with the DM to create an entirely new one:

You might want to tweak some of the features of a background so it better fits your character or the campaign setting. To customize a background, you can replace one feature with any other one, choose any two skills, and choose a total of two tool proficiencies or languages from the sample backgrounds. You can either use the equipment package from your background or spend coin on gear as described in chapter 5. (If you spend coin, you can't also take the equipment package suggested for your class.) Finally, choose two personality traits, one ideal, one bond, and one flaw.

If you can't find a feature that matches your desired background, work with your DM to create one.

In my experience, most players do not even try to customize their background, but it happened occasionally when we were playing 5e, and I personally tried to do it with all of my characters. Players new to D&D who start with the 2024 PHB might not even think to ask their DM about the possibility, which is a real shame.

Ability score increases are now tied to backgrounds, which is fine. I personally like my dwarves sturdy and taciturn and my elves lithe and frail, but I get why the culture has moved away from that. 

What's funny is that, beginning a few years ago, Wizards of the Coast did away with specified ability score increases entirely in favor of letting players build the exact characters they wanted, then tied the ability score increases to backgrounds in the 2024 playtest materials, and now...made them specific again? For example, a player who chooses the acolyte background can now only increase some combination of Intelligence, Wisdom, or Charisma (+2 to one and +1 to another, or +1 to all three).

In the 2014 rules, a player might choose to be a dwarf fighter because that race and class synergize well, then have their pick of interesting backgrounds with which to differentiate that character. Now, a player might choose to be a soldier fighter for the same reason, and it is instead the species which differentiates the character.

It's not necessarily better or worse, it's just...different. The race/species and background have just changed places. (Considering the fantasy bioessentialism of it all, maybe it is slightly better now. I just never particularly cared about that but I understand why some people do.)

Perhaps worst of all is that the 2024 PHB does away with traits, bonds, ideals, flaws, and background features (the latter are more or less replaced by origin feats, which are more mechanically useful but less interesting to me). 

I personally loved the TBIFs and used those suggested by each background to create my own for each character. Instead, the character creation chapter suggests one-word descriptors based on the character's ability scores and alignment, which is...fine. It's just kind of a little bit...boring? Uninspiring?

The background features were never terribly exciting - they mostly amounted to some way for the character to obtain free room and board in a settlement - but I felt that they were interesting little roleplaying tidbits. They invited the player and DM to collaborate on something together - the acolyte's home temple or shrine, the criminal's black market contact, or the entertainer's preferred tavern to perform at. It felt good to be the street urchin who could show the rest of the party a faster way through a busy urban neighborhood, or to be the sage who knew how to use the fantasy Dewey Decimal System.

Tools and Adventuring Gear

Every tool and bit of adventuring gear now has a little description about what a character can use it for. Tools list what items a character can craft with them. A character can use perfume to get advantage on Persuasion checks to Influence Indifferent Humanoids. Cute! But also very specific for some reason.

A character can use alchemist's supplies to identify a substance! Cartographer's tools to make a map! Jeweler's tools to appraise a gem! 

What is with these DCs?!

DC 15 to identify any substance? DC 15 to make a map of any area? DC 15 to appraise any gem? It goes on and on like this! A DC 20 check with a gaming set to "win the game". What game? Who is the character playing against? DC 20 to win any game against anyone?

Admittedly, this isn't the first time these DCs have appeared in a book, at least not with regards to tools. DCs to do a variety of things with tools were first printed in Xanathar's Guide to Everything. But now, they've infected the entire equipment list: DC 10 to apply makeup with a disguise kit (what?), DC 13 to wrap a chain around any creature, DC 10 to tie a knot with a rope. It's enough to make me go completely mad.

Of course the DM is well within their rights to change any of this nonsense. Of course. But the DCs are in the PHB! Why? That's DM-facing information - players literally do not need to know this information, and that goes for social interaction, and hiding, and equipment, and everything else. The section on making ability checks should tell the player how to make an ability check, then say "The DM determines the DC." That's all the player needs to know. The book is actively undermining the DM's ability to adjudicate these situations by implying that there are set-in-stone DCs for any of these things.

All of the tools also have an associated ability score, which mostly seem arbitrary. Painter's supplies require Wisdom, as do all gaming sets (dragonchess doesn't use Intelligence, and playing cards don't use Charisma, so no bluffing). All musical instruments use Charisma (dexterous fingers have nothing to do with playing a string instrument - it's all about convincing people that the performance is good through force of personality).

I've seen a few people praise the new equipment section, so I was especially startled when I read and saw just how bad it actually is.

Spellcasting Services

This is nice - there are suggested prices for NPCs to cast spells for the party. Of course, I'm once again wondering why this is in the PHB, and if it must be in the PHB, why there isn't some disclaimer that the DM can change it or decide that certain services aren't available. Still, it's nice enough, and the prices seem reasonable.

I also find it interesting that the table aligns pretty closely with my own ideas about what services are available in which types of settlements. I've always run it so that Tier 1 NPCs can be found in villages, Tier 2 NPCs in towns, Tier 3 NPCs in cities, and Tier 4 NPCs in metropolises, which is exactly what the table suggests (minus the metropolises).

Where Are the Downtime Rules?

I don't know. They're not in the 2024 PHB (crafting is, but none of the other options). 

The new DMG includes rules for player character bastions, so maybe they're lumped together with downtime rules there. Or maybe they've gotten rid of downtime altogether. I wouldn't be surprised.

That's all I've got for now. I've spent a little bit of time looking at how they changed the classes and spells and whatnot but honestly my eyes just kind of glaze over whenever I try too hard to read that stuff.

I'll probably do another post like this when I can get a look at the DMG. Hopefully it will give me less of a headache and I'll have more positive things to say, but I've already heard some things that make me think I, uh...might not like it. Lol.

An Update on the Adventuring Day

Apparently, the 2024 edition of the 5e DMG will not include the concept of "the adventuring day".

by Matt Ray

This isn't really much of a surprise to me. As my previous post detailed, the adventuring day has always been a controversial (and misunderstood) concept in 5e. People got hung up on the "6 to 8 medium to hard encounters" of it all, and ended up missing the forest for the trees - that is, that the adventuring day concept was simply demonstrating that player characters are intended to face a certain degree of challenge (measured in XP) between long rests.

What made the adventuring day helpful is that it attempted to quantify the attrition that player characters are expected to experience between long rests, when all of their hit points and abilities reset. The important thing wasn't the exact number of encounters or their difficulty, but simply that player characters should have at least three encounters (allowing for a minimum of two short rests) between long rests, and that these encounters gradually wear down the characters' resources. Attrition is the primary means by which the characters are challenged - not so much each individual encounter.

This was especially important in 5e, because characters recover resources (and especially hit points) much faster than in previous editions, where attrition could be drawn out over a longer period of time.

I felt that the adventuring day concept was - at the very least - helpful for planning dungeons. If I wanted a short dungeon, I could plan for it to have an adventuring day's worth of encounters. If I wanted a three-level dungeon, each level could have an adventuring day's worth of encounters. And in a dungeon environment, it has always made sense for there to be that number of encounters. The adventuring day worked perfectly well for me in dungeon-crawling scenarios.

The problem, as I see it, is that inexperienced 5e DMs don't run dungeons - at least not the kind that the adventuring day seems to encourage. The 5e DMG barely teaches dungeon design, and when it does, the guidelines are not great. (For example, according to page 296, 50% of dungeon rooms contain monsters - why yes, that does sound like an un-fun slog! Compare that to AD&D, where only 25% of rooms contain monsters.) 

Instead of running dungeons, most 5e DMs seem to Google "D&D battle map" until they find one that's like, a shattered causeway being held up by a petrified giant's hand while a volcano is erupting or something (which is cool I guess - I've never understood how people use these hyper-specific battle maps unless they are designing the whole adventure around them, but more power to them), then plop down some tokens on the map and voila - find a way to get the player characters to the big damn fight, and that's the session. At most, they are running Five Room Dungeons (which I have my own problems with).

Anyway, I thought it was funny that I just wrote a post about how the adventuring day is an unfairly-maligned concept which is actually kind of useful, then find out that, like many such things throughout D&D's history of editions, it is getting dropped in the 2024 rulebooks. What are we getting instead?

According to Christian Hoffer:

The new 2024 Dungeon Master's Guide contains a streamlined guide to combat encounter planning, with a simplified set of instructions on how to build an appropriate encounter for any set of characters. The new rules are pretty basic - the DM determines an XP budget based on the difficulty level they're aiming for (with choices of low, moderate, or high, which is a change from the 2014 Dungeon Master's Guide) and the level of the characters in a party. They then spend that budget on creatures to actually craft the encounter. 

But we already had that. XP budgets for encounters of different difficulty levels were already in the 2014 DMG, it just also told DMs how to string those encounters together in a combination that would challenge the player characters. Instead, the 2024 DMG includes advice about "encounter pace and how to balance player desire to take frequent Short Rests with ratcheting up tension within the adventure." 

Sure. I'm sure those guidelines will be very...thorough.

This certainly isn't the end of the world. After all, the adventuring day is a very new concept - it's not some sacred cow of original D&D. In fact, I'm sure most gorgnards and OSR people alike detest the very concept. 

But honestly, one of the things I really like about 5e is that the designers explicitly say "X characters of Y level can handle A monsters of B CR, and they can do that C number of times per day before they need to take a break." It's clinical and math-y, but I also find it very helpful in broadly gauging what my players' characters can handle. I miss it when I play AD&D because it takes a lot of the guesswork out of prep. 

Sure, it's only slightly more useful than "X characters of Y level can handle A monsters of B CR - you decide how many of those encounters you and your players want!" But, I don't see why as a game designer one would design an "improved" version of the game with...slightly less useful guidelines.

We'll have to wait until the new DMG is more widely available to see what these new guidelines actually look like. I am trying to reserve judgment, but I do find it funny how the "refined" rules for 5e, with 10 years of hindsight, seem to be eliminating rules and advice from the original rulebooks, replacing them instead with flat DCs for hiding and tying knots. I look forward to perusing the new DMG to see just how much they've gotten rid of. (For free, of course - there's no way I'm paying for these books.)

On the Adventuring Day

As far as I can tell (usual disclaimer: I am not very familiar with 3rd or 4th edition D&D, at least not from a DM's perspective), the term "adventuring day", as defined by D&D, originated with 5th edition. From the 5e DMG (page 84):

Assuming typical adventuring conditions and average luck, most adventuring parties can handle about six to eight medium or hard encounters in a day. If the adventure has more easy encounters, the adventurers can get through more. If it has more deadly encounters, they can handle fewer.

In the same way you figure out the difficulty of an encounter, you can use the XP values of monsters and other opponents in an adventure as a guideline for how far the party is likely to progress.

For each character in the party, use the Adventuring Day XP table to estimate how much XP that character is expected to earn in a day. Add together the values of all party members to get a total for the party's adventuring day. This provides a rough estimate of the adjusted XP value for encounters the party can handle before the characters will need to take a long rest.

Since XP is tied to Challenge Rating, and CR is (allegedly) tied to how difficult a monster or other challenge is to overcome, XP is used as a measure of how many and what degree of challenges a given group of player characters can overcome before needing to take a long rest - this collection of challenges per long rest is then called an "adventuring day". (Note that the DM is intended to award XP to the players for overcoming challenges of all kinds, and not just monsters - thus, an entire "adventuring day" could also be spent exploring a dungeon or talking to NPCs, so long as doing so is challenging and requires expending resources.) 

Because the adventuring day is simply a measure of how much challenge a group of characters can overcome before needing to take a long rest, an adventuring "day" needn't be a day at all. If a DM instead prefers to use the "gritty realism" resting variant detailed in the DMG (page 267), an adventuring "day" is instead any period interrupted by a week's worth of rest - it could be a day, a week, a month, a very long year, or even a merciless lifetime. Even if the DM uses the default rule - in which a long rest is 8 hours (6 hours of sleep and 2 hours of light activity) - an adventuring "day" can be much longer than a single day, if for some reason the characters are prevented from resting. Perhaps an antagonistic illusionist is haunting their dreams.

The point of all this is to illustrate that D&D 5e is a game that relies on attrition to create true challenges. Well-rested player characters have a lot of hit points and spells, and other abilities that they can use to soften, overcome, or utterly trivialize challenges - even "deadly" ones. However, they have a finite amount of these resources that they can use each "day". Challenges become harder as the "day" goes on, because fewer resources are available. Eventually, resource-depleted characters will encounter something truly challenging - life-threatening, even. 

This is how it's supposed to work, anyway - the player characters push their luck as their resources dwindle, and it's a matter of player skill to determine when would be the wisest point to stop and recuperate. But if all of a character's resources are regained on a rest, and characters are more powerful when well-rested, isn't it wisest and most prudent to simply rest after every resource expenditure? This results in what is frequently referred to as "the 15-minute adventuring day". 

Let me clarify exactly what I mean by that. I have seen some people get hung up on the "15-minute" part - that is, they get upset that, because a combat round lasts just six seconds, and most combats are over in just a few rounds, that all of the action of the adventuring day takes place in about 15 minutes. They are interpreting "the 15-minute adventuring day" literally. I don't care about that. It makes plenty of sense to me that 15 minutes of battling magical monsters, dodging life-threatening traps, and negotiating with orcish warbands is enough for someone to want to call it a day. Adrenaline is exhausting once it wears off.

What I'm referring to (and what people usually mean when they use the phrase "15-minute adventuring day") is the tendency of players to liberally abuse D&D's resting mechanics to ensure that they have the optimal resources for every encounter - that is, the characters use all of their most powerful abilities in their very first encounter of the day, then rest as soon as possible so that they'll have those abilities back for the next encounter. Because D&D's design is predicated on the assumption that attrition is taking place, this behavior often renders the game toothless - DMs complain that player characters are too powerful, or that monsters are too weak. Maybe that's true. Or maybe we just need to stop player characters from resting so much.

Was It Always This Way?

Before I get into it, I want to take a detour to talk about older editions of D&D. Do players of OD&D, B/X, and AD&D run into problems with the 15-minute adventuring day? Maybe. One thing that has never changed is that player characters get all of their spells back when they take a long rest. So even way back in OD&D, magic-users had good reason to beg for a break after they'd burned their one spell for the day and had to resort to feebly wielding their dagger.

Throughout the editions, player characters have accumulated abilities aside from spells which recharge on a daily basis. In AD&D, paladins can lay on hands once per day, and druids can shapeshift into a either a bird, mammal, or reptile once per day each. 4e codified "daily" powers alongside "at will", "encounter", and "utility" powers. This has been further complicated in 5e by the introduction of abilities that can be used a number of times per day equal to a character's proficiency bonus, or ability score modifier, or an even more arbitrary value.

The main difference between editions when it comes to this issue is the speed of natural healing. The amount of hit points recovered after a rest has ballooned over time, and the limitations on what constitutes a rest have grown much looser. In OD&D, B/X, and AD&D 1e, characters require complete rest (i.e., no adventuring activities whatsoever) to heal naturally. In 2e, characters can heal naturally while doing light adventuring activities (i.e. traveling from place to place), but heal more quickly when resting for a full day, and never heal naturally on a day when they engage in combat.

Natural healing begins to take on super heroic qualities in 3e, in which characters can finally regain hit points at the end of a strenuous adventuring day - and not just one hit point, but a number of hit points equal to the character's level. 

The super heroic trend continued with 4e, in which characters could use healing surges (equal to a quarter of their total hit points) to heal themselves at any time between combats (and sometimes mid-combat). Characters had different numbers of healing surges determined by their class, Constitution score, and other factors, with rangers, rogues, warlocks, and wizards having the smallest baseline number of surges (6) and paladins having the largest number (10). At the end of the day, an extended rest recharged all healing surges and hit points, meaning even the frailest characters had not just their usual hit points, but an additional pool of hit points equal to 150% of their maximum in nearly on-demand healing every day. The distinction between a full day of rest and a night's sleep at the end of a strenuous adventuring day - which had persisted all the way up to 3e - is lost to time at this point.

5e rolled back the extremes of 4e somewhat, replacing healing surges with hit dice. A character has hit dice equal to their level (rather than a number of surges determined by their class), the size of the dice is determined by the character's class levels (rather than being a 25% heal), and the character's Constitution modifier is added to each die. A short (1-hour) rest is required to use the hit dice (rather than being an at-will heal outside of combat). Essentially, the character has an additional pool of hit points roughly equal to 100% of their maximum (more or less, depending on how they roll) which they can use to heal in the middle of the day. Characters regain only half their expended hit dice on a long rest (compared to all expended healing surges in 4e). The healing of all lost hit points on completion of an extended/long rest survives from 4e.

The point is that the earliest editions of D&D use a different timescale of attrition. On day one of the adventure, the fighter may take only 5 points of damage, but when they rest that day, they regain no hit points. On day two, they take another 5 points of damage, and again cannot regain hit points from resting. There may be a cleric in the party aiding them with magical healing, but the point is that the damage compounds over time - no natural healing is possible so long as the fighter continues to exert their self each day. In 3e, the fighter can heal at the end of a strenuous day, but usually not completely. In 4e and 5e, the damage the fighter took on day one of the adventure no longer matters on day two, assuming that they were able to successfully rest - everything resets (except 5e's hit dice, which only reset halfway).

In other ways, the earliest editions of D&D don't rely on attrition to create challenges at all: the stench of the zombie lord kills the characters if they fail a saving throw; the touch of the wraith drains a level of experience from the affected character; rocks fall, everyone dies - no whittling away of hit points, spell slots, or per-day abilities necessary in any case.

I'm not suggesting that 5e needs slower natural healing, or to distinguish between resting at the end of an adventuring day and taking a week or more to rest completely between adventures. Instead, I hope to have demonstrated that "the adventure" (i.e., the days or weeks at a time during which the characters are adventuring, where little to no recovery is possible) has gradually transformed into "the adventuring day" (i.e., the single, brief unit of time during which the characters are adventuring, after which they recover rapidly and almost completely) - and how the adventuring day is more important to the game's assumptions of balance than ever before.

How Long is an Adventuring Day?

Well, a day is 24 hours. A long rest is 8 hours, so that leaves 16 hours of adventuring. Characters can travel 8 hours per day before needing to check for exhaustion, which leaves 8 hours for poking around dungeons or whatever. Solved.

This is wrong. For the purposes of this discussion, it doesn't really matter how long a day is. What matters is how challenging a day is. The quote from the DMG at the beginning of this post says that a "typical" adventuring day consists of "about six to eight medium or hard encounters". People get hung up on this. The more important part is this:

For each character in the party, use the Adventuring Day XP table to estimate how much XP that character is expected to earn in a day. Add together the values of all party members to get a total for the party's adventuring day. This provides a rough estimate of the adjusted XP value for encounters the party can handle before the characters will need to take a long rest.

The important thing is to determine how much challenge the party can handle before needing to take a long rest, and the challenge isn't measured in a number of encounters of X or Y difficulty, but in XP. Every encounter has an adjusted XP value determined by the number of monsters and the monsters' CR. Add up the adjusted XP value of each encounter/challenge until the party's daily XP budget is reached - that's an adventuring day. (The DM is also supposed to award XP for non-combat challenges, but assigning an XP value to these is more of a gut feeling with loose guidelines than a mathematical equation.)

Five 1st-level characters can handle 1,500 XP of challenges per day. That could be 6 medium encounters or 4 hard encounters, sure, but it could also be 12 easy encounters or 3 deadly encounters. It could be 1 easy, 2 medium, 1 hard, and 1 deadly. It could be a single encounter which is beyond "deadly" but still within the 1,500 XP budget. A common problem I see is 5e DMs planning for one "big damn encounter" per day, but still adhering to the easy/medium/hard/deadly encounter categories. Those categories are still assuming there will be multiple encounters per day! If the DM really wants to run one big damn encounter per day, they need to make that encounter much more robust, with an adjusted XP value nearly equal to the party's daily budget.

Technically, that is all true. But is it for the best? Another thing worth considering is that 5e introduces short rests. Every character gets all of their hit points and daily abilities back on a long rest, so every character is incentivized to long rest eventually. Likewise, every character can spend hit dice to heal during a short rest, but many characters benefit from short rests in ways which are unique to their class, and almost every class benefits differently, so every character is incentivized to short rest to a different extent.

On the one hand, bards, clerics, and druids, for example, regain some secondary features on a short rest: Bardic Inspiration, Channel Divinity, and Wild Shape (the primary feature of all of these classes is spellcasting, which is tied to long rests). On the other hand, fighters, monks, and warlocks regain their primary features every time they take a short rest. It's part of the balancing act between classes. Characters that recover their features on a short rest maintain their potency throughout the day - the longer the day (i.e., the more short rests they can take), the better they match up to the long rest-dependent classes. If the day is too short (especially if it consists of just one encounter), they pale in comparison, and any attempt at balance between the classes falls apart.

The 5e DMG says:

In general, over the course of a full adventuring day, the party will likely need to take two short rests, about one-third and two-thirds of the way through the day.

So if, for example, the adventuring day consists of 6 medium encounters, the party is expected to have two such encounters per short rest. Frankly, if I'm playing a fighter, monk, or warlock, I'm proposing a short rest after every encounter, so long as there isn't some time pressure. Even if I'm playing another class, so long as I get something back on a short rest, I'm taking one. It's fun to be a bard and have all of my uses of Bardic Inspiration to throw around every single encounter.

Let me just say now that I don't really do all this encounter planning and math when I'm prepping a D&D session. If I'm creating a dungeon, I'll stock it with a mixture of different types of monsters and encounter difficulties - there will be easy, medium, hard, and deadly encounters designed for a group of characters of the level I think is appropriate for the dungeon, trivial encounters designed for lower-level characters, impossible encounters designed for higher-level characters, indifferent, unfriendly, friendly, and helpful monsters in addition to hostile ones, traps of varying deadliness, and magical benefits of varying potency. I'm not making spreadsheets with XP values to plan my dungeons. It's not that deep.

Sometimes, the goal of the adventure at hand is to travel across the wilderness, find the tower where the harpies are roosting, and kill them. The characters might have a random encounter in the wilderness on the way there, but most likely, they're going to arrive with all their resources at hand. The tower won't be a complex mythic underworld with a variety of monsters, tricks, traps, treasure, and the like. The adventure will most likely big a big damn fight with a bunch of harpies, and that's okay too, so long as it's only sometimes.

The point is that characters are intended to face a certain amount of challenge (measured in XP) per day. Because some classes rely more than others on short rests to keep pace, it is better to spread the day's challenges out across multiple encounters. That being said, variety is the spice of life, and a mixture of different numbers and types of encounters per day is vital to an exciting game.

No Rest for the Wicked

The most common solution to adventuring day woes is to limit long rests in some way, forcing the characters to deal with attrition. Usually, this is done in one of two ways: either resting has to take place in a location that is some combination of safe and/or comfortable, or the "gritty realism" optional rule is used, extending the length of a short rest to 8 hours and a long rest to 7 days.

From what I can tell, Basic D&D is the only version that explicitly required resting in a particular location in order to recover resources:

Each day of rest and recuperation back "home" will regenerate 1 to 3 of his hit points for the next adventure.

One might imagine that past editions' "bed rest" rules require, well...resting in a bed, but they don't. From 2e:

If a character has complete bed-rest (doing nothing for an entire day), he can regain 3 hit points for the day. For each complete week of bed rest, the character can add any Constitution hit point bonus he might have to the base of 21 points (3 points per day) he regained during that week.

3e also defines "bed rest" as "doing nothing for an entire day". No literal bed required. Thus, I don't really think there's much precedent for forcing characters to sleep in a "safe" location (aside from the Basic D&D rule). It seems to me like a somewhat arbitrary limitation imposed by the DM to prevent resting in the dungeon. There are other ways to do that, though, which I'll get to later.

As for "gritty realism", I most often see this solution championed by true adventuring day believers who insist that it's "unrealistic" to fit that many encounters into a single day. How does one possibly solve that problem? Where might a party of adventurers face 6 to 8 encounters worth of monsters in a single day? Perhaps in some ancient complex of densely interconnected, labyrinthine rooms, where a vast array of monsters, traps, and tricks can lurk around any corner - a mythic underworld of sorts where the common sense laws of the natural world no longer apply? No, no, that simply doesn't make sense. That's not what they do on the actual plays I watch. Those people are actually playing the game, after all! How could they be doing it wrong?

Something I find interesting about "gritty realism" is this, from the DMG (page 267):

This puts the brakes on the campaign, requiring the players to carefully judge the benefits and drawbacks of combat. Characters can't afford to engage in too many battles in a row, and all adventuring requires careful planning.

This approach encourages the characters to spend time out of the dungeon. It's a good option for campaigns that emphasize intrigue, politics, and interactions among other PCs, and in which combat is rare or something to be avoided rather than rushed into.

This is not really an angle I had considered. I only ever see people discuss this optional rule as a way to fit an adventuring "day" worth of encounters more easily between long rests, not as a means to make players more carefully pick their battles, or to encourage more social interaction gameplay. It makes sense - if a combat lasts a minute but takes a week to recover from, the players may think twice before engaging, and if they're spending a week at a time in town, they may prioritize building relationships with NPCs in their home settlement.

Truthfully, aside from the atrocious name, there is a convincing case to be made for adopting the "gritty realism" rule variant. It does seem to fix a pacing problem that many people experience. I have a few quibbles with the rule and how it interacts with spell durations, preparing spells, exhaustion, and recharging magic items, but my greatest concern is that it breaks down in dungeon play. 

Dungeons are still, for me, the main focus of play in 5e, and they're exactly the place where the adventuring day makes sense. It is crystal clear to me that 90% of 5e's design was done with dungeon play in mind, even if the rules do a poor job explaining that or instructing DMs in how to run dungeons. "Gritty realism" makes sense if the adventure involves a few encounters on the way to the dungeon, a few encounters in the dungeon, and a few encounters on the way back. It doesn't make sense if the dungeon is a densely-packed collection of challenges designed to test the players all on its own.

Personally, outside of dungeon play, I don't really care all that much if combat is "balanced" or challenging. The challenge of traveling from place to place isn't fighting (at least, not entirely). The journey from place to place is interesting because the characters can find unexpected things, get mixed up in unforeseen side quests and shenanigans, make new friends and enemies, and the like. 

When they do end up fighting while on a journey, those encounters should be swingy. The journey to the tomb shouldn't be carefully preplanned to provide the appropriate amount of challenge for X characters of Y level. It's the wilderness - who knows what's out there? 1st-level characters can get wrecked by an ancient dragon if they're unlucky, and 20th-level characters can pick a fight with five goblins if they feel like it. The wilderness is wild. It has always been more dangerous than the dungeon.

The dungeon traditionally has levels, which correspond to the characters' levels. The dungeon is balanced - it's the site of the adventuring day. It's where the game actually works - unless the players decide to abuse rests, and thus break it. So without changing anything else about how resting works, how does one prevent this?

"YOU CAN NOT HAVE A MEANINGFUL CAMPAIGN IF STRICT TIME RECORDS ARE NOT KEPT" (AD&D 1e DMG, page 37)

Gygax comes off unnecessarily strong in the above quote, but he's not entirely wrong. I would revise the sentiment as such: "You can have a meaningful campaign without keeping strict time records, but it helps a lot to keep at least some records and solves a lot of the game's problems." Not pithy or prescriptive enough for the grognards I guess.

Some people advocate for putting every adventure on a time limit. There is an evil spellcaster in the tomb doing an evil ritual! The characters have to go through the dungeon and stop them before the ritual is complete! This happens on every adventure for some reason! This certainly works sometimes, but it's hard enough already to come up with quest hooks to propel a campaign forward. It's doubly hard to create some artificial time limit for every adventure. It also just isn't very fun for the players to never be able to slow down and simply live in the fictional world for a while.

Instead, I use wandering dungeon monsters to exert time pressure. In my games, I use a loose interpretation of AD&D's 10-minute dungeon turn concept. Here's my basic procedure for running a dungeon crawl:

1. The characters enter the room. I describe it.
2. Players ask clarifying questions and I answer with just enough detail to either spur the players to action or elicit more questions.
3. At least one player says they want to do something.
4. I ask every other player what their character is doing (even if it's "Just standing around" or "Waiting for so and so to do their thing").
5. I give each player new information based on what they're doing, call for ability checks, etc.
6. Repeat.

A 10-minute dungeon turn is when all of the characters are doing their chosen thing in that room. If the fighter is muscling open a sarcophagus, the wizard is studying a tome, the cleric is examining an altar, and the rogue is checking for traps, that all takes the same 10 minutes. Every turn, I make a wandering monster check, with a 10% chance of an encounter. On average, the characters can do ten things before they encounter a wandering monster.

This accomplishes two things: It puts time pressure on the players without needing to make every adventure a "stop the ritual" quest, and it also obviates the need to put arbitrary limits on resting in the dungeon. A short rest is 1 hour, which is six wandering monster checks. The characters have a decent chance of squeezing that rest in without having an encounter (I am considering changing short rests to 10 minutes, just to really juice those fighters, monks, and warlocks a bit). A long rest is 8 hours (and because of the one-long-rest-per-24-hours rule, the characters may need to wait around for a bit before they can actually start the rest). That is, at minimum, 48 wandering monster checks. Long resting inside the dungeon is an exercise in frustration, at best.

By comparison, in the wilderness, encounter checks are made every four hours, so resting outside the dungeon is a much safer bet. Okay, so the characters just leave the dungeon to rest after every fight. The problem is still there.

I fixed this by checking for wandering monsters every time the characters reenter a room they've already been through. If the characters have to backtrack through ten rooms to leave the dungeon, they will probably have an encounter on their way out. Then, they will probably have another on their way back in. I justify this by telling myself that monsters are more likely to be patrolling areas where the party has already caused a ruckus or are known to be using for ingress and egress. It also rewards the players for finding additional entrances and exits to/from the dungeon which allow them to go in and out without passing through as many rooms.

I've gone back and forth on whether or not wandering monsters should be drawn from the dungeon's key or should materialize out of thin air. I am inclined towards the latter because if wandering monsters are taken from the key, it actually becomes beneficial to run into them. If there are 20 goblins in one room, and the party randomly encounters half of them wandering the dungeon, they've now split the single difficult encounter into two easier ones purely by chance, due to a mechanic that is intended to be a disincentive.

I like wandering monsters that materialize from the dungeon itself because I like a mythic, otherworldly dungeon. I like to fill my dungeons with supernatural creatures like elementals, fiends, oozes, and undead - monsters that be conjured, seep through cracks in the floor, or manifest from the shadows themselves. I don't really have to justify it because the characters haven't explored the whole dungeon yet - the players don't know what's all in it, so they have no way of knowing where the monsters are coming from. If a specific type of monster has a lair in the dungeon, I'll cross them off the wandering monster table once they've been defeated in their lair, allowing player characters to eliminate threats as they progress.

Okay, so the characters fought through ten rooms in the dungeon. Let's say two rooms had preplanned encounters (medium and deadly), and the characters encountered a wandering monster while exploring (hard). They decide to rest, and leave the dungeon the way they came, having another encounter (medium) on their way out. Assuming they are our aforementioned 5 1st-level characters with a daily budget of 1,500 XP, that gets us to 1,375 XP (250 x 2 = 500, + 375 + 500 = 1,375). Not 1,500, but very close - some delves will be easier and some will be harder (and remember, not all of the encounters will be with hostile monsters, and characters have abilities which allow them to avoid combat with hostile monsters, so players can still pick and choose their fights, and often expend resources when doing so).

While They're Sleeping

The characters have made it out of the dungeon and get to rest. They might have two to four wilderness encounter checks while making camp and resting - depending on the terrain, they will have fewer or more checks with a lesser or greater chance of having an encounter. They go back into the dungeon the way they went the previous day and maybe have one wandering encounter. So what? They cleared out a bunch of rooms the day before. The dungeon is easier now.

But it isn't, necessarily, because the dungeon and its denizens do things while the characters are away. There's plenty of commentary and differing philosophies about restocking dungeons. Generally, it's most common in old school play, where dungeons are tentpole megaplexes intended to support entire campaigns on their own. However, I've found that restocking works plenty well in 5e and with smaller dungeons. 

Now, restocking makes a lot more sense in editions of D&D where natural healing occurs much more slowly. If the characters spend a week recovering in town, it's understandable that the dungeon would change - if they spend just a day camping directly outside, not so much. This is another decent argument for the "gritty realism" rest variant, but again, I like my dungeons to be filled with supernatural creatures that can magically materialize at a moment's notice, so the sudden repopulation of monsters doesn't bother me.

I already linked some pretty well-considered ways of restocking, but I like to keep it simple - basically, every room the characters have "cleared" has a 10% chance of being restocked, usually with monsters, but sometimes with traps (it is almost always something bad, because the point is to introduce some new challenge into the cleared space, unless the characters have allowed some friendly faction to move in). The new monsters are drawn from the wandering monster table, excluding those monsters whose lair within the dungeon has already been cleared.

Using the aforementioned example of ten cleared rooms, on average, one would be repopulated. The characters are still making progress through the dungeon, but now on their way back in, they are probably dealing with one wandering monster and one restocked room. Then, they still have to deal with the rest of the dungeon, which they haven't touched yet.

There are also more common sense ways of restocking - if the remaining monsters are aware of the characters' initial foray, they may move around to different areas, set traps, make common cause with other monsters, and the like. I like to be surprised as a DM, so I randomize all of this and try to create a fictional justification for it after the fact.

I personally prefer not to award XP for defeating wandering monsters and restocked monsters, because both are intended to be avoided. Between wandering monsters, restocked rooms, the time-consuming nature of combat in D&D, and little to no XP being gained, this probably sounds like unnecessary cruelty and a big hassle for the players, and in a way, it is. 

By encountering these monsters, the players have already failed at the challenge, which was to use their time efficiently and avoid doing extra work. If the players are getting unlucky on the die rolls and end up dealing with a bunch of extra encounters, I might throw them a bone, but the bulk of XP is for clearing the dungeon, not clearing it again because they didn't go far enough the first time. I'm always very clear with the players: "This is when I'm checking for wandering monsters" and "If you want to leave, you might want to try to find a shortcut out first" and "If you leave now and rest, there might be more monsters when you come back".

The trick is this: if I want to (and I don't always want to), I can decide how many "days" I want the characters to have to spend in the dungeon before "clearing" it. If it's meant to be a small dungeon, it will probably have an adventuring day's worth of challenges. If it's a three-level dungeon, it will probably have three adventuring days' worth of challenges (one per level). 

The point is for the characters to push themselves, and for the players to feel motivated to get a full "day" in. After they "clear" a level, maybe that level is safe enough to rest on, and won't be restocked. I'll say "Hey, you cleared this area. It's safe if you want to rest before pushing on." The players are rewarded for actually pushing themselves to complete a full day of adventuring, and for clearing the dungeon efficiently. If they dilly dally too long or backtrack for breaks too often, they're only hurting themselves by having unnecessary encounters.

Again, I don't always abide by this by-the-numbers, balancing act approach to adventure design, but I do think it's helpful to understand the concept of challenge by attrition and the unit of measure that is the adventuring day. A little time pressure goes a long way towards pushing the players to play the game the way it was designed, and it turns out that when they play it that way, it works pretty well.

On Illusions

I recently ran a handful of sessions in a 5e game in which the player characters had earned the enmity of an arch-illusionist (more on that later). An illusionist is like most good Batman villains - not a particularly powerful adversary in combat (Batman will usually beat the piss out of the Joker, the Riddler, or the Mad Hatter once he gets his hands on them), but extremely frustrating up until the moment they can finally be confronted. I wanted my antagonist illusionist to feel like that. 

With that in mind, I started familiarizing myself with the school of illusion, thinking of how best to make this scenario as difficult for the player characters as possible. It was time for them to experience the terrifying power of light and sound.

Illusions are something of a prickly topic in D&D. Players whose characters rely on illusions require a DM who is willing and able to roleplay their monsters and NPCs with fidelity to the in-character/out-of-character knowledge divide. DMs whose monsters and NPCs rely on illusions require players who are willing and able to do the same. 

There are a variety of spells which create images, sounds, smells, tastes, and tactile sensations of differing complexity. These spells create sensations which sometimes appear real to all who see them, and sometimes appear only in the minds of a chosen creature or creatures, and it can be challenging to keep track of which spells do what. 

Illusions are only as powerful as we allow them to be. To take away their power, one only has to stop believing in them...right?

The Illusionist

The illusionist, as we know it today, did not exist in the original, 1974 version of D&D. While a magic-user earned the title of "Conjurer" at 3rd-level, "Enchanter" at 7th-level, and "Necromancer" at 10th-level (all names of what would later become "specialist" wizards), there is no level at which an OD&D magic-user becomes an "Illusionist".

That's not to say illusions did not exist in OD&D - phantasmal forces, invisibility, hallucinatory terrain, and projected image all appear in Book I: Men & Magic.

In 1975, Peter Aronson created the Illusionist class for D&D, which was later published in Strategic Review and Dragon magazine. Around the same time, Len Lakofka, in his fanzine Liaisons Dangereuses, put forth four "specialties" for magic-users - one of which was the Illusionist. 

Between Aronson and Lakofka is probably the origin for the very idea that spells in D&D could be sorted into "schools", and that magic-users could choose to "specialize" in one of them. So, for all those people who are aggrieved by the inconsistencies present in the schools of magic, or by wizards later being pigeonholed into specialization as a concept - blame them! I for one really enjoy the classification of schools of magic - even when it confounds me - as well as the concept of specialist wizards.

The Illusionist officially appears as "a sub-class of magic-users" in AD&D 1e, with Gary Gygax presenting a retooled version of Aronson's concept. Illusionists gained access to a spell list which was largely distinct from that of the magic-user, including dozens of new spells which were either classified as illusions or were illusion-adjacent (1e being the first edition in which spells were sorted into schools in the PHB or its equivalent), many of which would go on to become iconic spells which survive to this day - others would be collapsed and consolidated.

In AD&D 2e, the Illusionist appears as an example of a specialist wizard - wizards could now specialize in any of the eight schools of magic, all of which had similar benefits, but different restrictions tied to the character's race, ability scores, and the types of magic they could utilize. Like all specialists with their associated schools of magic, the Illusionist was better as memorizing, casting, resisting, learning, and researching illusions. The 2e Illusionist could be only a human or a gnome, needed a Dexterity score of 16 or higher, and was unable to cast spells from the schools of abjuration, evocation, and necromancy.

Illusionists appeared in 3e and 4e, but I don't find those editions' conceptualization of the class to be particularly interesting or worth discussion. Moving on.

In 5e, wizards can choose Illusion as their Arcane Tradition. Doing so allows them to copy illusion spells into their spellbooks more quickly and for less gold, automatically grants the minor illusion cantrip (an improved version of it, which allows the wizard to create both a sound and an image with a single casting of the spell), allows the wizard to change the nature of an existing illusion (i.e., without casting the spell again), create an illusory duplicate of their self to force an attack against them to automatically miss, and make an illusory object temporarily real (the example being the illusion of a bridge across a chasm, which is "real" just long enough for the wizard's allies to run across it).

As with 5e's version of the Necromancer, it's worth pointing out that - with the exception of the improved minor illusion and the 14th-level Illusory Reality feature - none of these features encourage 5e's Illusionist to actually use illusion spells. Sigh. As with all 5e wizards, shield, counterspell, and fireball are the Illusionist's bread and butter - but hey, they get minor illusion for free, so...there's that.

Anatomy of an Illusion

What exactly is an illusion? From 2e:

Illusions deal with spells to deceive the senses or minds of others. Spells that cause people to see things that are not there, hear noises not made, or remember things that never happened are all illusions.

And from 5e:

Illusion spells deceive the senses or minds of others. They cause people to see things that are not there, to miss things that are there, to hear phantom noises, or to remember things that never happened. Some illusions create phantom images that any creature can see, but the most insidious illusions plant an image directly in the mind of a creature.

Pretty similar. The 5e definition touches on the distinction between "phantom images that any creature can see" and images planted "directly in the mind of a creature". The 2e PHB actually goes into far greater detail regarding illusions (in Chapter 7, in a section that definitely should have instead been in the DMG):

Spells of this school fall into two basic groups. Illusions are creations that manipulate light, color, shadow, sound, and sometimes even scent. Higher level illusions tap energy from other planes, and are actually quasi-real, being woven of extradimensional energies by the caster. Common illusions create appearances; they cannot make a creature or object look like nothing (i.e., invisible), but they can conceal objects by making them look like something else.

Phantasms exist only in the minds of their victims; these spells are never even quasi-real. (The exceptions to this are the phantasmal force spells, which are actually illusions rather than phantasms.) Phantasms act upon the mind of the victim to create an intense reaction - fear being most common.

I say that this section should have instead been in the DMG because it goes into great detail about adjudicating illusions - believability, monster and NPC expectations and reactions, when a saving throw is warranted, how to handle "damage" from illusions, whether illusions can scare a creature to death, when to roll for system shock, and the like (I'll get to this later). 

As is often the case with AD&D, it's verbose, but not at all unhelpful or unwelcome. There is even more text on adjudicating illusions in Appendix 2 of the PHB - really, a section titled "Adjudicating Illusions" is in the Player's Handbook!

3e goes into even greater detail when distinguishing the types of illusions. There are figments, glamers, patterns, phantasms, and shadows:

A figment spell creates a false sensation. Those who perceive the figment perceive the same thing, not their own slightly different versions of the figment. (It is not a personalized mental impression.) Figments cannot make something seem to be something else.

A glamer spell changes a subject’s sensory qualities, making it look, feel, taste, smell, or sound like something else, or even seem to disappear.

Because figments and glamers are unreal, they cannot produce real effects the way that other types of illusions can. They cannot cause damage to objects or creatures, support weight, provide nutrition, or provide protection from the elements. Consequently, these spells are useful for confounding or delaying foes, but useless for attacking them directly.

Like a figment, a pattern spell creates an image that others can see, but a pattern also affects the minds of those who see it or are caught in it. All patterns are mind-affecting spells.

A phantasm spell creates a mental image that usually only the caster and the subject (or subjects) of the spell can perceive. This impression is totally in the minds of the subjects. It is a personalized mental impression. (It’s all in their heads and not a fake picture or something that they actually see.) Third parties viewing or studying the scene don’t notice the phantasm. All phantasms are mind-affecting spells.

A shadow spell creates something that is partially real from extradimensional energy. Such illusions can have real effects. Damage dealt by a shadow illusion is real.

Someone who is more well-versed than I in the workings of 3e could probably explain why all of these distinctions are necessary or how they interact with the game's other rules.

5e makes no attempt to delineate different types of illusions. As is often the case in the most current edition, the spells' individual descriptions simply say what each does in game terms - that is, whether the spell creates a sensation which can be observed by anyone (and which sensation or combination of sensations it produces) or a sensation entirely within the minds of the affected creatures, whether it can be interacted with, whether it imposes some condition, whether it can cause damage, how to overcome it, and the like. One can assign most of the 5e illusions to any of the above categories: minor illusion is a figment, disguise self is a glamer, hypnotic pattern is (isn't it obvious?) a pattern, phantasmal force is a phantasm, and creation is a shadow.

Conjuring Enchanting Illusions

Illusion is one of the more consistent schools of magic in 5e. The spells found there are about what one would expect. There are some oddballs, and they are usually spells which would fit into the latter three categories presented in 3e - patterns, phantasms, and shadows.

Patterns and phantasms are illusions which affect the mind to "create an intense reaction - fear being the most common". Sounds kind of like an enchantment, no? Examples include fear, hypnotic pattern, phantasmal killer, and weird. It makes sense to me that these spells are in the realm of illusion, because it's the images being created which cause the mental effect. Whereas enchantments affect the mind directly, illusion can affect the mind only to the extent that the target is susceptible to the illusion itself.

Fear in particular is often brought up as being an oddball. In AD&D, the wizard "sends forth an invisible cone of terror", which, yeah, that doesn't really sound like an illusion at all. But in 5e, the caster projects "a phantasmal image of a creature's worst fears". The spell affects a 30-foot cone, so presumably each creature is seeing a different image, assuming that their fears are different. I've always thought of this as being like that scene in Fellowship of the Ring where Gandalf gets really tall and bellows at Bilbo "I am not trying to rob you!" He instills fear in Bilbo, but he does it by changing the way he looks and sounds.

Shadows, on the other hand, create "something that is partially real from extradimensional energy", which sounds a lot like conjuration. The "extradimensional energy" is usually drawn from the Plane of Shadow, and is magically woven into the illusion to make it quasi-real. Conjuration, on the other hand, involves "the transportation of objects and creatures from one location to another" or the creation of "objects or effects out of nothing." Conjuration transports or creates real things - shadow magic does not (it uses extradimensional energy to make an image temporarily real). Phantom steed is shadow magic, but find steed is conjuration. Creation is shadow magic, but create food and water is conjuration.

There are a few other oddballs, like color spray, which creates "flashing, colored light", yet other spells which create light (light, daylight) are evocation, as is prismatic spray (a more notorious spray of multicolored light). 3e classifies color spray as a mind-affecting pattern, so it's meant to be similar to hypnotic pattern.

Magic mouth is also odd, since nothing in the spell's description identifies the mouth as being illusory. It can make something like a painting or a statue appear to speak, but there's nothing saying that the mouth itself isn't real. 3e classifies the spell as a glamer, so the intention is clearly for it to be a trick of light and sound, not a real mouth that is conjured or an ordinary object like a statue that is magically altered.

I was surprised to see that silence is an illusion (3e classifies it as a glamer). I could see an argument for it being an abjuration (a magical barrier that traps sound), but I suppose it makes sense as a glamer (a spell which can make something sound like something else, or, in this case, make something sound like nothing at all).

Lastly, simulacrum - this is a doozy of a spell, and I'm not going to get into the weeds on it, but basically, the caster creates "an illusory duplicate" of a creature, which is "partially real and formed from ice and snow". However, the illusory duplicate is a creature (a construct, specifically) with its own hit points and all of the abilities and statistics of the creature it is based on. 3e classifies the spell as a shadow, so I guess shadow magic is used to give the duplicate a semblance of life.

This is a spell people love to whine about. A lot of DMs don't like that a high-level caster can make an almost exact duplicate of themselves or someone else in the party. A party of all wizards (or bards, with Magical Secrets), could hypothetical double the size of their party if they all cast this spell. I think it's iconic and awesome, and something that high-level casters should be able to do. 

Besides, the spell is plenty limiting - even if a character can get their hands on all that ruby dust, how readily available is the ice and snow to make a snow man? Sounds to me like every campaign needs a calendar with distinct seasons and a system for randomly generating weather!

Real Eyes Realize Real Lies

The way in which illusions work in D&D has changed a lot throughout the editions. Most notably, there has been a great deal of consternation regarding belief in illusions. In OD&D, phantasmal forces can only do damage if "believed to be real". In AD&D, phantasmal force grants a saving throw only to creatures which "disbelieve" the illusion. Allegedly, this led to AD&D players shouting "I disbelieve!" at anything and everything they encountered, on the off chance that they were experiencing an illusion and could benefit from a saving throw.

2e cracks down on this nonsense to a reasonable extent. From Appendix 2 of the PHB:

Disbelief must be stated by the player, based on clues provided by the DM. Players stating disbelief must give a reason for disbelief based on sensory information available to the character. Failure to give such a reason results in failure to disbelieve. The DM can impose additional requirements or delays in recognizing illusions (such as Intelligence checks) as needed, such as when one player is obviously parroting a discovery made by another. Disbelief automatically forfeits a saving throw if the effect is real.

Chapter 7's section on illusions states "disbelieving is an action...and takes a round". So, shouting "I disbelieve!" at anything and everything is not exactly a foolproof strategy. It takes a full round to attempt to disbelieve an illusion (the character foregoes attacking, casting spells, and perhaps even moving or defending themselves, as they focus on the supposed illusion and attempt to see through its tricks), and if the "illusory" fireball happens to be real, the character forfeits their saving throw against it - they were too busy trying to think the fireball away, and got blasted in the face by it.

On top of all that, to even attempt to disbelieve an illusion, the player must provide the DM with some sort of logic for doing so. This is usually tied to believability, which is exhaustively detailed in Chapter 7:

The key to successful illusions or phantasms is believability, which depends on three main factors: what the caster attempts, what the victim expects, and what is happening at the moment the spell is cast.

The type of image chosen by the caster affects the reaction of the victim. ... Spellcasters are well-advised to create images of things they have seen.

The next important consideration is to ask if the spell creates something that the victim expects. ... The key to a good illusion is to create something the victim does not expect but can quickly accept.

This then leads to the third factor in the believability of an illusion, how appropriate the illusion is for the situation. ... The best illusions reinforce these expectations to your character's advantage. 

The text is littered with helpful examples for each of these three factors, which I've left out for brevity. Suffice to say, in typical 2e fashion, there's a lot of verbose advice given, which seems largely meant to help the DM adjudicate things in a way informed by "common sense".

2e also provides some guidance for situations in which some creatures are affected by illusions while their allies are not:

In many encounters, some party members will believe an illusion while others see it for what it really is. In these cases, revealing the truth to those deluded by the spell is not a simple matter of telling them. The magic of the spell has seized their minds. Considered from their point of view, they see a horrible monster (or whatever) while a friend is telling them it isn't real. They know magic can affect people's minds, but whose mind has been affected in this case? At best, having an illusion pointed out grants another saving throw with a +4 bonus.

2e also touches on Intelligence as a factor in disbelieving illusions, though it isn't very specific: 

Intelligence is the best defense against illusions. Low and nonintelligent creatures are more vulnerable to illusions, unless the illusion is completely outside their experience or the illusion touches on an area of the creatures' particular competence. 

One has to consider how this would affect an animal, such as a wolf. It has low Intelligence, but heightened senses. A wolf may be startled by the illusion of a bear, but if it doesn't smell like a bear, the illusion won't be convincing. Thus, the limits of believability must be determined by the effected creature's experience and senses.

Furthermore, creatures with sufficiently high Intelligence are immune to low-level illusions. A creature with an Intelligence of 19 is immune to 1st-level illusions, and a creature with an Intelligence of 25 is immune to 7th-level illusions.

Others have taken the idea of Intelligence as a determining factor and run with it (interestingly, this post suggests instead that low-Intelligence creatures ought to be less susceptible to illusions - they simply don't fully process them). Using Intelligence this way might be useful for DMs who don't trust themselves to be unbiased when determining whether their monsters and NPCs might reasonably come to disbelieve an illusion, or those who simply want more mechanical heft when adjudicating this sort of thing.

That brings us back to 5e, where Intelligence plays much more of a role in disbelieving illusions. Many illusions in 5e do not involve a saving throw at all.  Generally, an illusion in 5e only requires a saving throw if it can damage the target or inflict some condition (phantasmal force, fear, hypnotic pattern, phantasmal killer, dream, or weird). The saving throw has nothing to do with disbelieving the illusion, but simply resisting the illusion's effects.

Because of 5e's very literal rules language, there are spells which explicitly state conditions under which illusions are automatically revealed. Minor illusion, disguise self, silent image, major image, hallucinatory terrain, seeming, programmed illusion, and project image are all revealed to be illusions when physical contact is made.

These same spells, with the addition of phantasmal force, all allow creatures to use their action to make an Intelligence (Investigation) check to attempt to see past their illusory nature. The DC is equal to the spellcaster's spell save DC. This is, essentially, 5e's version of the "disbelieve" action. Like in 2e, a character forfeits the bulk of their turn in order to attempt to disbelieve an illusion. 

I would also use 2e's "common sense" guidelines to determine whether a check is possible in the first place - the player would have to give me some sort of justification as to why their character would disbelieve the illusion, although a player is unlikely to forfeit their turn anyway unless they have good reason. 

Similarly, if one creature has already seen through the illusion or is unaffected by it, and alerts an allied creature to its presence, I would give the affected creature advantage on their Intelligence (Investigation) check to see through it (based on 2e's suggestion of +4 to the saving throw).

I would also utilize 5e's concept of "passive" skill checks to give additional clues to players. A passive skill check value equals the characters skill check modifier plus 10. A 1st-level wizard with 16 (+3) Intelligence and proficiency in Investigation would have a an Investigation modifier of +5, and a passive Investigation score of 15. 

Thus, if this wizard encountered an illusion created by a monster or NPC with a spell save DC of 15 or less, they would "passively" notice something "off" about the effect. They would not automatically be immune to the illusion, but it would be enough to justify an Intelligence (Investigation) check to see past it on their turn (if they're willing to use their action to attempt to do so, as normal).

Scared to Death

Can illusions do damage? Is damage done by illusions "real" damage, which persists after the illusion has ended? Can illusions kill? This is probably the biggest, most confusing change in illusions from edition to edition. 

In OD&D, "Damage caused to viewers of a Phantasmal Force will be real if the illusion is believed to be real," but also, "the illusion will continue unless touched by some living creature." If the orcs believe that the phantasmal unit of archers is real, does the illusory volley of arrows "damage" the orcs, or does the illusion cease to be once the arrows touch the orcs?

In AD&D, phantasmal force explicitly does damage: "the magic-user creates a visual illusion which will affect all believing creature which view the phantasmal force, even to the extent of suffering damage". An illusionist who concentrates on the spell can cause it to "react appropriately" to contact with an opponent, allowing it to persist even after being touched. 

However, phantasmal force is a visual illusion only. I don't see how a visual illusion alone can convince the orcs that they are being fired upon with real arrows, since they wouldn't feel the arrows (or even hear them whistling through the air). Perhaps the visual component is so convincing that the target's mind tricks the orcs into believing they have taken damage? Much and more must be read between the lines.

Tom Moldvay's Basic D&D states that "[Phantasmal Force] never inflicts any real damage. Those 'killed' will pass out, those 'turned to stone' will be paralyzed, and so forth. These effects wear off in 1-4 (1d4) turns."

2e provides a broader ruling:

Illusions are spells of trickery and deceit, not damage and destruction. Thus, illusions cannot be used to cause real damage. When a creature is caught in the blast of an illusionary fireball or struck by the claws of an illusionary troll, he thinks he takes damage. The DM should record the illusionary damage (but tell the player his character has taken real damage). If the character takes enough damage to “die,” he collapses in a faint. A system shock roll should be made for the character. (His mind, believing the damage to be real, may cause his body to cease functioning!) If the character survives, he regains consciousness after 1d3 turns with his illusionary damage healed. In most cases, the character quickly realizes that it was all an illusion.

When an illusion creates a situation of inescapable death, such as a giant block dropping from the ceiling, all those believing the illusion must roll for system shock. If they fail, they die -killed by the sheer terror of the situation. If they pass, they are allowed a new saving throw with a +4 bonus. Those who pass recognize the illusion for what it is. Those who fail faint for 1d3 turns.

Illusions are a powerful thing in AD&D! In 2e, phantasmal force is a 1st-level spell, affects a 500 square-foot area - at 1st-level, plus 100 square feet per additional level! - and can create an effect which can instantly kill every creature caught within it. The limits regarding believability and the affected senses - the spell creates only a visual effect - are still present, but those don't seem particularly difficult to overcome for a savvy player.

Phantasmal killer creates an illusory 4 HD monster in the target's mind. If the illusory monster makes a successful attack roll against the target, the target dies instantly. Only one attempt to disbelieve is allowed, and the spell lasts for 1 round per level of the caster - that's a lot of attempts to instantly kill any creature!

Weird causes targeted creatures in a 20-foot radius to see "phantasmal images of their most feared enemies". Even those who successfully save versus spell are paralyzed for a round and lose 1d4 Strength for a turn (10 rounds). Those who fail have to actually fight the phantasmal creatures with an entire separate combat playing out entirely in the minds of the affected:

Actual combat must then take place, for no magical means of escape is possible. The foe fought is real for all intents and purposes; affected creatures that lose will die. If a creature's phantasmal nemesis from the weird spell is slain, the creature emerges with no damage, no loss of items seemingly used in the combat, and no loss of spells likewise seemingly expended. The creature also gains any experience for defeating the weird, if applicable.

Although each round of combat seems normal, it takes only one-tenth of a round. During the course of the spell, the caster must concentrate fully upon maintaining it.

Appendix 2 of the 2e PHB suggests some limits on what kind of phantasmal monsters can be created and how to handle their combat statistics:

Monster Special Attacks: Before the caster can effectively duplicate a monster's special attack, the wizard must have undergone it (a wizard cannot conjure up the twinkle in a medusa's eye correctly without actually experiencing it - i.e., having been turned to stone by one).

Option: Illusionary monsters attack using the wizard's attack values. This would be a subtle clue that the monsters are fake.

Option: Extend the spell level control to monsters - the caster can create monsters only if the total monster Hit Dice are equal to or less than the caster's level (an 8th-level caster could convincingly do one hill giant, two ogres, or four 2nd-level fighters).

Thus, a 20th-level wizard could force the affected creatures to face a phantasmal 20 HD very old red dragon. It's unclear to me if the affected creatures experience the phantasm together or separately - that is, would a party of affected player characters fight the phantasmal dragon as a unit, or would each of them need to battle the dragon individually?

Modern editions of D&D have simplified things, for better or worse. It is readily apparent which illusions are capable of doing damage, because they say as much. Unfortunately, the damage-dealing capability of these spells has been greatly reduced. 

In 5e, phantasmal force does 1d6 damage per round (if the target fails the initial Intelligence saving throw, if concentration is maintained, and if the target doesn't successfully disbelieve the illusion). An instant-death illusion of a falling stone block it is not, but the caster could make it look like, I don't know...individual bricks are falling from the ceiling and plunking the target on the head every round.

Another wrinkle is that the spell does 1d6 psychic damage, but the creature "perceives the damage as a type appropriate to the illusion". Since the spell allows the caster to create an effect which occupies a 10-foot cube, the caster could create an ogre and have it do what seems to be bludgeoning damage, but what ogre does 1d6 damage? 

There's also the issue of damage resistances and immunities. If a raging barbarian is resistant to bludgeoning damage but not to psychic damage, how do they react to taking phantasmal bludgeoning damage which is actually psychic damage? If I tell the player, "You take 4 bludgeoning damage" and they say "I'm resistant to bludgeoning damage, so I instead take 2 damage", do I say "Well actually, it's psychic damage" or "No, you're not resistant to this damage...for reasons." 

It's entirely unconvincing, and almost any player will see through the ruse immediately. Maybe this is a feature, rather than a bug. Maybe phantasmal force is meant to be an obvious illusion, since it is a low-level one. Maybe.

Phantasmal killer now frightens the target if they fail the initial Wisdom saving throw. Then, at the end of each of their turns, the target repeats the saving throw, taking 4d10 damage on a failed save - not even close to instantly killing most creatures at 7th-level, which is when a caster can first learn it. What's in a name?

Weird works similarly, but affects all creatures in a 30-foot-radius sphere. It still does 4d10 damage per round, despite being a 9th-level spell. Meteor swarm, by comparison, does 40d6 damage in four separate 40-foot-radius sphere. Power word kill does effectively 100 damage. Wish can do basically anything.

In the 2024 version of 5e, all three spells have received some love. Phantasmal force now does 2d8 damage per turn, phantasmal killer does the same damage but does so immediately upon a failed save (rather than requiring a second failed save), and weird does 10d10 damage initially and 5d10 damage on subsequent turns. It's a step in the right direction, but still rather underwhelming compared to what AD&D offered.

Illusionist wizards in the 2024 version will also have a few conjuration spells always prepared, which they can cast as illusions. This is similar to AD&D's idea of using illusions to create phantasmal creatures, rather than having phantasms do arbitrary damage determined by the spell. I was somewhat opposed to this at first, as it makes the Illusionist feel like a Great Value Conjurer, but the more I've read about how illusions worked in AD&D, the more I've come around to it.

I'm not sure how to feel about all of this. Clearly, the illusions of D&D's early days - with even the weakest spells having the potential to instantly kill anyone who failed a saving throw - were far to powerful, while also being nebulous enough to be entirely defanged by DM fiat and "disbelieving" players. The illusions of modern D&D are much more mechanically hard-coded - as is almost everything else in the game, at least when it comes to doing damage in combat - but feel totally lacking in potency compared to what came before.

Is there a way to split the difference? Damned if I know.

So You've Pissed Off an Illusionist

I've been running my buddy's 5e campaign as a guest DM for a month or two now. I was burned out from running my own game, and so was he, but I felt up to the task of designing a couple of quick adventures for his campaign, where the player characters are 9th-level - much more exciting than 4th-level AD&D, in my opinion.

I start planning all of my adventures by determining the primary antagonist. For this adventure, I rolled up an archmage. Because I love specialist wizards, all of my wizards have a school in which they focus their studies. For this one, I rolled a d8 and got a 6, which according to alphabetical order meant that my archmage was an illusionist.

I started to think about how to design an adventure around an adversarial illusionist, and it came to me very quickly. Really, just look at the spells available to an illusionist and think of all the trouble they can get up to.

I decided that this arch-illusionist had it out for the party - the circumstances are too specific to get into here, but there was a good reason (in the eyes of a megalomaniacal archmage, that is). Looking at the list of illusion spells I see disguise self, illusory script, Nystul's magic aura, creation, dream, seeming, project image, and others.

Here's what happened: While the party was returning to the city after their most recent adventure, they started having terrible dreams, which prevented some of them from resting, leading to accumulating levels of exhaustion (one of 5e's nastiest mechanics). 

When they got back to civilization, they found that they suddenly had a bad reputation. Shop owners had closed their doors to them, stating that the party had sold them phony goods or paid in counterfeit coins and gems. Apparently, they had forged documents to get access to a royal gala, and had angered a number of important NPCs there.

The party knew that something was amiss, and started contacting the few friends they had left in the city. The party eventually learned that people who looked exactly like them had been going around town sullying their reputation. Some people claimed to have seen the party traveling to and from a minor noble house's estate.

The party asked around and found out that this minor noble house had recently dismissed many of their servants and guards. The dismissed workers reported that the family had begun behaving strangely - they played elaborate pranks on each other and the staff, enjoyed different food and wine than usual, stopped entertaining guests, and the like.

The party went to the estate to investigate, and found that while it appeared ordinary at first glance, it was all an illusion - the estate had become a rundown frat house. Mercenary thugs were magically disguised as noble knights, goblins as the noble children, and a cabal of illusionists as the noble family. In a secret cellar, they found the real noble family locked in a cage, trapped in a magical dream state, believing themselves to be at a high society gala with the city's greatest movers and shakers.

At the center of the deception, of course, was the arch-illusionist. He appeared in every room of the house, taunting the party until they struck him, at which time he disappeared, and a magic mouth appeared to taunt them into finding him elsewhere. In AD&D style, I gave the illusionist the ability to summon phantasmal monsters for the party to fight - they could either "kill" the monsters (exhausting their resources in the process) or "disbelieve" them to determine that they were illusions (this being a master illusionist, that was no easy feat).

The party finally found the illusionist and battled an illusory dragon he had created. They cornered him, slew him, and...he melted into a pile of ice and snow. A magic mouth appeared again, swearing that they would never find him. He was everyone, and no one. He could be anywhere he wished, and when next they found him, it wouldn't be him at all. Since the party had overcome his house of illusions, he made a gentleman's agreement to let bygones be bygones, but he warned them - he would be watching.

And that is the power of illusionist. They can be anyone. They can make someone see, hear, smell, taste, or feel anything. They can make themselves appear anywhere - even in someone's dreams. The illusionist can force someone to battle enemies that aren't there, trap them in a prison within their own mind, or scare them to death. When the illusionist finally shows up, it might not be them at all. And just like that, they're gone.