On Magic Items and Treasure Hoards

I find the official method for determining what magic items are included in a treasure hoard in 5th edition Dungeons & Dragons to be pretty unsatisfying. Basically, the Dungeon Master rolls d100 to determine the type of treasure hoard, depending on the "Tier" of the characters - Tier 1 is levels 1-4, Tier 2 is levels 5-10, Tier 3 is levels 11-16, and Tier 4 is level 17 and up (Dungeon Master's Guide, page 133). 

If there are magic items present, there will be one or 1dX magic items from Magic Item Table Y. My issue with this is that the magic item tables in question (page 144) are organized mostly by rarity (page 135) and by whether the items are "Minor" or "Major" items (I can't seem to find where the distinction is laid out, but generally speaking, "Minor" items are either consumables like potions and scrolls or more permanent items which have less significant (i.e. non-combat) utility effects - lots of "meme" items like the bag of holding, alchemy jug, immovable rod, and the like - whereas "Major" items tend to be weapons and armor, rods, staves, wands, and the like - many of which have some impact on the game's math or grant a combat ability).

So what's the issue with this method of organizing the tables? Well, I didn't really realize there was a problem until around 14th-level of my most recent campaign, where one dungeon's treasure hoard contained a +2 suit of armor, an Oathbow, and a staff of power. This in and of itself is not a problem, as these are perfectly acceptable treasures for a 14th-level party to acquire (bafflingly, the party sold the armor, gave the bow to an nonplayer character follower - who became a DPS juggernaut as a result - and the character who took the staff retired from the campaign immediately afterward, but that's besides the point). 

The issue became apparent when I prepped the next dungeon, which contained nothing but potions and scrolls. The first dungeon had a handful of great, iconic, useful items because when I prepped it I happened to roll the treasure type which directed me to roll on the magic item table consisting only of great, iconic, useful items. The second dungeon had a handful of consumables which the party promptly forgot about because when I prepped it I happened to roll the treasure type which directed me to roll on the magic item table consisting only of easily forgotten consumables (there's something to be said of players' tendency to forget about consumables and/or their hesitance to use them, but again, besides the point).

Another time, I might detail my method for determining the contents of treasure hoards, which includes magic items (the gist of it is that I use 5e's treasure tables in combination with Courtney Campbell of the Hack & Slash blog's excellent treasure tables in the Hack & Slash Blog Compendium II). For now, let's assume the DM is preparing a treasure hoard containing X magic items. If the DM is running 5e, their roll on the Treasure Hoard table tells them how many items are present and what magic item table to roll on, so they end up with a handful of the same type of item.

Contrast this with the original 1974 edition of  D&D - the monster encountered determines the "type" of treasure found (Book II: Monsters and Treasure, pages 3-4), and the type determines the chance for magic items (page 22), some of which are specific (i.e. "Weapon, Armor, or misc. weapon", "1 Potion", "1 Scroll", etc.), while others are general ("any 3"). In the case of the latter, the DM then refers to the following table (page 23):

Each of the magic item types then have their own unique tables on the following pages.

Advanced Dungeons & Dragons (both 1st edition and 2nd edition) handle it similarly - each monster has a "type" of treasure in its lair, the type refers to a table, and the roll on the table determines if and how many magic items are present. A subsequent table is used to determine the type of these magic items:

AD&D 1e DMG, page 121

AD&D 2e DMG, page 183

The result of this system is not only a greater variety of types magic items in a given treasure hoard, but because the subsequent tables for each type of magic item aren't organized with the characters' levels in mind, it also allows the chance for low-level characters to stumble across something truly special, game-changing, and character-defining.

But, I understand why "new school" DMs might be hesitant to entirely dissociate character level from the power level of magic items they might find the way earlier editions of D&D did. Which is why I'm going to suggest an entirely different magic item table from another 5e book.

Fizban's Treasury of Dragons, published in 2021, is the rare post-Tasha's Cauldron of Everything (published in 2020) Wizards of the Coast book that I actually recommend (TCE is where WotC truly lost me with lazy, sloppy products of dubious worth to DMs - not to say there were not similar products before that - but perhaps that's a topic for a later date). FTD is a good book and an obvious resource for any DM who wants to include dragons in their game with "dragons" in the title. But it also contains a hidden gem - a second, more secret, better magic item table than that found in the DMG.

The table is in Chapter 4: Lairs and Hoards, on page 72. Like other treasure tables, it uses a d100 roll. The DM is meant to roll on the table corresponding to the dragon's age category (wyrmling, young, adult, or ancient). This is similar to how the tables in the DMG are broken up by tier, or character level.

Assuming the DM knows how many magic items are present (FTD similarly suggests a number of magic items in a given dragon's hoard, again determined by the dragon's age category, but the DM might prefer to use the DMG's tables as to number of items, or some other method of determination like I do with Courtney Campbell's), the DM can roll d100 however many times on the appropriate column of the table in FTD (treating wyrmling as Tier 1, young as Tier 2, adult as Tier 3, and so on). The DM can decide if the column they're rolling on is determined by character level or by the "tier" of the adventure site where the treasure is found (if the DM is running a sandbox, for example, where characters can presumably attempt to go anywhere and do anything, Tier 1 characters might end up trying their luck in a Tier 2 adventure site, and end up with Tier 2 treasures as a result).

This method produces a spread of magic items which is more diverse in terms of rarity and distribution of minor and major items. A Tier 1 party (or a party adventuring at a Tier 1 location) is still most likely to find common minor items, but there may be rarer items and major items mixed in as well. Another option, which I haven't playtested, is to treat a roll of 100 (or the DM's preferred number) as a reroll on the next highest column of the table, so that even greater items occasionally appear as well.

With both this system and that laid out in the DMG, it's still impossible to randomly find an artifact-level magic item. The DMG states that "artifacts only appear when [the DM] wants them to", hence why they're not included on random tables. Maybe this is good advice and what the DM wants, or maybe it isn't. I know from my own prep style that I'll probably never remember to deliberately place artifacts in my game. If the DM wants a chance for artifacts to appear randomly, they might treat a roll of 100 on any column of FTD's table as an artifact. They may want to follow up a roll of 100 with a subsequent d100 roll to confirm. Or, a roll of 100 might only result in an artifact when it's rolled on the highest (ancient/Tier 4) column of the table.

I would go so far as to also use the tables from past editions (2e is mine of choice) to determine the type of each magic item (i.e. potions and oils, scrolls, rings, etc.) so that these are also diverse - however, some knowledge of what constitutes a minor or major item in 5e would be required. For example, most or all consumables like ammunition, potions, and scrolls are considered minor items, and there are only a handful of minor armors, rings, rods, shields, staves, and wands. "Wondrous Items" (formerly known as "Miscellaneous Magic") can be either minor or major depending on what they do.

A secondary benefit of this system is that, while the DMG's magic item tables (understandably) include only those magic items published in the DMG, this method allows the DM to include magic items from those other sources published since the DMG's release. Using D&D Beyond (or some other filterable or searchable database of 5e resources, for those of us who don't give money to WotC anymore 😉 ), the DM can filter or search for minor or major items of the appropriate rarity and type to choose from. Frustratingly, at some point WotC stopped identifying items as minor or major (the aforementioned table from FTD uses the terms, but the items in the book itself are identified as neither - like I said, lazy, sloppy, of dubious use, etc.), so the DM will have to use some judgment as to what constitutes a minor or major item in some cases.

And that's it! Hopefully this post communicated what I intended, which is that the treasure and magic item tables in the DMG provide same-y, unsatisfying results (for me - maybe they're great for every other DM). I enjoy a bit more variety in the type, rarity, and power level of magic items my players find, and I don't mind doing a bit of extra work to get the results I enjoy seeing in prep. There are undoubtedly a lot of DMs who simply include the items they want to be in their games, whether because they're cool, important to the setting or "story", or particularly suited to a specific player character, and those DMs probably don't notice these issues at all. My prep style is randomization plus imagination, and I'm always looking for ways to make my randomization methods yield results that are satisfying and inspiring.

On Experience

Rewarding experience points is out of fashion among modern Dungeons & Dragons players, and the designers of D&D seem to recognize this. Although the 5th edition Dungeon Master's Guide, published in 2014, lists experience points as the default mode of advancement ("most often the reward for completing combat encounters", page 260), a quick perusal of any recent Wizards of the Coast adventure module or discussion among 5e Dungeon Masters will quickly reveal that the prevailing form of advancement is what people now call "Milestone Advancement" (what they mean is "Level Advancement Without XP" (page 261), specifically "Story-Based Advancement", but sometimes "Session-Based Advancement", and other times "player characters level up when the DM decides they do", or "Fiat-Based Advancement" - referring to all of the above interchangeably as "Milestone Advancement" seems to have won the hearts and minds of 5e players, regardless of how the DMG itself defines "Milestone"). It's gotten to the point that I wonder if XP will even be in not-OneD&D (NOD&D)'s forthcoming DMG.

People seem to prefer "Milestone Advancement" for a variety of reasons. Story-Based Advancement ties character advancement directly to the advancement of the "story", signaling to the players that they're making progress through the pre-determined plot (as defined by the DM or adventure). Session-Based Advancement makes each session count, even if nothing of note is accomplished. Fiat-Based Advancement grants DMs the authoritarian power to arbitrarily decide whenever the players advance, usually because they or the players are bored at the current level and the DM is ready to design/the players are ready to defeat the next tier of challenges. 

XP tracking is viewed as another onerous burden upon the player characters, like tracking encumbrance, torches, or food and water. Nobody likes doing math and bookkeeping (i.e. using the calculator on your phone to add one number to another number and write it down), and it's a real pain to erase one number and then write down another one (all that erasing puts an unseemly amount of wear and tear on the pristine character sheet - although a large contingent of players are using some form of virtual tabletop or digital character sheet at this point).

But I like XP. I like it for a lot of the reasons that people like other methods of advancement. Tying XP gain to quest goals gives players the same feeling of progress as Story-Based Advancement and incentivizes attendance the same way Session-Based Advancement does (if XP is only granted to characters whose players show up, which I don't always do, but it is useful for open table games). Unfortunately, XP does take away some measure of the DM's draconian prerogative to advance the characters anytime they feel like it, assuming the DM isn't the type to handout fistfuls of miscellaneous XP on a whim.

The main reason I use XP is because it smooths out level advancement. Sure, with the other forms of advancement you can work out a structure so that the number of "pips" needed to level up increases alongside the PCs' level (such that "Tier 1" characters require a single bump per level, "Tier 2" characters require two, "Tier 3" characters require 3, etc., or any number of more complex structures). Most WotC adventures do this, and I have no doubt that each DM who doesn't use XP has their own structure they prefer, but it doesn't hit the same way the XP curve does.

That XP curve, by the way, does have some wonkiness in 5e. For example, it takes 21,000 XP to go from 10th- to 11th-level, but only 15,000 to go from 11th to 12th. After that, it takes 20,000 to advance to 13th and 20,000 again to advance to 14th. It takes 30,000 XP each to advance to 16th, 17th, and 18th (Player's Handbook, page 15). In short, the curve has a very random dip right in the middle and then stops being a curve. I know fellow DMs who have taken it upon themselves to "fix" the curve, but it's never bothered me enough to implement it in my games. 

Treasure is Its Own Reward

I would be remiss not to mention XP-for-gold, the old-school method of advancement which has been discussed by OSR bloggers for a long time. The idea is that the players will seek out whatever gets them XP. In old-school games, the players will prioritize obtaining gold, usually by pulling it out of a dungeon. Exploring the dungeon to discover its hidden wealth becomes the primary goal of the players, and their method of dealing with the dungeon's challenges - be they monsters, traps, tricks, or the layout of the dungeon itself - is entirely up to the players, so long as they get the treasure. If they can do it in such a way as to minimize the risk to their characters, all the better.

Similarly, Story-Based Advancement incentivizes the advancement of the "story", Session-Based Advancement incentivizes attendance, and Fiat-Based Advancement incentivizes...appeasing the DM, I guess.

Were I to run OD&D or AD&D, I'd certainly use XP-for-gold. There isn't much guidance for awarding XP for anything else. I've considered using XP-for-gold in 5e, but it doesn't quite work. WotC adventures are notoriously light on treasure (not that I run their adventures), and 5e's treasure hoard values only increase by "tiers", i.e., all treasure hoards from levels 1 to 4 are approximately equal in value, then bump up at 5th-level, then again at 11th and 17th. This means that characters in 5e would advance quickly at the beginning of each tier of play, then slow down at the end of each tier until they bump up to the next one. Compare this to older editions, where dungeon level corresponds (more) directly to character level and includes treasure hoards appropriate to that level.

OD&D Book III: The Underworld and Wilderness Adventures, page 7

I don't like XP-for-gold in 5e for another reason: treasure it its own reward. Gold and magic items are an incentive in addition to XP. Gold increases player character power by granting them access to better equipment, strongholds, social influence, and the like. Magic items increase PC power by granting conditional abilities and other benefits in addition to those they gain by advancing in level. XP is also being granted for navigating the monsters, traps, tricks, and dungeon and wilderness environments which stand in the way of the PCs and treasure, so it doesn't make sense to me to award XP for both overcoming those challenges as well as obtaining the treasure itself.

Murderhobo-Based Advancement

A common refrain when criticizing XP in 5e is that it's only earned through combat, which incentivizes the players to kill every creature they come upon, even if the creature is friendly or non-threatening! The players could go on a rampage in every settlement, murdering all of the villagers to accumulate that sweet 10 XP from all those helpless CR 0 Commoner stat blocks. After all, they don't get XP unless they fight, and surely they would get XP for killing all those civilians, right?

I attribute this way of thinking to that terrible aforementioned quote from the DMG, "Experience points (XP) fuel level advancement for player characters and are most often the reward for completing combat encounters." Surely this quote is actually saying that XP is only ever the reward for completing combat encounters, right?

The DMG actually does callout awarding XP for "noncombat challenges", just somewhat unhelpfully: 

"You decide whether to award experience to characters for overcoming challenges outside combat. If the adventurers complete a tense negotiation with a baron, forge a trade agreement with a clan of surly dwarves, or successfully navigate the Chasm of Doom, you might decide that they deserve an XP reward.

As a starting point, use the rules for building combat encounters in chapter 3 to gauge the difficulty of the challenge. Then award the characters XP as if it had been a combat encounter of the same difficulty, but only if the encounter involved a meaningful risk of failure."

Okay, sounds good. Wait, why are we using the rules for building combat encounters to gauge the difficulty of a noncombat challenge? How do I do that? Does the Chasm of Doom have a stat block?

Actual Milestones

The solution for me was when I discovered what "Milestone" actually means, according to the 5e DMG:

"You can also award XP when characters complete significant milestones. When preparing your adventure, designate certain events or challenges as milestones, as with the following examples:

• Accomplishing one in a series of goals necessary to complete the adventure.
• Discovering a hidden location or piece of information relevant to the adventure.
• Reaching an important destination.

When awarding XP, treat a major milestone as a hard encounter and a minor milestone as an easy encounter."

Forget that first part. I only care about the last sentence. How much XP is a hard encounter worth? What about an easy encounter? What about everything else?

XP-for-Challenges

This is one of my favorite tables in the 5e DMG (page 82), and it's the only thing I reference when determining how much XP the player characters have earned. Forget monster stat blocks. You can forget "Milestones" as defined by the DMG as well. You can forget them as defined by everyone else, too. This is all a 5e DM needs. 

Player characters don't earn XP by defeating monsters, or advancing the story, or by their players showing up, or making the DM happy. They earn XP by overcoming challenges. That might mean fighting monsters or using some other means to bypass them, neutralizing traps, puzzling out tricks, traveling to a location for the first time, discovering a secret, completing a quest (or advancing the "story"), or any number of things. If the PCs did something non-trivial, they probably get XP for it.

How much XP do they get? The table above breaks it down. Each PC gets XP according to their level and the difficulty of the challenge overcome.

How difficult is the challenge? The DMG defines each level in combat terms:

Easy. An easy encounter doesn't tax the characters' resources or put them in serious peril. They might lose a few hit points, but victory is pretty much guaranteed.

Medium. A medium encounter usually has one or two scary moments for the players, but the characters should emerge victorious with no casualties. One or more of them might need to use healing resources.

Hard. A hard encounter could go badly for the adventurers. Weaker characters might get taken out of the fight, and there's a slim chance that one or more characters might die.

Deadly. A deadly encounter could be lethal for one or more player characters. Survival often requires good tactics and quick thinking, and the party risks defeat.

This can be applied to combat encounters as written. This method has the added benefit of circumventing 5e's wonky Challenge Ratings for monsters. The weirdness of the CR system in 5e is well documented. The shadow (Monster Manual, page 269) is far deadlier than its CR 1/2 indicates. An elemental myrmidon (Mordenkainen's Tome of Foes, page 202-203) is considerably less dangerous than a regular elemental (MM, page 124-125), but the myrmidon's CR is two points higher. Plenty of times in my games I've intended for an encounter to be a certain difficulty level, only for the players to trounce it resoundingly or struggle more than I expected. 

Using this method, the PCs earn experience not based on the CR of the monsters they encountered, but on the actual difficulty of the fight (or other method by which the monster was avoided, neutralized, or overcome). If the big bad boss fight was resolved unexpectedly easily, such as by a wish neutralizing the enemy in the first round, then that's an easy encounter regardless of the enemy's CR. If the characters could easily get past the guards by killing them, but instead engage in a lengthy and clever social interaction or subterfuge, they get XP based on how difficult it was to accomplish that, not based on the guards' CR.

Similarly, for traps and tricks, the PCs earn experience based on the difficulty of overcoming the trap. If the trap was merely a setback and the PCs lost few hit points, they earn experience as for an easy encounter. If they spent some resources to overcome it (including healing resources), it counts as medium. A trap that nearly kills a PC is hard, and a trap that does kill a PC is deadly.

For other types of encounters, the DM needs to use their judgment. A ruin which is visible from the road just outside town is easy to find (but might have a secret entrance which is hard to find), whereas navigating trackless jungle or high altitude mountain peaks to find a different ruin might be hard or even deadly. Finding a single clue in an investigation might be easy, but putting the clues together and conclusively identifying and apprehending the culprit might be worth deadly XP. A quest which is resolved in a portion of a session is easy, but a big adventure which takes months of game time is worth deadly XP at its conclusion.

Sometimes, the players won't accomplish anything in a session. This has happened in my games when the players take a session between adventures to decide what they want to do next, or spend that time carousing or shopping in the city. I like for my players to feel that every session is "worth it", even if nothing of major consequence was achieved - after all, they took time out of their day to show up, and often commit to do so every week, sometimes for years at a time, regardless of what's on the agenda for that week's session. In these cases, I usually give each PC XP equal to an easy encounter, as a sort of consolation prize for showing up, sometimes sitting through bookkeeping or roleplaying that doesn't concern their character at all.

The system is flexible, and that's why I like it!

On Total Party Kills

Total Party Kills are a funny thing in Dungeons & Dragons. The "killer Dungeon Master" is a myth and a trope as old as the game itself, originating in the earliest days of D&D when the game was more lethal and DMs were encouraged to be more adversarial, and becoming popularized in the imagination of the play culture due to internet discussions and, eventually, memes. DMs love to lean into the trope too, in spirit, if not in practice. If you were to form your conception of D&D based purely on the "dndmemes" subreddit, you'd probably think that every DM was out to get the player characters and that PCs were dying left and right in every session and campaign. DMs are constantly excitedly talking about the next way they're going to "get" their players' characters, yet stories of actual TPKs treat them as some monumental, rare, unforeseen, and often unfortunate outcome. 

After all, a TPK means the game is over.

Right?

"Dead Adventurer" miniatures from Tiny Furniture

The popularity of 5th edition D&D and associated actual plays like Critical Role has polarized the player base on the topic of TPKs. After all, D&D is a game about telling a "story" about heroic characters who overcome the odds to vanquish evil and save the realm, and they can't do that if they're dead.

Right?

(As an aside, I don't think that this is what D&D is "about", and it certainly isn't the assumption in my games, but it does seem to be the popular way for modern players to approach D&D and 5e in particular.)

This "to TPK or not to TPK" discussion is also roughly as old as the hobby itself. Gary Gygax wrote the following in the Advanced Dungeons & Dragons (1st edition) Dungeon Master's Guide (Conducting the Game, page 110):

Interestingly, this passage draws a distinction between PCs who "die through no fault of [the player's] own" and those who die because the player has "done something stupid" or "not taken precautions".

It's also worth noting that the idea of story-focused campaigns in which PCs more or less have "plot armor" is at least as old as AD&D 2nd edition's Dragonlance modules, in which the players control characters from the novels who are unable to die for story reasons. While these modules may have started the trend towards TSR's and Wizards of the Coast's D&D adventures being "story"-focused, I still feel that the 5e play culture is uniquely occupied with the preconceived notion of "story", even if this tendency is present to some extent in all fantasy TTRPG players due to the outsized influence of fantasy epics like The Lord of the Rings (which is not necessarily the type of fiction D&D was originally attempting to ape, hobbits and balrogs aside).

Total Party Defeat and Campaign Collapse

Although TPK stands for "Total Party Kill", suggesting that all PCs are dead, it's worth it to discuss Total Party Defeats as well. A TPD is when the PCs are soundly defeated in combat but not necessarily killed. PCs always have the option of subduing rather than killing monsters and nonplayer characters, and that's a two-way street. TPD'd characters might be imprisoned, ransomed, robbed, or put into any number of detrimental positions depending on the monster or NPC's goals. There are other TPD conditions as well, which don't necessarily result in direct harm to the PCs, such as the villain achieving some goal - killing a key NPC or securing a powerful magic item. These conditions don't end the game. Why should a TPK?

I can't think of any time that I've TPK'd my players, but I can think of a few TPDs. On one occasion, the PCs were double-crossed by a cambion and a rival party of adventurers. Ambushed at their camp while resting after a harrowing fight with plague-ridden demons, two PCs were killed, two fled, and the fifth was taken prisoner. On another occasion, three were petrified by beholders while three barely escaped. In a different campaign, the PCs fell to infighting over a magic item.

Only the latter ended in a Total Campaign Collapse, which was because the in-game defeat was so fractious for the real-world players that we felt the only way to continue playing was to start something fresh (the perils of immature, adolescent gaming). On the other two occasions, the party regrouped and liberated, resurrected, and unpetrified their companions and eventually overcame the enemies that defeated them.

Well that's all well and good, but the PCs can't exactly do all that if they're all actually TPK'd, so surely the campaign must end in a TCC.

Right?

Total Communication Failure

There has been some consternation as to whether a TPK is a failure of the DM or of the players. It can be one or the other or both, but doesn't have to be either. There is such a thing as good DMing and good play. 

A good DM provides their players with information to make choices. As it relates to TPKs, this information is mainly concerning threat level. In an "old-school" game, threat level is usually consistently determined by the level of the dungeon the players are exploring, whereas the wilderness's danger is more unpredictable due to the nature of random encounters. In more linear games (and often, adventure modules), players can usually trust that as long as they are doing what they're "supposed to do", the threat level will be appropriate for their characters. 

This is more challenging in a sandbox game, where areas of varying danger will be scattered about somewhat randomly. Occasionally, threats will be telegraphed by some other means, such as by having distinct regions which are obviously more or less dangerous, or by having more dangerous areas located further from civilization. A good DM will make the players aware of these game assumptions or rules of thumb. In a more random sandbox, a good DM will use rumors and other delivery methods to convey information about threats to the players.

Good players use information about threats to make wise decisions, whether by taking game assumptions into consideration or by making an effort to gather in-game information before embarking on a given adventure. They will use their own judgment to determine whether to attempt to explore the next level of the dungeon or venture out into the wilderness, engage in "side quests" before advancing the "story", travel to the next, more dangerous region, or choose to embark upon either the adventure where they'll be fighting goblins or the adventure where they'll be fighting ogres.

A failure by the DM is a failure to convey information. A failure by players is a failure to make decisions that correctly take that information into account.

Total Imagination Failure

There is an idea that a TPK is due to a failure of imagination on the DM's part - every monster or NPC should have a motivation other than killing the party, so if the PCs are defeated, they always have a chance to live another day and fight again. But sometimes monsters are just monsters - the dragon's motivation might be simply to eat someone (or four-to-six someones); the undead abhor life and seek to snuff it out at every opportunity; the ooze just dissolves things.

I opt to take the choice out of my hands by determining monster and NPC motivations somewhat randomly, depending on alignment, reaction rolls, and random tables to determine personality traits, ideals, bonds, flaws, and the like. Whether my monsters and NPCs fight to kill, subdue, or achieve some other end isn't strictly up to my imagination, but some combination of imagination and randomization.

In the same vein, a TPK could be a failure of imagination on the players' part - they never imagined the possibility of fleeing, for example.

The true failure of imagination is the failure to imagine what comes after the TPK.

In a typical D&D game, the afterlife is a real and somewhat tangible thing. Depending on the setting, the PCs' souls are whisked away to some other plane of existence, where they may or may not maintain agency, and where further adventures are surely possible. 

In my recently-concluded 5e game, I planned for a true TPK from the outset. In the event of a TPK, each character would be spirited away to an Outer Plane corresponding to their alignment - the Chaotic Good barbarian would find themselves in the Beastlands, the Chaotic Neutral wizard in Limbo, the Lawful Evil ranger in Gehenna, the Lawful Neutral paladin in Acheron, the Neutral Evil warlock a lemure in Hell, and the Neutral Good bard in Elysium. They would have their work cut out for them if they intended to traverse the Great Wheel or head to Sigil to regroup, but adventure would await them wherever they went, and with good play there was always the possibility of returning to the mortal realm. We never got there, but I was ready for it if we did.

As always, DM and player communication and collaboration is key. If the goal is to play a story-focused game with high narrative stakes in which character death will seriously derail the campaign and cause it to collapse, or the DM wants to try their hand at being the "killer DM" or run a Tomb of Horrors-style adventure, those expectations should be communicated at the beginning of the campaign. Both DM and players should be on the same page as to whether the PCs are destined heroes or fodder. If a TPK does occur, it warrants a discussion between both parties as to whether the game should continue in one manner or another, either with the same characters in a new situation, new characters in the same setting, a new setting, or a new game system entirely.

On Character Creation

It behooves any Dungeon Master to write out their character creation rules for their players. I include these rules in the same document as any house rules I might be using. Obviously, if character creation in the game adheres to the rules laid out in the Player's Handbook, this isn't strictly necessary. 

While character options (races, classes and subclasses, backgrounds, feats, spells, etc.) presented in later supplements (and in the PHB, page 163, in the case of feats and multiclassing) often include a disclaimer that they are options only at the DM's discretion, in my experience, players will often assume that every published option is available to them regardless of the campaign unless the DM makes the (un)available options explicit. Thus it's important to have at least a bulleted list with general guidelines (or, a short conversation between DM and players). 

This post summarizes the character creation rules I use in 5th edition Dungeons & Dragons (and some thoughts on how to use them for other versions of the game).

Choose a Race

The player can choose what race their character will be, or they can roll on a table to randomly determine this. The table is as follows:

d20

1-13      Human

14-17    Roll d6 (1-2: Elf; 3-4: Dwarf; 5-6: Halfling)

18-19    Roll d10 (1-2: Dragonborn; 3-4: Gnome; 5-6: Half-elf; 7-8: Half-orc; 9-10 Tiefling)

20         Anything else

If the player chooses to randomly roll, they can add +1 to an ability score of their choice after determining ability scores.

The conceit of this table is thus: 

• I want the races in the Player's Handbook to be most common (thus why "Anything else" makes up only 5% of the table).
• Of those in the PHB, I want humans to be most common (65%).
• Of the remaining races, the latter five are specifically called out in the PHB as being uncommon, so elves, dwarves, and halflings must be the "secondary" races to humans (20%), where the latter five are tertiary (10%).

A DM can easily take these same percentages and change the races associated with each to suit their setting. If the game is OD&D, I would omit everything except humans, elves, dwarves, and halflings. In AD&D, I would also include gnomes and half-elves, and the final 5% might be the various race options found in the edition's many splat books (if I wanted to ruin my game by allowing players to make firbolgs, pixies, ogre magi, and the like, which I've done in the past). 

My guiding star is that whatever's in the PHB should be most common, and all supplemental options should be relatively rare, but this can vary by setting. For example, I have a separate setting where these "common" races are all but extinct, and "monstrous" races are the predominant inhabitants, and I use the more granular d100 to create a table with a greater array of race options inspired by the Monster Manual. The point of providing a table like this is that the players can instantly get an idea of what races are most common as well as what others are present in the setting. The exact composition of races and their rarities is important only to the individual campaign.

The actual percentage breakdowns are modified from certain tables in AD&D, namely random encounter tables (where common, uncommon, rare, and very rare monsters are typically broken up almost exactly like this, 65/20/10/5). I use this same breakdown for my own random encounters. 

At the DM's option, the percentages provided to players can also be used when creating NPCs, ensuring that the racial distribution of NPCs is similar to that of player characters. 

I grant a +1 to an ability score for rolling on the table because I want to incentivize players to randomize character creation. Most of my character creation rules attempt to incentivize this. This is an attempt to make party composition more interesting and varied than the characters the players might tend to create. (Interesting and varied, in this case, is relative to the usual choices made by the players - obviously this method creates a lot of human PCs, which is certainly less varied, though not inherently less interesting, than a party consisting of a tiefling, dragonborn, automaton, half-hag, and owl person. But if the players in that group always make those exact characters, this method is meant to encourage them to try something else.)

I've been in too many games where each player essentially makes the same character every time, and I wanted to try to break up those habits. If the players want to continue playing the same types of characters anyway, they can still do that, of course.

Determine Ability Scores

Players can determine their characters' ability scores using the standard array or point buy methods presented in the Player's Handbook (page 12) or by rolling dice. If the player chooses to roll, they can roll 4d6 and drop the lowest die six times or roll 3d6 six times. If the player chooses to roll 3d6, they can choose a feat of their choice at 1st-level (in editions which don't include feats, I would just choose whether players roll 3d6 or 4d6, or I would offer both and grant some other appropriate benefit for choosing to roll 3d6).

Rolling 3d6 may seem odd in 5e, but I think 5e characters are robust enough to cope with having more average ability scores. I have had a few players take me up on this in the past - the extra feat is too tempting to pass up. More new-school players who are concerned about having "bad" ability scores are free to choose 4d6 or one of the more dependable ability score generation options from the PHB.

Whether the player rolls 4d6 or 3d6, they can either assign the scores in any order they like, or assign the scores in the order they are rolled (Strength, Dexterity, Constitution, Intelligence, Wisdom, and Charisma in 5e, but the order might differ by edition, for example in OD&D I believe it would be Strength, Intelligence, Wisdom, Constitution, Dexterity, and Charisma). 

If the player chooses to roll their ability scores in order, they can add +1 to an ability score of their choice after determining ability scores (again, I want players to roll their stats in order, so I'm trying to incentivize it by offering a mechanical benefit in exchange).

The player then adds any ability score adjustments from their chosen race.

Players can also choose or roll to determine their character's age category, as follows:

d20

1-13     Young                    

14-17   Middle Aged    -1 Str, -1 Con, +1 Int, +1 Wis

18-19   Old                   -3 Str, -2 Dex, -2 Con, +1 Int, +2 Wis

20        Venerable         -4 Str, -3 Dex, -3 Con, +2 Int, +3 Wis

If the player chooses to roll to determine their character's age category and gets a result of Old or Venerable, they can add +1 to an ability score of their choice after determining ability scores (the benefits to Intelligence and Wisdom don't balance out the penalties to Strength, Dexterity, and Constitution in these age categories, so I'm offering another mechanical incentive to mitigate the hurt a bit).

The age categories are pulled from AD&D, and may seem overly punitive for a 5e game, however all this really amounts to is that older characters are less likely to be physically-oriented and more likely to be clerics, druids, and wizards, which I think is fine.

I have thought at times that players should choose their class before determining their age, thus allowing for the possibility of an "old fighter", for example, however I'm of the opinion that character creation represents only a snapshot of the character at a moment in their life. After all, according to Gary Gygax, Conan's class levels change with age.

At this time, the player adds any ability score adjustments accumulated from rolling to determine race, age category, or rolling ability scores in order. No ability score can exceed 20 (if the DM wants a more "balanced" 5e experience, they might cap it at 16 instead, but I personally don't care if a 1st-level character starts with a 20 - the faster the players maxes out their character's important ability scores, the sooner they can consider actually interesting ways to use their ability score improvements, such as by choosing feats or preparing to multiclass by increasing secondary or tertiary ability scores).

Choose a Background

I have no special rules for this, but one could easily use the percentage spread I've used so far to create common, uncommon, rare, and very rare backgrounds determined by the setting, if they wish.

Choose Languages

When choosing which languages a player character knows, exotic languages count for two language proficiencies. Thus, if a background allows you to learn "any two languages", a player can choose for their character to know two standard languages or one exotic language.

Standard and exotic languages are laid out in the PHB (page 123), but a DM can further customize these lists by adding or subtracting languages or switching standard and exotic languages according to the setting. For example, in my setting where "common" races are extinct, those races' languages are considered exotic, and other, traditionally exotic languages are considered standard.

Choose a Class

At this point, character creation can proceed according to the chosen system without much or any changes. In 5e, I use the following prerequisites, meaning a player character has to have at least the following score(s) in a given ability in order to choose the class (these are based on the ability score requirements for multiclassing into a given class, which are included in the 5e PHB):

Barbarain    Strength 13

Bard            Charisma 13

Cleric          Wisdom 13

Druid          Wisdom 13

Fighter        Strength or Dexterity 13

Monk          Dexterity and Wisdom 13

Paladin        Strength and Charisma 13

Ranger        Dexterity and Wisdom 13

Rogue         Dexterity 13

Sorcerer      Charisma 13

Warlock      Charisma 13

Wizard        Intelligence 13

The DM could take this a bit further, using something like the ability score requirements from AD&D if for example they wanted bards, druids, monks, paladins, rangers, and the like to be especially rare. I haven't done this in my 5e games, but I like the idea. I'd have to come up with something for warlock, since as far as I know they're the only class to never appear in an edition where ability score requirements were the norm (I didn't DM 3e or 4e, so I could be wrong).

That's pretty much it. The point of this is to customize character creation to facilitate the kind of game I want to run. I like truly randomly generated characters over carefully crafted ones, and I don't mind more or less powerful characters that may result from rolling ability scores, aging penalties, free feats, and extra ability score bonuses. 

Rolling ability scores in order produces more varied characters with interesting arrays (the buff wizard or the smart fighter). It allows for more interesting multiclass combinations that you wouldn't normally see, and starting with higher stats allows players to spend fewer of their ability score improvements on maxing out their primary ability score, which allows for more robust character customization much earlier in the level progression.

It's not for everyone, and I've had more than one person claim that this somehow completely "breaks" 5e, and that's fine. They can find another game to play or run their own. The customizability of procedures like this is one of TTRPGs' greatest strengths!