One (single magic item) to rule them all

I've been thinking for a while now about the excessive multiplication of magic items in D&D. 

Whether through random tables or official modules, players quickly become overloaded with items. And this ends up stripping away all the special character these items should have. Many times I've asked myself (and others) if this aligns with the literary works that inspired D&D (Appendix N etc). 

Sometimes Moorcock's works are cited, since Elric has several magic items and so does Corum. there are examples on Tolkien too. But I think there's a distinction to be made there:

It seems like most of these stories have just one magic item that clearly stands out above the rest: Stormbringer, in Elric's case, Excalibur, in King Arthur's case, or the Holy Grail, in other Arthurian stories, or The One Ring, in Tolkien's case, the decks of cards in the Amber series, and so on.

And D&D-based literature seems to follow the same line, from The Crystal Shard, which gets its own book, to Dragonlance, which lends its name to an entire series.

Of course, even when there's just one protagonist item per campaign or setting, there are several other items as well. In Tolkien's case, for example, we have Andúril, the (quite powerful) Palantíri, and even Sting and mithril armor. Stormbringer itself has its twin sister sword, which gets less attention in the book.

However, what definitely does not happen in most of the literature mentioned is the constant accumulation of magic items by a single character, or quickly swapping those items for better ones. Every magic item carries a story, which is why simply discarding it seems to diminish its value and importance.

In addition, such legendary items should play a central role, not something you find in a random dungeon or that is easily traded. Think of the Arkenstone in The Hobbit, which has monetary value but also a sentimental/historical values that may cause or resolve a war, and ends up feeling more relevant than The One Ring at least for this book.

So why do D&D PCs get so many (disposable) items?

One of the reasons this happens is that, in older versions of D&D, magic items worked as a way of making PCs not only more powerful but also more unique. Often, equipment was what differentiated one character from another, in the absence of feats or even talents that could be chosen with new levels. Two tenth-level fighters, for example, would differ greatly from one another depending on the items they carried.

Now that modern D&D has feats and new class talents as you level-up (well, even BECMI and AD&D eventually got their skills, proficiencies, etc.), the magic item treadmill feels unnecessary.

But having only "meaningful magic items" is not an easy thing to apply at the table, precisely because D&D players are so used to, and expect, receiving more and more magic items. 

Another problem is that D&D wizards essentially have no clear ceiling on their power. A D&D wizard with enough XP has more power than a Palantír (or a dragon, making dragonlances somewhat pointless), so many magic items may become worthless as wizards level up.

Maybe one way would be to separate legendary items from more disposable ones more clearly. Something I've been using in my (science-fantasy) campaign is making some of the supposedly "magic" items mass-produced. You could do something similar by saying that a +1 sword is just better quality and can be safely traded for a +2 sword without any regard for its story or "magic".

Since I don't like writing a post about a problem without at least trying to offer a solution, here are a few attempts:

Multiplication. A magic sword that doubles or even triples its damage, for example. If a regular sword deals 1d8+1 damage, the magic sword deals 2d8+2. But if the character, for whatever reason (strength, proficiency, etc.), deals 1d8+5 with a common sword, the magic sword deals 2d8+10. That way, it's ALWAYS meaningful, but it's far more dangerous in skilled hands.

Absolutes. Items that simply bypass the rules in an absolute way: this item is completely immune to magic, or it's basically indestructible like "the One Ring", or it simply negates any possibility of resurrection, or ignores armor completely. D&D itself already has a few items like this, they just need to be part of the setting's history instead of thrown into a lsit of artifacts.

Items that grow with the character. The longer the character keeps the item, the more they learn from it and bond with it. Something I've suggested before, but which would also work with sentient items like Stormbringer, or mysterious items like the Claw of the Conciliator, whose effects, even by the end of the book of the same name, aren't 100% clear (which is also the case with The One Ring and Stormbringer).

"Serious" fumbles I can enjoy

As I’ve said before, I usually don’t like fumbles or critical failures in combat, at least in more "serious" campaigns; they make fighters look foolish.

Worse, the higher level a fighter gets, the more attacks they make; if every natural 1 is a fumble, fighters end up failing far too often. When you’re rolling 4 or 5 attacks per round, one of them is almost guaranteed to be a ridiculous blunder. Critical failures do happen in real life, but not nearly as often as a single die face suggests.

The idea of a saving throw to confirm whether the fumble actually happens is a decent (mathematical) fix; but with multiple attacks and multiple saves you end up with lots of rolls that don’t lead anywhere.

Instead of focusing on the character, we could focus on the weapon or the environment. Keep fumbles, but only in situations that are genuinely risky; and the effects shouldn’t make the character look like an idiot, but highlight the limitations of the weapon or the setting instead.

For example, a longsword needs space to be effective. In a cramped tunnel it still works (you can use half-swording, etc.), but it’s suboptimal; that could cause the fumble. You could even build a table of things that might go wrong on a natural 1, but only if it makes sense in context. If there’s no additional danger, then nothing funny happens.

Another option is to give the enemy an opportunity to strike with an advantage; maybe you overextend, make a reckless swing, and miss, opening yourself up to a counterattack. That way the focus isn’t on your “stupid mistake,” but on the danger you’ve exposed yourself to.

Let's try to combine both ideas. Here is how an actual rule I might use would look like:

---

When you roll a natural 1 on an attack, it is always an automatic miss. In addition, it can cause serious consequences if you’re in a risky situation:

• Close allies: If an ally is too close to your target (e.g., shooting into melee or attacking a grappled foe), roll again to see if you hit your ally.

• Tight spaces: If the area is too cramped for your weapon, you strike a wall and take a –4 penalty on your next attack with that weapon.

• Flails and chains: If you’re using a flail, roll again to see if you hit yourself (half damage).

• Fragile weapons: If you’re using a low-quality weapon or one unsuited to the target (e.g., a common blade against a stone creature), your weapon may break or lose its edge (–1 damage until repaired).

• Dangerous stunts: If you’re attempting a dangerous stunt, such as jumping form a higher point etc., you failed catastrophically. Fall prone, take damage, save for half.

• In all cases: You lose your footing, expose yourself, and take –2 AC until the start of your next turn, unless you spend an attack to regain balance.

---

By the way, this is the best rule I have for "shooting into melee". Being  a good archer improves your chance to hit an ally close to targe (an incompetent archer is more likely to miss both entirely), and this also takes into account your ally's armor.

Snakes & Ladders: The Problem of Linear Progression in D&D

Among games that remain popular to this day, one of the oldest in the world is called Snakes and Ladders, a game that practically everyone knows (see image below). Movement is essentially linear, from start to finish, but there are two types of shortcuts/detours: ladders, which accelerate your progress toward the goal, and snakes, which send you back a few spaces. It is a simple mechanic, but one that creates real tension between advancement and setback.

The Goal of D&D and Its Structural Problem

Without getting into the debate about what exactly the goal of D&D is, certainly one of the central purposes of its characters is to become more powerful, that is, to rise from level one, where they are basically beginners, all the way to becoming legendary heroes. Gygax himself (if I'm not mistaken) has stated that this was the goal of the game: to transform starting characters into heroes like Conan or Elric.

The problem is structural. In game terms, the route from level one to level twenty always moves in the same direction. You only gain XP, you only gain levels, you never lose them. It is a one-way street.

The older designers did not entirely ignore this problem. Hence the mechanic of undead creatures that drain levels, creating an effect similar to the snakes in the original game. The Dungeon Master's Guide also included situations that could temporarily reduce your levels, such as exhaustion after a grueling march, or situations of permanent loss such as aging, which can only be reversed with very powerful spells.

Of course, XP is not the only measure of a character's power. With experience, a character gains abilities and hit points, and some abilities can be lost temporarily (often for no more than a day). HP can also be lost, and in older D&D this could have more lasting consequences (perhaps weeks rather than days) than in modern versions. Finally, one can lose money (whether because raising the dead comes at some cost, or through theft, etc.).

It must be acknowledged that some RPGs, such as Call of Cthulhu, famous for its downward spiral into madness, and other horror RPGs in general (Unknown Armies, certain versions of Kult, etc.), exhibit quite clear and lasting negative effects in their systems.

In any case, aside from the issue of undead creatures draining levels, there are not many truly permanent losses in D&D. The only real loss available is the death of the character, which even then can easily be circumvented by a resurrection spell. In most versions of D&D, there is no universal mechanism for lasting negative consequences: the loss of an attribute, the amputation of a limb, a permanent scar (except as an optional rule in the DMG to save a PC reduced to zero HP, an option I find quite appealing). There are no significant setbacks, save death itself. Using the original metaphor: it would be as if every snake on the board led back to the first square.

There are other tools in the game that attempt to work around this limitation, such as creatures that destroy equipment (even magical weapons), that petrify characters, or that impose curses and other lasting effects. But these are isolated/specific solutions, not structural ones; they only happen with a very small percentage of creatures and situations.

Another problem is that D&D players, accustomed to constant and uninterrupted advancement, tend to dislike losing special items, levels, or anything they have already incorporated as part of their character. They may even come to see this kind of loss as a form of cheating.

In modern D&D, there is debate over whether characters can even die at all, so the line of advancement becomes, in the end, a conveyor belt with no way back: you can only move forward.

Source: Wikipedia.

The Parallel with Video Games

It is worth noting that older video games, also divided into "levels" (or stages), followed a logic quite similar to that of Snakes and Ladders. Reaching the end of a stage meant advancing to the next, but dying at any point sent you back to the beginning of that stage, not to the beginning of the entire game, and never (or almost never) to a previous stage. Alongside this, there was a system of permanent loss: a limited number of "lives" which, once exhausted, forced you back to the very beginning of the game (that is, "level one").

Some more modern and less linear games that generate enthusiasm among RPG players (such as Dark Souls, Bloodborne, and Darkest Dungeon) play with this question of real loss in more sophisticated ways. Death in Dark Souls does not send you back to level zero, but it frequently causes the loss of all accumulated XP. In the case of Darkest Dungeon, where you control groups of characters, the death of a hero is a permanent loss for the team, something no spell can undo.

Why This Matters

Even for those who prefer a narrative in which characters never die and always advance, that advancement would be more interesting if it were not purely linear, that is, a little more like Snakes and Ladders. In other words: on the road to heroism, you always lose something.

High-level characters would carry, in their character sheets and their stories, a series of losses, obstacles, and scars. This enriches both narrative and simulation, whatever the goal of the players at the table may be.

In more narrative games influenced by story games, setbacks make the narrative more interesting (and there are books and RPGs at least partially built around this concept of gain and loss, such as Hamlet's Hit Points, Fate, etc.). In games that aim to reproduce cinema, literature, or the hero's journey, an advancement without setbacks makes little sense either.

In games focused on challenge, obstacles become more complex and dangerous when there are real chances of loss. Likewise, in games that seek simulation or immersion, such dangers make the game more believable and realistic.

In short: from any angle, questioning (and perhaps overcoming) the unidirectionality that governs character advancement in D&D may be an experience not only valid, but necessary for the evolution of the hobby.

(Solutions? Probably the subject of a future post. However, I will say permanent wounds are a great idea, IMO, as they could happen in any combat not just against very specific monsters. This could take the form of scars, diminished abilities, etc. Alternatively, expand level loss to include other situations besides undead and marching, maybe as a general fatigue mechanic, so that wounded/tired MUs cannot cast their best spells and tried fighters attack a bit slower).

Additional reading:

https://methodsetmadness.blogspot.com/2024/12/nothing-to-lose-but-their-lives-stakes.html

https://methodsetmadness.blogspot.com/2023/09/ad-dmg-cover-to-cover-part-ix-pages-100.html

Minimalist (?) turn undead, plus a reflection on playtesting

Here is my minimalist version of turn undead, which precludes the need for a turn undead table, and uses 1d6 instead of 2d6 plus another 2d6 plus table:

Turn Undead: Clerics can repel or destroy undead. Turning is attempted once per turn, in lieu of an attack. Turned undead flee by any means available and will not harm or contact the cleric. To turn undead, roll 1d6, add the cleric's level, and subtract the target's HD (e.g., 2 for zombies). A result of 5+ succeeds; 10+ simply destroys them permanently. The roll result also indicates the total HD of undead affected (minimum one creature, maximum 20 HD affected). For example, if you roll 11 against zombies, 5 of them are destroyed; against skeletons, 11 are destroyed.

This is the type of rule I want for my game; maps reasonably close to the original B/X (at least to my liking), but a bit simpler, faster, leaner, easier (it also expands to RC levels).

(BTW I can take no credit for it as apparently Delta wrote something similar more than a decade ago; since I take much inspiration from his blog, I might have read it at some point).

In practice, however, I found that this is not enough for even the simplest games. If using this rule (or even the original B/X rules), the players will certainly ask simple questions like "how often can I turn?", "how far", "for how long", etc. It happened in my last campaign.

And the text simply doesn't say. The Rules Cyclopedia adds a much longer text (and table) - but not many answers either. Same in the AD&D PHB.

5e D&D, on the other hand, clearly answers all these questions (30-foot radius, 1 minute or until the creature takes damage, etc).

I'm probably adding such details to my own game since they were obviously needed at my table. So my version might even look a bit longer than B/X, which wasn't my original goal. 

Old school D&D seems to work very well in practice; people often say it is because Gygax etc. had immense wargaming playtesting experience. But I have a feeling that old school GMs often relied on their experience and rulings over having things spelled out in the book, which some people may appreciate but certainly brings endless problems when you don't have much experience with a system and need to learn from the book.

In other words, these games were likely playtested by people who were familiar with wargames, instead of given to newbies to see how understandable they were.

Modern D&D is much more complex (and even verbose and repetitive at times) but often better explained. And, to be honest, I don't think you can get "minimalist points" by omission and incompletion. If the book needs a "good GM" to work, it is not a great book, as most GMs are average by definition (or, at the very least, the book cannot take much credit for the rules if the GM has to create most of them).

Anyway, I keep looking for my ideal D&D - say, something as simple as B/X but as clear as modern D&D. This, I hope, is one step in that direction.

The critical hit "checklist"

As you might have noticed, I really like critical hits, but I dislike complexity. Playing systems such as Rolemaster with extensive crit and fumble tables, where a bad roll could mean your character trips over an invisible turtle (really!), was fun but also slow and frustrating.

This is another idea for critical hits. 

My aim is to pile all weapon complexity onto critical hits, so we don't have to write it down in every character sheet. It makes critical hits lengthy and detailed, but something you can easily change or ignore if you dislike it.

Maybe critical hits activate on a natural 20 (maybe 19-20 for certain fighters), or maybe on a margin of success of 10 or more. Whatever method you use, I think this table would simplify things. Each line can represent an idea that applies to lots of weapons, and you can just skip the lines you don't use.

You start with the idea that a critical hit not only gives you maximum damage, but — if it doesn't outright kill your target — gives you a percentage chance of doubling it.

The chances start at 5% for each point of difference between your d20 roll and your target number (e.g., with ascending AC, if you have a total of 23 against AC 13, you start with a 50% chance).

Then you go through a small checklist, skipping the items that do not apply:

• Using two hands gives you +10%.
• Armor (or tough/brittle bodies) gives swords and axes -10%.
• Shields give swords, spears and arrows -10%.
• No armor (or soft/flexible bodies) gives swords +20% and axes +10%.
• Lack of blood and functioning organs gives swords and spears -20%.
• Spears give you +10% otherwise (i.e., if there are blood/organs).
• Oozes give blunt and small weapons and missiles -20%.
• Large foes give blunt and small weapons and missiles -20%.
• Polearms are combinations of weapons, so it depends. Other weapons might fall somewhere in the sword (slashing) spear (piercing) or mace (blunt) categories. 

You get the idea. You can tweak the numbers, but potentially the entire "weapon versus armor" and "weapon versus large" tables could be included here — and since critical hits are rare, all this complexity only comes up occasionally, keeping the game fast the rest of the time.

When you get the percentage (if positive), you roll the chances of doubling your damage. If you roll doubles, you triple damage. Such a critical hit should always be described in detail, and someone killed in this manner will often suffer a gruesome death (decapitation, disembowelment, etc). If you miss the percentage roll, you still add +2 damage if you roll doubles.

If you like to use fumbles, a natural 1 could give you a fumbling chance. At least here they wouldn't apply to 5% of all attacks. Fumbling chances could be affected by circumstances such as:

Using a flail.

Shooting into melee.

Fighting with large weapons in small spaces or using bows indoors.

Uneven or unsafe ground.

Either way, the goal is the same: keep the complexity tucked away where it rarely surfaces, but it exists and always stays somewhere in the back of the players' heads, making weapons feel more grounded and detailed without slowing your game down.

Maximum Damage

We've been playing with a lot of ideas using D&D weapons and margins of success. Yesterday, a new one occurred to me: maximum damage.

Let's say weapon damage is determined by margin of success, but the maximum damage is unchanged from the original game. So a d4 becomes 4, a d8 becomes 8, etc.

This has several benefits before we get any deeper:

• You don't need a damage roll.
• Damage raises steadily with level, especially for fighters.
• Armor becomes even more significant.
• Armored duels
•  feel a bit slower and more realistic, with lots of wounds.
• Magic users could use swords etc. but it would usually not be worth the effort.
• We'd get more granular weapons (e.g., maximum damage 5, 7, or 9).
• Even a goblin can potentially survive a hit of a +3 sword, although this is very rare.

If we rule that a natural 20 doubles or triples maximum damage (or just raises it by 10 points or whatever), even a dagger can be deadly in the hands of a very skilled fighter against an unarmored foe.

Conversely, we can introduce laser guns with 15 or 20 maximum damage, and he stormtroopers would still have a hard time actually killing someone with it in a single shot.

I'm tempted to add your attack bonus to maximum damage too, so Conan can occasionally kill a sorcerer with a punch.

We'd have to consider how backstab works. Maybe a flat +4 bonus to attack and maximum damage.

Also, how do exact hits work? Maybe 0 damage, maybe 1, maybe some special effect, not sure.

One downside I can see is that average damage is a bit higher than usual even for low-HD creatures, although this is somewhat countered by the fact that in my games, PCs don't die at 0 HP. The fact that high-level warriors deal lots of damage is a plus for me, especially because I don't like having many magic weapons.

The fact that people get used to thinking about margins of success could also be useful if you use this for skills and other checks, which I like.

Another problem is that big foes such as giants and dragons could have their damage significantly diminished if we use this system exactly as written, never being able to actually reach their "maximum damage". Maybe that's a problem for another day, but I can imagine we could have a size multiplier of ×2 to ×4 (maybe "3x6" instead of "6d6"), which could create an interesting effect: a giant will not often hit the fighter, but when he does the damage is massive — one mistake might cost him dearly!

I think I originally had this idea years ago, when playing bell-curve systems, which might have different, but interesting, effects.

There are probably another issues I'm missing, but so far I really like this idea... let me know what you think in the comments!

Note: the GM Day's sale* is on, and most of my books are included! 40% off many titles! Here are some of my picks from past sales. Some deep discounts here, including the Dolmenwood Campaign Book looks really cool and it is 50% off! Maybe I"ll get it to take a look...

* Affiliate link.

How minimalist can D&D characters be?

As I've said before, this is the amount of information I'd like D&D characters to have — and that would have to be enough in actual play for something like 80% of the rolls, to minimize the time spent doing math and checking the book:

Nice, huh? Class, level, abilities, a couple of magic items or spells, and you're good to go. Most PCs have a little more than a dozen pieces of information (Name, Alignment, Level, Class, 6 abilities, AC, HP, weapons, and armor), plus spells for some.

Realistically, however, even the lightest versions of D&D need more information than that. For example, can you recall each saving throw from memory? Unlikely, but this is easily solved by reducing all of them to a single saving throw (say, roll 1d20 + level, target 20, or 16, etc.).

What about THAC0? Same thing. I'm happy with leaving the attack bonus equal to level for fighters, half level for everyone else, which is a huge simplification from D&D. But that's two extra bits of information. And usually, you need ranged and melee values, which rely on more information than just level (so you need to add strength modifier).

And ability scores? You have the six of them, but you need modifiers. You could commit the modifiers to memory, but you use them often enough that is is easier if you write them down. Well, maybe not all of them; since you already have AC, HP, and languages, you can ignore Dex, Con and Int modifiers most of the time (which is, by itself, an interesting idea - why keep these modifiers in the character sheet?). But you need Charisma mods for reaction, at least in theory, and Wisdom for saves.

Strength modifiers are needed to attack and damage - and in AD&D, this can mean two additional numbers. Notice these stats lack weapon damage too, something you'll use all the time.

Notice tat at the very least we could ignore all "+0" modifiers so we'd only need to add two or three digits, not six new ones.

We do not have much equipment here either; it is likely that a real PC has at least half a dozen items or more, not only weapons and armor. I'm counting "sword +2" as a single piece of information, not two.

[On a side note, maybe in a low magic D&D setting, "sword +1" could be a personal trait for a warrior instead of a magic weapon. This could incorporate your strength bonus and make "weapon specialization" a lot easier.]

So maybe we'd have a minimum of 30 pieces of information for each PC... but there is more!

Spellcasters have spells, which is straightforward enough (if not for the fact that they could in theory pick new ones every day, and clerics have access to their whole list - notice that the cleric here has no spells memorized). Thieves have skills - again, a bit hard to memorize, but can be easily replaced by rolling 1d20 and adding level (once you get some customization, more information is needed). Fighters have their weapons specializations and extra attacks - and they need this stuff.

And that is assuming each PC can only have one class.

In the end, we could have more than 100 pieces of information. Look at this AD&D sheet:

Of course, much of it is redundant, or rarely used, but it still muddles the sheet.

Sigh.

In the end, this post ended up doing the opposite of what I intended when I wrote its title...

The answer, I think is that D&D characters could be a lot more minimalist than they are, but it is not an easy task.

We could start by cutting all ability scores in half (only one number, no modifier), reducing all saves to one single save, and streamlining all skills... but I've been to this road before. There is no end to this, other than ending with something that doesn't resemble D&D anymore.

Maybe this much complexity is fine if I let the players handle it. 

I guess I have no easy answers today.

RPGs vs. Wargames - Zooming in and out

I have written about this before. Today I'm not discussing if RPGs are wargames or not. I'll just notice that there's a tension between the wargame and RPG perspectives, and will discuss how it manifests in Chainmail (wargame) and D&D (RPG).

The main difference is one of scope:

- Wargames are (traditionally, although there are exceptions) focused on battles between dozens to thousands of people; each player controls many people.

- In RPGs (traditionally, although there are exceptions), each player (except the GM) plays a single character/"role".

D&D was spawned from Chainmail (in OD&D, Chainmail appears as required material), among other games, and we can see the shift in focus as the game progresses. There is a "zooming in" of sorts.

Here are a few examples.

Alignment

Originally, alignment was about factions/teams. In modern D&D it is linked to personal philosophy, outlook, and behaviors. This shift creates some confusion and is discussed here.

Weapons

Chainmail had very important weapon versus armor rules that couldn't be ignored. Maces are better against plate, and daggers are a lot better against unarmored people. There are magic weapons, but not many details, you can get a bonus due to a generic "magic sword" or "magic arrow", for example.

When you get to D&D, the focus starts shifting to individual weapons. Magic weapons get more detailed (flaming swords, then vorpal swords, mace of disruption, etc.) and swords get deeper personalities and stats of their own. This tendency will continue through editions, with an ever-growing number of singular magic weapons.

The weapon versus armor table, on the other hand, is included in AD&D but often ignored in actual play (even by Gygax). In 2e, it gets simplified, and from 3e onward is nearly forgotten. The individual weapon is more important than weapon type. 

Ability scores and other stats

Chainmail does not use ability scores or many individual stats. Most creatures are defined by type/HD, AC, and attacks. In  OD&D, ability scores are present but not as important, and there are cases when Dexterity 7 is equal to Dexterity 14 in most circumstances; class and level are way more important. But soon "ability checks" become popular, and ability tables gain more detail; there is an effort to make each single point important. In modern (post-2000) D&D, ability scores are almost as important as class and level.

One interesting anecdote is how some classic D&D characters get names that are jokes/puns, simple anagrams of their player's names, or derived from class and level - the famous "Melf" is a "Male elf" abbreviation. Compare this to Drizzt, for example; a rebel drow that is opposite to whatever drow originally represented.

Hit points and level

In Chainmail, creatures are defeated or not with one or a few "hits". Hit points are created precisely because players get attached to their characters. This causes a "hit point inflation" eventually. Individual advancement, which doesn't quite exist in Chainmail, becomes an important focus of the game.

Dungeons and the battlefield

Wargames are often set in open spaces. Tight dungeons require a tighter focus. In old school D&D, this tension is often resolved by giving weapon range and movement different meaning indoors and outdoors (from feet to yards IIRC), which I find to be an elegant solution, but it later editions simplify things to make them equal despite the environment - often assuming that you're in tight environments anyway and even focusing on "grids" and "squares", especially in 4e.

Large battles

Large battles are assumed in Chainmail. In early D&D, the fighter gets some tools to fight hordes of weak creatures; this is expected. In modern D&D, this becomes somewhat of an special case. By 4e, you get "minion" rules to facilitate large battles. 

Individual monsters

In modern D&D, even lowly monster get endless variations, so that these creatures can also be individuals. The stat-block get bigger. By 3e, creatures have ability scores of their own. You also get more detailed rules on how to interact with them on an individual basis, maybe negotiating and so on. Most intelligent monster will have names, personalities and particular interests, which were not as relevant in old school D&D.

In conclusion

Wargames and RPGs are not necessarily incompatible, and some believe that RPGs are a subset of wargames. 

I do believe some "hybrid forms" or tools that allow you to "zoom in and out" are fun and will give you that "Appendix N" feel; Conan is sometimes in dungeons and single combat, and sometimes fighting or ruling over hordes and kingdoms.

Realizing there is a tension between the two perspectives may be useful to choose what rules to apply to your own games.

Additional reading:

https://methodsetmadness.blogspot.com/2024/02/wargames-storygames-and-rpgs.html

https://methodsetmadness.blogspot.com/2024/04/character-death-in-rpgs-war-games-and.html

https://methodsetmadness.blogspot.com/2023/06/weapon-versus-armor-from-ad-to-bx.html

https://methodsetmadness.blogspot.com/2017/03/does-d-require-miniatures-3e-versus-4e.html

https://methodsetmadness.blogspot.com/2025/03/brief-mass-combat-idea.html

https://methodsetmadness.blogspot.com/2015/03/on-alignment-part-i-alignment-origins.html

I want LESS!

Most of my time playing and reading RPGs has been ruled by the “undisciplined pursuit of more.”

I played several RPGs and took pride in trying new ones. I favored "universal" systems where I could create endlessly detailed PCs and do anything I wanted. I also collected monster manuals for my D&D-ish needs.

This process has been somewhat useful in helping me discover my tastes and needs.

However, there has also been much waste. I read only a fraction of the RPGs I buy and play an even smaller fraction of the RPGs I read. Nowadays, I'm embracing "the disciplined pursuit of less."

In other words, I want less: fewer monsters, fewer spells, fewer magic items, fewer stats, and simpler systems.

This partly reflects my appreciation for minimalism, but it goes beyond that. I believe that having fewer elements makes each one more important and meaningful.

While having a multitude of monsters is fun, each Monster Manual contains more creatures than entire worlds like Middle-earth, Barsoom, Lankhmar, or the Hyborian Age (not to mention most horror and sci-fi settings). It becomes virtually impossible for PCs to understand each individual monster with any depth.

Take dragons, for example. The dragons of mythology and literature are often unique individuals, like Smaug, Fafnir, Tiamat, Drogon. Each is distinct and memorable. But the 2e MM (my favorite!) has about more than a dozen dragon types. If a D&D party sees Smaug, it is just another red dragon (they don't even need to interact to know that he is chaotic, since he is red; but this is another issue).

[Another thing I've been considering is how adding more monsters to a game doesn't make it any different from "standard" D&D, but replacing existing monsters creates a completely new kind of setting. Take Curse of Strahd, for example: it includes few "demihumans," and even the elves are distinct from the familiar elves we're used to. This seems to hold true for most of my favorite settings and modules, and I think I might never run an adventure containing orcs again.]

Magic items are the same: Excalibur or Stormbringer are memorable, and so is Sting. In D&D, a first-level party often has dozens of magic items. Eventually, they discard some of them as they reach higher levels. This abundance devalues magic items and magic in general.

I feel the same way about rules.

I’ve run a few 5e campaigns. 5e is a more "complete" game than B/X. However, it requires ten times the page count to achieve this. So, I’ve been asking myself: is 5e ten times more complete than B/X? And the answer is no. Same goes for AD&D.

[Sure, I could use a one-page RPG. In the end, this ultimately comes down to a matter of taste.]

Spell selection has also been a headache, leading to imbalance and analysis paralysis.

I like customizing characters, but I don’t need dozens of classes and races. I really enjoy the simplicity of being able to say the bandit leader is a "fighter 5" and leaving it at that.

When you have fewer elements, you can connect them more meaningfully. For example, elves resist ghouls. The undead are raised by demons. All aberrations come from other dimensions, while monsters are created by mages. Etc.

And to be honest, this would make way more likely that my players would even REMEMBER most of this stuff.

In short, many of my current issues with D&D could be solved by just having less.

Additional reading:

https://methodsetmadness.blogspot.com/2024/05/unlimited-d-x-limited-d.html

https://methodsetmadness.blogspot.com/2018/01/the-illustrated-war-of-earth-and-sky.html

https://methodsetmadness.blogspot.com/2025/01/magic-item-inflation-in-ad.html

https://methodsetmadness.blogspot.com/2023/10/skeleton-with-bow.html

https://methodsetmadness.blogspot.com/2015/01/minimalism-elegance-and-multipurpose.html

Monster taxonomy and organization

It seems that the new 2025 Monster Manual organizes EACH  monster alphabetically. This means a "Green Dragon" is found under "G". In the 2014 Monster Manual (and most MMs before that), all dragons were found under "D", for dragon. The same happens with giants, demons, etc.

I think this is an awful decision.

Not that this is simple. One reason there's so much debate over monster classification (and issues like orcs being inherently evil) is that taxonomy itself is complex. In the real world, we classify living beings into categories like Domain, Kingdom, Phylum, Class, Order, Family, Genus, and Species. Can a green dragon interbreed with a red dragon? There is no "right" answer, but I'd guess they can. What about a gold dragon? Applying real taxonomy to fantasy creatures is not easy or ideal.

But let's look at this from a practical standpoint

Should we have one single entry for each species? That is impossible because we need many entries for humans (bandits, clerics, druids, etc.).

Besides real taxonomy, how can we organize monsters?

One alternative I really like is monster type. This is one of the main points of my Teratogenicon. Undead have LOTS in common to each other, and if you ever want to create your won, looking at existing undead is more useful than calculating CR.

There are other practical reasons to use monster type.

First, let's assume you are new to D&D, and you don't really know the difference between demons and devils. Or maybe you vaguely remember playing 2e and you don't even KNOW there are monsters called either of these things.

You go looking for an explanation in the MM. You turn to "D" and... there is nothing.

Can you see the problem?

On the other hand, let's say you're an experienced DM and you want to build your own dungeon! This is going to be a hellish cave, full of demons... Now let's find some o populate it! Where do you look for them? Again, you've got nothing. At best, the MM has a list of "fiends" that include demons, devils and others.

And what if - unimaginable though it is - you forget the name of a particular demon you once saw and want to use as the villain?

The only way this is useful is if you use strictly for reference. You never create your own adventures, but maybe you're running a module that lists "1d4 green dragons" on the encounter table and you have to check it in the MM (that is not great either; most adventures should provide you with the relevant stats to avoid page-flipping and book-flipping, but modern D&D is so crunchy that this is nearly impossible).

This might be a radical, but I think a good MM could be divided in 20ish chapters, including the 14 monster types with a few subdivisions. For example:

Aberration

Beast

(Giant beasts)

Celestial

Construct

(Golems)

Dragon

(True dragons)

Elemental

(True elementals)

Fey

Fiend

(Explaining differences between demons, devils, etc.)

Giant

(True giants)

Humanoid

(maybe separate species from professions)

Monstrosity

Ooze

Plant

Undead

(maybe corporeal/incorporeal)

To clarify, "true" dragons, elementals and giants have that word in their names: e.g., Green Dragon, Hill Giant. To make things even clearer, D&D could use different names for wyverns and trolls. For example, "draconians", "dragon-like", "draconic creatures", "gigantic humanoids" (notice that troll is a "giant" but "giant bat" is not).

Calling a wyvern a "dragon" makes the idea of "natural language" impossible, since you'd have to explain (or assume) the meaning of the word "dragon" every time you find a "dragon-slaying sword", etc.

There are a few obvious problems to this approach.

First, the monstrosities are so numerous that the alphabetical approach just feels easier. In addition, they are not always easy to separate from aberrations (gricks and grells - what are they?). In fact, when I wrote Teratogenicon I had to go back to 3e to find a good definition of aberrations.

Some subdivisions would need further reflection. Should dragons be listed alphabetically, or should chromatic dragons be separated from metallic? Not sure.

But, from a learning or world-building approach, this would be nearly perfect. 

It also gives the MM a more "in universe" feel. When an average peasant sees a "dragon", "green" is not the first thing that comes to mind. Similarly, a "death knight" is an undead first, and for the untrained eyes it is not that different than other ghosts or apparitions. People will just run and call this place "cursed"!

The "monster type" division, therefore, is also teleological.

In addition, you could easily create an alphabetical index of each creature for easy referencing (with page numbers, of course), and this list could include both "dragons" and "green dragons", under D and G. You could add page numbers to modules and encounter tables too, but maybe that'd be too much to ask...

In conclusion, I dislike the new organization. It makes it more difficult to find some monsters and put them into proper context. It makes the game less coherent and more difficult to learn. I will not say I have the perfect answer, but I can say I find the former approach preferable to the current mess.

Reflections on RAW, RTFM and game design

It is common knowledge that several rules are simply ignored in many RPG systems.

Instead of playing RAW (rules as written), people often play the game with several changes they have invented or found elsewhere.

AD&D is a good example - apparently, not even Gygax used all the rules that it proposed (most famously the weapon versus armor tables, maybe weapon speed). But this is true for a number RPGs, and it definitely includes the current version of D&D.

Some rules are ignored simply because they are BAD. But that's not what I'll discuss here.

Let's assume we have some GOOD rules that are ignored by many (maybe most) tables. We could even imagine that ignoring them will make a worse/more unbalanced game.

If your game breaks because of that... who is to blame?

Well, most people would say you are at fault. Especially if you are a "RAW purist" - someone who believes RPGs should be played exactly as written.

You should "Read The Fucking Manual" (RTFM), as people say.

I'm not so sure this is the case.

Let's try an analogy. 

A doctor orders you to take a medicine daily.

Many people will automatically say it is obvious that taking it is your responsibility.

But I can BET that if this is a pill to treat an advanced case of dementia, or it is a medicine in form of a big suppository for a mild disease, many people will simply skip the medicine.

And this is a DESIGN PROBLEM.

Likewise, if your games have rules that work in theory, but often get house-ruled in practice, maybe this could be a design issue.

Maybe the rules are too burdensome, fiddly, for anyone to actually use.

And yes, sometimes popularity is about quality - especially in this case. 

You already bought the book, and decide to play the game, so if a particular rule is often ignored, it probably means it is bad or too cumbersome, obscure, etc.

Maybe they tried the rule and didn't like it.

Maybe they didn't even try - partly because the designer hasn't been able to sell it in the manual. 

If people ignore an IMPORTANT rule, maybe part of the reason is that the designer failed to emphasize it enough.

Another example that occurred to me is buying my grandma a new air fryer.

At first, she was not sure how to use it. She does sometime struggle with the remote.

Fortunately, the manual is about 2-pages long, and buttons have been reduced to the minimum.

Good design is also about ease of use.

Maybe calling grandma stupid for not being able to use the remote and telling her to "RTFM" accomplishes nothing.

Maybe the remote COULD have a simpler design.

If you write a game, you should at least consider it.

AD&D DMG cover to cover - Part XI, p. 174-215 (Appendices C, D, E - Random monsters)

We've been reading the original DMG - the ultimate DM book! - but from a B/X and OSR point-of-view.

Check the other parts of this series here.

Today we discuss random monsters!

APPENDIX C: RANDOM MONSTER ENCOUNTERS

This includes extensive tables for random monsters. These are bigger, more detailed, and overall a bit better than the weird B/X / OD&D tables. Whales are not encountered in any kind of "water", but only in "deep water", etc.

"the only monsters which are included are those in MONSTER MANUAL. Two notable exceptions to this are those the mezzodaemon and nycadaemon which are found in the AD&D module D3, VAULT OF THE DROW (TSR Games, Inc.). If you do not have this module, simply ignore results calling for these monsters and roll again." 

A weird choice, but okay; the author found these two creatures important enough to be part of the core.

This section includes encounters in dungeons, outdoors, water, underwater, airborne, astral, ethereal, and also psionic encounters, whatever these are.

I'll admit this looks like it is too much for me. Underwater adventures are maybe 100 times less common than forests, at least in my campaigns.

First, there are random dungeon encounters. I do not think this is a great idea but the tables are detailed enough that they may help you create your own dungeon, with a proper theme and hopefully some coherence. 

There is a big focus on balance here; in theory, players can only find the strongest dangers if they travel deep enough (alike wilderness encounters, where they can suddenly face a couple of dragons). This has indeed some "mythic underworld" vibe, with little regard for naturalism/realism/etc: the deeper you go, the bigger and more numerable the monsters become. You can find a dozen bandits on level 1, but there is 120 of them if you find them on level 10.

You can also find adventurer NPCs, each extremely detailed, including random magic items. It is not clear how - and why - are these tables different from the ones in the appendix P.

The book recommends you prepare several parties/NPCs in advance. Looks like a lot of work, but fortunately we might have some tools like this one to make it automatic.

Underwater encounters are simple enough, but detailed - they are "are divided into those which occur in fresh water and those in salt water (seas and oceans). Each division is further broken down by depth - shallow and deep water encounters". Not much to comment here, and not much use for me as noticed above.

ASTRAL & ETHEREAL ENCOUNTERS are next. These are completely baffling to me. The explanation might be elsewhere; I might have read and forgot about them, or skipped it (probably they are the result of some spell?). The glossary indicates there is an explanation in the PHB, so its my fault for not reading it first. 

After some research, it seems these pertain to a certain spell, so maybe they should be include in that context. Like underwater encounters, I feel these won't be used often.

In any case, these are evocative and very interesting. It makes astral/ethereal travel feel dangerous and exciting.

PSIONIC ENCOUNTERS may happen if PCs are using psionic powers - these apparently can attract demons and other entities, which is sinister. These seem to manifest out of thin air (since the yellow mold doesn't move IIRC), so I'm not sure why the book suggests "Roll until an appropriate encounter occurs, ignoring inappropriate results" for this particular table only.

OUTDOOR RANDOM MONSTER ENCOUNTERS is the meat of the chapter. It has tables and subtables for Inhabited areas, Uninhabited areas, Castles, multiple types of terrain in various climates (artic, subartic, temperate, etc.), plus some setting conditions like "faerie", "Pleistocene" and "Prehistoric".

Well, Pleistocene is part of "Prehistoric", but here it means "Age of Dinosaurs", as indicated by the table. The book adds: "Feel free to devise your own encounter matrix for Jurassic, Triassic, or other period with non-aberrant creatures.". 

Why are there no mountains, hills river or seas in the age of dinosaurs? No idea. Probably it is the other way around: in D&D-land, you'll only find dinosaurs in these places.

Pleistocene conditions are somewhat to Sub-Arctic Conditions, without fantasy creatures such as trolls, etc.

I can IMAGINE the Pleistocene/prehistoric tables could be combined for a pulp/S&S campaign, but then you'd also need a river/sea table without nixies, hobgoblins and such. as written, maybe they are meant to apply to certain "lost world" parts of your settings - despite dinosaurs and mammoths being found in the regular tables too.

Putting everything together looks like a bit of a headache, but hopefully this too can be automated (I am sure there is an online roller somewhere, please let me know in the comments!). This one is based on 2e.

AIRBORNE RANDOM MONSTER ENCOUNTERS is short and sweet: "simply use the appropriate

OUTDOOR RANDOM MONSTER ENCOUNTERS table [...] but an encounter occurs only if the creature indicated is able to fly or is actually flying."

CITY/TOWN ENCOUNTERS are meant for unexplored cities, basically. They seem to happen incredibly often ("every three turns"), probably because you meet people all the time in a city, but many will simply ignore the party. 

Checking that often must be a bit of a hassle in practice; maybe we could just check a few times a day for "memorable" encounters that are likely to approach the party.

Also worth noticing that ordinary people seem to be a small percentage of encounters. I'd assume there are more, but unlikely to make memorable encounters. As written, these tables make cities extremely  dangerous, full of demons, undead, and bandits, maybe even more than the cities of S&S like Lankhmar.

BTW, this is where you can find the infamous "harlot table" that describes encounters with "brazen strumpets or haughty courtesans".

We also get ANOTHER table to generate magic items for NPCs, for reasons I cannot fathom.

APPENDIX D: RANDOM GENERATION OF CREATURES FROM THE LOWER PLANES

This is, basically, a generator of random demons, devils, etc.

I LOVE this chapter. This is a precursor to Teratogenicon and all similar books.

Basically, it makes each creature weird and unique, from head to toe, including stats. Here is one example created by this generator:

Demon #1

---------------------------------------------

Frequency: Uncommon

No. Appearing: 3

Armor Class: 0

Move: 15"

Hit Dice: 9

No. of Attacks: 3

Damage: 3-9 (Mouth), 2-12 (Each Arm), 

Special Attacks: Summon/Gate, Spell-like Abilities, 

Special Defenses: Acid Immunity, Weapon Immunity, Cold Immunity, 

Other Abilities: None

Magic Resistance: 45%

Intelligence: High

Size: L

Psionic Ability: Nil

Strength and To Hit/Damage Bonuses: 18 (00) (+3/+6)

---------------------------------------------

Appearance:

Head: Human-like  / Knobs

Overall Visage: Wrinkled - Seamed

Ears: None

Eye Color: Metallic

Eyes: Huge, Flat; Two-Eyed

Nose (If Necessary): Slits Only

Mouth: Tusked; Tiny

Bipedel Torso: Ape-like

General Characteristics: Short and Broad

Tail: None

Body Odor: Urine

Skin: Leathery/Leprous

Skin Color: Reddish

Back: Normal

Arms: 2Hands: Taloned

Legs and Feet (As Applicable): Suctioned

Pictured by Grok using data above.

This technique is great to keep things fresh and keep players guessing, although all fiends share some traits (e.g., magic resistance).

Teratogenicon extends this reasoning to other creature types: undead, aberrations, monstrosities, etc.

APPENDIX E: ALPHABETICAL MONSTER LISTING

A list of monsters and their stats. Probably based on the Monster Manual. No stats for the mezzodaemon, but more than 20 lines for hydras with varying number of heads.

Overall, these appendixes are good, despite some redundancies, weird choices, and mixing things of dissimilar importance without clear distinction, which seems to be a common trend in the DMG.

NOTE: there is a California Wildfire Relief Bundle on DTRPG. It has lots of Savage Worlds (including Savage Worlds Adventure Edition) and a couple of OSR games. "By This Axe I Hack!" and "There and Hack Again" are the most interesting to me.

Contains affiliate links. By purchasing stuff through affiliate links you're helping to support this blog.

AD&D DMG cover to cover - Part X, p. 169-173 (Appendices A, B - Random dungeons/wilderness)

The project is back after a long hyatus!

We've been reading the original DMG - the ultimate DM book! - but from a B/X and OSR point-of-view.

Check the other parts of this series here.

The meat of the book is finished; now we move on tho the appendices, and these are almost as important as the main text (in fact, there is often no clear reason for why something is in the appendix rather than another chapter).

Today we discuss random dungeons and wilderness!

Unfortunately, I'm not a big fan of those; I invite anyone who has more experience with them to share it with us in the comments!

Anyway.

APPENDIX A: RANDOM DUNGEON GENERATION

This part contains brief advice on how to make a dungeon and multiple tables to generate one randomly.

I am not a big fan of random dungeons; I find them nonsensical and often cliched.

I wrote about the subject here.

Last time I need a "dungeon" I took a castle map online and populated the rooms in an interesting, coherent manner - according to a theme (haunted castle) - and I was very satisfied with the results.

Can the Appendix A provide something more flavorful?

I'd have to try them in practice. I never actually did; I remember seeing some examples online, but nothing impressed me. Please let me know about any examples you have!

But let's see those tables.

There are 5 "start areas", which seem too weird and not sufficient. With the exception of number 4 (maybe?), they simply do not resemble any actual building, nor do they give an "mythic underworld" vibe (dungeons always start with stairs and they have lots of stairs - maybe they are "created by a mad mage" stuff).

In the same manner, the tables indicate long corridors and lots of 45º degrees passages.

In short, this produces a very specific type of dungeon, ideal for mapping on a square grid, but not much else. Your result will look like a "vanilla D&D dungeon", but not like a cave, castle, ruin or spaceship.

The tables that fill these rooms are a bit better; basically, they add monsters, treasures and traps. Enough variation to make things interesting.

Appendix I adds more interesting stuff to dungeons. Why not put it closer? No idea. But we'll get there!

(BTW: my own book Dark Fantasy Places has some ideas on these topics).

Curiously, the book indicates that "the random dungeon generation system is easily adaptable to solitary play", which is a play-style that became much more popular in recent years.

In short, maybe a cool mini-game, but doesn't seem to create great dungeons, unless the DM adds a lot of input.

EDIT: as waywardwayfarer note in the comments, there is an app inspired by the appendix A. Try it for yourself to see if this is the kind of game you want: https://www.blogofholding.com/dungeonrobber/

APPENDIX B: RANDOM WILDERNESS TERRAIN

This is similar to dungeon generation, but shorter. The first paragraph explains that:

If a wilderness expedition moves into an area where no detailed map has been prepared in advance, the random terrain determination system below can be utilized with relative ease for a 1 space = 1 mile, or larger, scale. In using it, however, common sense must prevail. For example, if the expedition is in the north country the forest will be pine or possibly scrub, while in tropical regions it will be jungle. Similarly, if a pond is indicated in two successive spaces, the two should be treated as one larger body of water. The Dungeon Master must also feel free to add to the random terrain as he sees fit in order to develop a reasonable configuration. In any event, the DM must draw in rivers, large lakes, seas, oceans, and islands as these features cannot easily be generated by a random method.

I don't think random wilderness is a good idea, for several reasons: first, it often produces incoherent maps. It would also take a HUGE time and effort unless you automate it - even a small area requires hundreds of rolls. In addition, mountains can be seem from many miles away and it'd be absurd for the PCs to suddenly find one.

You'd be better off just drawing your own map in advance without any help - which is FAST and EASY. 

It wrote extensively about this here.

The DMG seems to recognize these limitations and only suggest you use this "where no detailed map has been prepared in advance".

There is only a couple of tables here: terrains and inhabitation. Only 10% of hills (and ZERO percent of mountains) contain forests, which seems weird. The DM probably has too choose those, or he'll get hills with forest in the middle of the desert.

Again, my Dark Fantasy Places has some additional tables that might be useful.

If you use 1 hex = 1 mile, the map seems a bit crowded (one hex out of six has something, from single dwellings to cities of 10,000 people). 

Still, it does a decent job of balancing small hamlets, cities, castles and ruins. This is evocative stuff. Unlike mountains, you COULD conceivable find a small castle/ruins in the woods by accident, and it could be the beginning of a great adventure.

BTW, if you want LOTS of tables about the subject, you can check this post in Knights & Knaves Alehouse. Again, it looks like too much to tackle without automation but it is FULL of cool ideas!

Coming next... RANDOM ENCOUNTERS AND MONSTERS!

Contains affiliate links. By purchasing stuff through affiliate links you're helping to support this blog.

Dead-end mechanics

When I was analyzing the 1e DMG - a project I plan to complete in 2025 - I noticed it has several interesting sub-systems. 

While I prefer simpler games, I can see the value of having interesting detail to your games, no matter how idiosyncratic (for example, AD&D seems obsessed with polearms, while swords do not seem to get the same attention).

There was something bothering me about AD&D and it was not the bits that felt unnecessarily complex. The system felt a bit disjointed and I couldn't quite explain it until I gave it a name: dead-end mechanics.

This is not something specific to AD&D - it can potentially happen in any RPG, and it might be closely related to what makes RPGs unique.

AD&D is a good example only because it has so many moving parts.

Let's try a definition: dead-end mechanics are parts of an RPG system that do not meaningfully interact with other important parts, especially when those other parts are thematically related.

This definition can be improved, but let's give some examples.

Example 1. Disease, ears and hearing

There is a curious idea in AD&D that you can check monthly to see if the PCs suffer from parasites or mild ear disease. While this feels un-heroic and not particularly exciting, I can see that it would give a campaign a gritty/realistic feels and a sense of urgency (TIME must always have a COST).

[It can also add some gravitas to fights against giant bats, rats and even wolves].

However, the effects of such diseases on usual dungeon activities are not always described. There would be some obvious solutions - for example, diminished chance of listening to doors or greater chance of being surprised - but the book simply does not address this.

If you lose hearing in BOTH ears, the results on "listening to doors" are obvious, even if not described - we'll get to that later. In any case, the results on surprise are not clear.

Similarly, there is no exact consequences to the loss of an eye for ranged attacks, for example.

(Another curious example here is venereal disease. Since the game includes no benefits to intercourse, this just feels disjointed from everything else. Pendragon, for example, has rules for lust and descendants, which might suggest a bigger focus on such issues. Also, the game mentions herbs/gems that ward off disease, but not how they actually affect disease rolls).

Example 2. Time - initiative, segments, weapon speed

My main frustration here is how complicated and disjointed the initiative/segments rules feel.

There is simply a lack of obvious connection between surprise, weapon speed and thief skills. They don't seem to communicate... but they obviously should!

Can a thief surprise a foe with his silent movement? Should a fast weapon be ideal for this job? Can you get more attacks with a fast weapon if your foe is surprised? I'd say "yes" to all of those, but the book either doesn't make it clear or indicates that the answer is negative.

There is also also no clear connection between the speed of melee weapons and ranged weapons, and they seem to work differently in the surprise segments for no apparent reason.

Example 3. Levels, abilities and dungeon/wilderness skills

This is not an AD&D thing, but something common to most versions of TSR D&D. 

As the PCs level up, they get better at fighting and surviving, but they do not seem to improve in any other dungeon activity: listening to doors (or breaking them down), find their way in the wilderness, hunting, etc.

The thief is the main exception here, since most of his skills are dungeon related and get better with level. 

Curiously, the ranger does not have many special skills in the wilderness: he is not better than any other PC when finding his way in the forest. He can cause upraise and avoid being surprise in any environment, but it is unclear how this fits with the vaguely similar abilities of thieves and even halflings.

Example 4. Drowning

I don't remember the source of this; could be some version of Labyrinth Lord. 

But the rule was something like "if you try to cross a river in plate armor, you have 90% chance of drowning".

Just flat 90% (or whatever). Your strength, level or class do not matter. HP? Save versus death? No. You just drown.

Are dead end mechanics even possible?

While I find these mechanics undesirable, I do think there is a natural limit to dead end mechanics.

As I mentioned above, it is obvious that if you lose hearing in BOTH ears you cannot "listen to doors", although the game does not say that, nor does it describe what happens if you lose hearing in ONE ear, which is much more statically probable.

This has something to do with the uniqueness of RPGs.

RPGs give fluff and crunch a peculiar bond, to the point that fluff IS crunch and vice-versa.

This is a long discussion, but in short, in RPGs a spear will NEVER be identical to an axe, even if both deal have the same damage, weight and cost. An axe will ALWAYS be more useful to take down a door even if the game doesn't say so.

My point is: if taking down doors is a frequent activity, the game should address this difference explicitly.

How to write better mechanics

My ideal game would have a big level of integration between the different rules. In the most frequent cases, this should be explicit to make the GM's job easier.

Modern D&D sometimes does a better job at connecting various mechanics. For example, a Constitution saving throw relies on ability AND level AND class. But, sometimes, it creates MORE problems of this kind, like the fact that 5e D&D has at least TWO unrelated ways of disarming opponents (one of them optional), with no clear relation. 

I must mention Quidditch as a negative example (despite not being a big fan of Harry Potter). While there is some nuance, it often feels like two guys are playing an entire different game that has a flimsy relation to the rest of the players and a huge possibility to make all other efforts void.

["A Seeker catching the Snitch ends the game and scores the successful Seeker's team an additional 150 points (15 goals). As the team with the most points wins, this often guarantees victory for the successful Seeker's team." - source].

But maybe I can express this point visually, using the two images in this post.

The spheres represent game mechanics. The biggest ones are the most important/common. They are connected by lines; e.g., Constitution and level should both have direct lines to HP, but Constitution does not necessarily have a direct line to level or saving throws.

The first image in this post represents a disjointed game: no clear center, with some important mechanics disconnected from others.

The second image is closer to my idea: the most important spheres are near the center and strongly connected; disconnected mechanics are few and unimportant.

I think there is more to be said about the subject, but I'll leave it here for now.

As a suggestion, I'll say we must consider what are the central mechanics of a game (maybe abilities,  levels, classes, maybe also time, money, XP, encumbrance, etc.) and how they related to each other. Dead-end mechanics should be rare.