I must create a system, or be enslaved by another man's. I will not reason and compare: my business is to create.

- William Blake

Friday, June 28, 2024

12 odd things about B/X wilderness encounters

While I was writing Basic Wilderness Encounters (now a silver bestseller!), I noticed lots of strange things in the B/X encounter tables I hadn't noticed before.

The B/X tables are inspired directly by OD&D tables. Notice that the OSE tables are identical.

In my book, I addressed/fixed some things to give the tables a more coherent feel, and I quite like the results.

Anyway, here is the list. Let me know if you have other items to add!


1. A dragon encounter is about 100 times more likely than finding a group of halflings. They are more common than wolves and hawks. Dragons are encountered in all environments regardless of color; i.e., a green dragon encounter is not more likely in forests despite the fact they live in forests.

2. Adventurer groups are incredibly diverse, considering how difficult is to find dwarves and other "demi-humans" in the tables. Any adventurer in a group of experts in 36% likely to be a "demi-human".

3. Expert adventurers, up to level 10-12, still adventure in the wilderness (despite many people thinking you should start "domain play" sat level 9 and stop adventuring.)

4. There are parties of high-level clerics, MUs, and fighters, but no thief parties (thieves only appear in more mixed parties, or as groups of "bandits" that apparently have no thief skills/talents). This might be because there were no thieves (as class) in the original D&D.

5. Mountains are extremely dangerous - at least 50% chance of encounter per day, and 25% of encounters are with dragon-like creatures (mostly dragons but also hydras, etc.), not to mention the possibility of a 36-HD giant Roc.

6. Swamps are dangerous too... Troglodytes are horrific; 2 HD, 5d8 appearing, camouflaged and always murderous. Desert and "barren" encounters are just as dangerous and almost as frequent (2-in-6, like forests), which I find odd, as desert fauna should be scarcer.

7. Swarms of common insects only appear swamps, but killer bees are everywhere.

8. The "unusual" table contains basilisks, which are more common than bears or ordinary hawks - or Halflings, etc.

9. In fact, there are NO ordinary bears in the tables, only cave bears and werebears.

10. Pirates (Chaotic, Morale 7) and Buccaneers (N, ML6) are nearly identical. They are also the largest groups; fleets can have more than 200 people in it.

11. You can find sharks in lakes and rivers, and even whales aren't uncommon. I replaced some with big alligators (of similar HD), that were sorely lacking.

12. Giant scorpions dwell in deserts according to the monster description. But they can't be found in the desert table. Only in swamps, jungles, forest, plains and settled lands.

Anyway, check Basic Wilderness Encounters if you want some reflections about these tables and 1000 entries to simplify you job as a DM. It is half the usual price if you get it in one of our bundles!

Monday, June 24, 2024

10:1 combat (B/X, Chainmail, and OD&D)

There are innumerable ways to do mass combat in D&D (I even tried to write my own). 

In old school editions, however, I am not sure it is even necessary to have a mass combat system.

On the contrary, I've been thinking a separate mass combat system might become a problem.

Take Chainmail, for example. Many people that play OD&D or AD&D like the idea of using CM for mass battlers, but I find this a bad idea.

In Chainmail, the mace is FIVE TIMES better than a regular sword against AC 2. In AD&D, since you can hit even negative AC with a 20, the mace is only slightly better than a sword.

This is the first thing that comes to mind, but there are many other differences - IIRC, a 4th-level fighter is much stronger in CM, for example, and using 2d6 will necessarily produce different results from a  d20.

My problem with this is that switching systems like that changes the assumptions about the game, to the point of changing the results of a fight depending oh the system you're using.

But what if you just use the same system with a different scale?

For example, say 100 knights are battling 150 berserkers. You could just run 10 knight "units" against 15 berserkers "units", as if they were individuals, and assume the results would be similar. 

I.e., a "10:1 scale"; one knight represents 10.

The rules would be exactly the same... with a few exceptions.


An unit tests morale upon losing one HP and again when losing half HP.

Notice that usually morale is checked when one combatant dies. If taken literally, this would indicate hat an army of 100 might flee is attacked by an army of 100 inflicting a single causality. I dislike that, and I only check morale when 10% of an army/unit is lost.

Interaction between scales

What happens when a unit of 10 knights attacks an ogre or even an exceptional PC?

Reverting back to the usual system is not a bad idea - it is easy enough to roll 10d20, etc.

Although I do think you should decide beforehand how many people can attack a single target at once (I like four, maybe twice as much for spears).

Likewise, the number of units attacking other units is limited and depends on geography.

When combat is resolved (because the ogre is slain or inflicts a causality and the knights fail a morale test), you can stay in 10:1 or 1:1 scale as appropriate.

Other scales/my experiences with mass battles

Of course, you can use other scales as appropriate. You could use 5:1, 100:1 or 1,000:1... The idea is finding a number of units that makes you comfortable. 

I have little experience running mass battles like that. But my PCs recently fought about 50 goblins in two or three waves, with the help of half a dozen NPCs, and it went very well. 

I once ran a 5e combat against 100 skeletons or so, which was also very easy as they came in groups of 4-5 thru the windows and the fighters would only miss them on natural 1s.

I think it would be hard to run more than 10 or 20 units at once, but within that limit I'd want to have as many unis as possible.

In addition, I'd have to consider unit types; if I have 1000 identical knights, I can use 100:1, but if I also had 50 archers, I'd definitely prefer 50:1. 

In any case, you can "zoom" back and forth as needed.

Well, for now, this is just brainstorming. Let's see if I can out it in practice.

Thursday, June 20, 2024

The Tombs of Atuan

The Tombs of Atuan is the second book in the Earthsea cycle. If you liked the first one, this is a decent sequel, if a bit slow and underwhelming. I've read it maybe a decade ago and revisited it last month.

The book is about Tenar, who gets taken from their parents as a child to become the new high priestess of the "Nameless Ones" - supposedly for being the reincarnation of the former priestess - in the Tombs of Atuan, a religious complex of a few buildings over a big, mysterious dungeon.

The first few chapters show a child trying to navigate the duties of a high-priestess, dealing with isolation, insecurity, friendship and jealously from other priestesses. Halfway through the book, a peculiar prisoner gets caught in the dungeon, and it is up to Tenar to decide what to do.

Like the first book, this is a coming-of-age story, adequate to young adults, although a bit simpler than the first. Tenar is taken against her own will, and while her decisions might be difficult for the character, they look a bit predictable for the reader. 

This book very deliberately avoid having any action. There is no swordplay, flashy magic and even the "monsters" are mere shadows. Unlike the first book, there is not much travelling either. Instead, the focus is on the characters, their dilemmas and feelings.

For D&D players, the book might be worth the read for the portrayal of the labyrinthine dungeon and the process of navigating it in the dark.

The book feels very true to the first book in the sense that the themes follow naturally. Ged was looking for a name in the first one, and here he has to face the Nameless (hinted in the first book). Like Tenar, he was taken as child, but they had different mentors. Ultimately, both have to learn responsibility in order to grow.

It is another short, easy-to-read book, that I'd recommend if you like the first one - even if I liked the first one much better. In any case, I enjoyed it and proceeded to (re)read the third book in the series - which I'll review soon.

Wednesday, June 19, 2024

Sandbox detour

Yesterday, something happened in my game... that might be worth discussing briefly.

The PCs in my sandbox were hunting a few goblin tribes. I had the entire area (hexes) and a couple of goblins caves prepared in advance (adapted from a published module).

But when the game started (right after they had slain one of the goblin tribes - there were two more to go), the PCs suddenly decided to abandon the quest and traveled to the nearby haunted ruins, which they thought would be more profitable.

I had already decided on a module for those ruins... but I hadn't read it.

They had two or three nearby places to go (in addition to goblin territory). The thing is, I'm unable to memorize them all.

[Also, notice, it is USELESS to memorize them; now, the goblin "plotline" became partially obsolete because they will NOT be able to rescue certain hostages anymore, which changes the whole thing... there is little use in preparing too much for things you'll never face].

Sometimes I just need a dungeon that I can read while I'm running it. Or, even better, a complete sandbox like Qelong or Curse of Strahd - it allows me to familiarize with the whole setting at once.

The ideal sandbox would have a significant number of "IFs" so I didn't have to come up with my own. "If the PCs refuse, the hostages die", etc.

[Notice Qelong is surrounded by mountains and Ravenloft by mists; an island would work too. It seems the ideal sandbox is somewhat limited: "we are playing in this area, if the PCs leave the adventure will turn into something else, which requires some time to prepare"].

The game turned out well, although it felt clunky reading and refereeing at the same time. I made a few mistakes (e.g., said a door was open when it was locked) and it took a bit longer to describe each room.

Fortunately, it was a short session, and I had my Basic Wilderness Encounters with encounters for EVERY biome in the setting, which made the wilderness part a breeze.

I will familiarize with the dungeon better for next week, since they are unlikely to change course soon, so things will be smoother.

What I DIDN'T want to do is "improvise" in any way.

I had two goblin lairs ready to go. Should I have used THAT maps for the ruins? 

NO! I want my setting to be a real place, not a Schrödinger's simulation.

Could I have rolled a new map randomly?

No, that would be equally bad. 

See, they had two nearby dungeons to go, each with its own "backstory". If I generated them randomly as they explore it, their choice would simply not matter.

Anyway, it is fun to have a sandbox and allow them to go anywhere. One of the players started asking, "wait, why don't we BUILD something?".

I have no plans for this type of campaign. Might be fun, I don't know.

One thing I do know is that the players can surprise me - and this makes running the game way more fun for me.

Monday, June 17, 2024

Milestones WITH XP

A quick thought exercise on the subject...

The two most popular (and somewhat diverging) ways of leveling characters are milestones and XP.

I assume you know how XP works; it is assigned according to treasure acquired, monsters defeated, etc. If you get enough X, you level up.

Milestones allow you to level up upon the completion for a particular goal: saving a hostage, slaying a particularly important creature, surviving a dungeon, etc.

Most people prefer using one way or the other, but it is easy to see you could combine them: for example, the GM may arbitrarily assign XP rewards for certain goals.

I like using milestones myself, because it allows me for a bit more control over the pacing of a campaign,  and also because it requires less bookkeeping. 

But I disregard XP entirely: PCs have levels and they level up when appropriate (usually upon finishing a module or two). 

In order to do that, I had to balance the classes using feats and a house rules.

This is also the case for my free "OSR Minimalist" game. I always enjoy getting feedback on that!

But what if you WANT to keep distinct XP costs and STILL use milestones to avoid writing down every coin acquired and every monster slain?

There are a few ways to do that.

The simplest way is probably choosing a PC at random to level up, but every PC has the same XP. 

If you have 4 PCs, all on 1st level, roll 1d4. It the thief is selected ts to level 2 (1,200 XP), no one else levels up. But if the elf is selected (4,000 XP), everyone else levels up too... and the thief gets to level 3 immediately!

As you can see, some nuance is lost here - some levels will be skipped.

Alternatively, just use 500 XP per milestone until 10.000 XP, then 5.000 until 100.000. 

This will give the fighter a level for each 4 milestones, more or less.

After 100.000 XP, you can give 20.000 to 50.000 XP per milestone, depending on how fast you want PCs to level up to "name level" and beyond.

Anyway, this is purely theoretical - I think I'd prefer to use XP or milestones as a primary method and maybe some exceptions.

There is probably a better solution out there... let me know in the comments!

BTW: my latest book, Basic Wilderness Encounters, is now added to my discount bundles. You can buy ALL my books for around 30 bucks!

Thursday, June 13, 2024

The importance of cleaving mooks

I recently wrote a couple of posts about how B/X Fighters are too weak and how to fix them.

There is one thing I want to elaborate further: "cleaving" (or sweep attacks, etc.). 

As it often happens, there is a great post in Delta's blog about the subject, which covers most of the topic.

This is an ability the fighter had in OD&D and AD&D, that is missing from B/X. Here is how OSRIC describes it:

"Fighting the Unskilled: When the fighter is attacking creatures with less than a full hit die (i.e. less than 1d8 hit points), the fighter receives one attack for each of his or her levels of experience, e.g. a 4th level fighter attacking goblins would receive 4 attacks per round."

Delta's describes the importance of this rule when fighting hundreds of goblins. Notice that the number appearing for goblins is:

- In B/X, 2d4 in the dungeon and 6d10 in the wilderness. 
- In AD&D, it is 40d10.

This rule is useful not only to give the (very) high level fighter a chance against a goblin horde, but also to balance the power of fireballs (and similar spells) somewhat.

There is also a narrative importance to this rule that is worth mentioning.

In fiction, the "main villains" are often surrounded by low level "mooks" that must be defeated first by the protagonist. In RPGs, I've seem PCs ignore all mooks over and over again to attack the villain directly, as if the main villain was all that mattered. 

Which, fair enough, can be a good tactic, especially in the absence of "opportunity attacks" and the like.

However, it makes me wonder if the reason is not the lack of a cleaving rules, which would often encourage the fighter to make 10 attacks against the goblins instead of going directly for the goblin king/captain - which is a bit less likely to die in a single blow and usually deals significantly less damage than 10 goblins combined.

If you decide that the death of a leader will cause some penalty to morale, this introduces and interesting tactical choice.

Fighting hordes of mooks at once is also something that Appendix N characters like Conan and Elric can do.

In addition, let's remind that hacking down hordes of ores feels awesome for the fighter.

Amazing art by Dean Spencer.

My main issue with "Fighting the Unskilled" is that it lacks nuance. A 10th-level fighters has 10 attacks against goblins, but only one or two against orcs. 

Notice that orcs also appear in greater numbers in AD&D than B/X, but are not subject to these extra attacks.

Another difficulty is how these "extra attacks" combine with multiple attacks. If a 7th level fighter has 3/2 attacks, how many attacks does he get against goblins? Seven? 8/7? 

Weapon proficiency confuses things further.

This is why I like the "cleave" solution, from 3e, as explained in Delta's blog:

3E D&D

CLEAVE  [General] You can follow through with powerful blows. Prerequisites: Str 13+, Power Attack. Benefit: If you deal a creature enough damage to make it drop (typically by dropping it to below 0 hit points, killing it, etc.), you get an immediate, extra melee attack against another creature in the immediate vicinity... (PHB p. 80)

You can wield a melee weapon with such power that you can strike multiple times when you fell your foes. Prerequisites: Str 13+, Power Attack, Cleave, base attack bonus +4 or higher. Benefit: As Cleave, except that you have no limit to the number of times you can use it per round. (PHB p. 82)

3E D&D introduced the concept of "Feats": special powers that may be chosen as characters advance in level. Fighters get additional, bonus Feats (more than any other class), and the two listed above are on their Bonus Feat applicable list. That said, not every Fighter gets the power; they must make a deliberate choice to pick up the ability. At the earliest, a Fighter might have Cleave at 1st level, and Great Cleave by (you guessed it) 4th level.

Most of us interpret this as a reworking of the rule from earlier editions; if a Fighter (with the Feat) battles very weak creatures, then they're likely to get a chain of attacks that puts many of them down. And many of us prefer the continuity of this mechanic -- unlike in 1E, where there's a huge quantum collapse between fighting "up-to-7-hp" creatures vs. "up-to-8-hp" creatures, the benefit here will more smoothly be usable against 2 HD or 3 HD creatures, just less frequently.

(Of course, I added similar feats to Old School Feats and I use them in my games.)

I'll add that this kind of cleave works smoothly with multiple attacks, magic weapons, weapon specialties, etc: the higher your attack and damage, the higher the chance to use this.

[One small aside: one thing I haven't considered in this post or the one before this is the 5e cleave method of "damage overflow": if you reduce an enemy to 0 HP, any excess damage is dealt to a nearby foe of same AC or lower. This is also a very interesting solution because it is simple, fast as smooth, while it also gives meaning to high damage rolls against goblins, etc.]

In short, if I were to add ONE single ability to B/X fighters, it would be this "great cleave". 

I do not think it would be enough, mind you - I still think they deserve extra attack to improve damage output against solo monsters  - but it would be a great start.

Monday, June 10, 2024

Fixing B/X fighters

Now I'll analyse some solutions - these are already implemented in AD&D, BECMI, and other games, but I'll discuss some and show which ones I use in B/X.

Better numbers

Easiest way of improving Fighters without making them more complex is just raising their numbers. E.g., giving them 1d10 HP or a better attack progression (AD&D), better saves, etc. 

This is a matter of fine tuning and taste; I like just giving fighters +1 per level as it simplifies things and it's just intuitive to me that a level 9 fighter has +9 to-hit (in my games, a 9HD monster also has +9). 

I'm not a fan of 1d10 HP because it leads to some HP inflation, but it is not bad either.

More damage

This is also "better numbers", but deserve special attention because, unlike HP and THAC0, damage doesn't usually raise with level (unless you use magic items).

A fighter that gets extra attacks increases damage output, but I always felt that Conan should be able to kill a sorcerer with a single blow occasionally.

We could give the fighter a damage bonus (level/3, round down?). 

Or just say that if you hit AC by 10 or more, damage is maximized or doubled. This will improve damage gradually with attack bonus, encourage the use of bigger weapons, give more importance to armor, and also give the thief a boos because of backstabbing (which is well deserved).

Better items

This is not exactly a solution, but a concern. Magic items in B/X are extremely important to Fighters, but not as common in my own games (which have a lowish fantasy, S&S feel). 

This is my own fault, of course - but I thought I should mention this here so you don't fall in the same trap. 

In other words, if you are limiting magic items (especially the intelligent swords of B/X), give the fighters something else to compensate.

Fighting minions and "cleaving"

The OD&D and fighter can attack multiple times against a horde of foes with less than one HD (e.g., goblins). I find this so important that I will write a small post on the subject next. 

My main objection to this is that it is binary; maybe the fighter can attack ten times against an horde of goblins but only once against a bunch of orcs.

"Cleaving" is a great substitution: when you reduce an enemy to 0 HP, you get an extra attack. This will often allow a high-level fighter to attack multiple goblins and even get some advantage against a group of reptile man. No calculations needed and it just feels cool.

Weapon specialization

Weapon specialization as described in AD&D and BECMI cover a lot of these grounds: better numbers, more damage, more attacks, etc. Which is good and also allows fighters to be archers, duelists, etc., giving players more archetypes to play with.

My issue with making weapon specialization a focus is that fighter often rely in magic weapons they've found, so they cannot always choose which weapon they'll use, and it is sometimes disappointing to find a cool magic axe that deals less damage in your hand than an ordinary sword.

My solution is that a "weapon master" gets a bonus with ALL weapons, and a SMALL bonus to a single fighting style, as suggested here:

Weapon master. You get +1 to attack, damage, and AC. Choose a type of weapon or a fighting style (blades, two-handed weapons, light weapons, missile weapons, dual wielding, sword and shield, grappling, etc.). When using this style, you get +2 to attack, damage, or AC (choose one when you pick this feat), instead of the usual +1.

Weapon maneuvers

BECMI has some cool weapon maneuvers (mostly starting at 9th level for fighters and mystics). 

The one I remember the most is smash: you get -5 to hit but add your entire Strength score to damage! From the Rules Cyclopedia:

For example, a Strength 17 fighter ( + 2 to attack and damage) using a sword +2 ( + 2 to attack, 1d8 + 2 damage) would perform a smash this way: He rolls to hit with a net penalty of -1 ( + 2 + 2-5). If he hits, he rolls 1d8 + 21 (17+ 2+ 2) for damage!

1d8+21 damage?!? With this, Conan can DEFINITELY kill a sorcerer with a single blow!

The only problem with this maneuver is that it is so obviously overpowered for strong fighters that you will never make a regular attack again!

Notice this was partly incorporated in 5e as a feat: a simple -5 to-hit and +10 damage. Other OSR games have their own rules for special maneuvers (more notably DCC, LotFP, and Low Fantasy Gaming).

Multiple attacks

This is the most notable and popular of fighter enhancements, present in almost every edition other than B/X. It enhances damage, number of targets, options of maneuvers (maybe you can trip AND bash).

One small issue is that the second attack DOUBLES damage output, which is an abrupt boost (in 5e, for example, a 3rd-level fighter has some hope against a 4th level one, but neither can beat a 5th level fighter).

The third attack enhances damage by 50% so it is not as dramatic.

AD&D has "one and a half" attacks, which is a decent solution if you prefer less dramatic changes. The second attack comes every other round, but if you don't like this solution make the second attack deal half damage, or only take place when you roll an odd number for the first, etc.

My favorites - to use with B/X

I'm running a game that is mostly inspired by B/X, although with many house rules

Here are some of my favorites, for fighters:

* +1 THAC0 per level.
Feats that include cleaving, multiple attacks, weapon specialization, etc. - the fighter gets more feats than other classes.
* Simple weapon maneuvers. Usually, -4 to-hit and choose a benefit. Examples: +4 to damage, or AC (against a single foe), or to an ally's AC, or to cause an effect in addition to damage (e.g., trip, disarm, etc. - the enemy usually gets a save). But it doesn't come up often.
* If we get to level 10, I'm giving fighters 10d8 HP and so on (explained here). If a 12th level magic-user causes 12d6 with a fireball, I think it is fair to give the fighter 12d8 HP.
* Strength affects how much you can carry.
* Natural 20 deals maximum damage.

One thing I haven't used in the current campaign is "cleave" (as no one picked the feat), but I'm feeling I should have given this to everybody. This is what my next post is about.

Saturday, June 08, 2024

Are B/X fighters too weak?

Yes they are.

Before comparing them to other classes, let's take a look at other versions of D&D.

In AD&D, fighters have several abilities that are ignored in B/X:

- 1d10 HD.
- Better attack progression.
- Better types of ordinary armor.
- A few perks if they have high abilities (more HP, damage, etc.)
- Extra attacks as you level up.
- Extra attacks against low-HD foes.

AD&D came before B/X, but BECMI came soon after.

Arguably, BECMI "fixed" many things in B/X I perceive as flaws (e.g., the cleric), and it also added several toys for fighters: weapon maneuvers, proficiency, etc.

All other editions (2e, 3e, 4e, 5e, and even 0e) have fighters that are stronger and more varied than B/X. Even some basic-inspired games such as DCC and LotFP make fighters more "special".

The fighter is DISAPPOINTED!

Now lets quickly compare a B/X fighter with other classes.

- A cleric with 50,000 XP can raise the dead once a day. He gets the equivalent of "THAC0 17" and 7d6 HP.
- A high level cleric might have fewer HP than a fighter, but he can heal himself compeltely in three  or four days of rest, while the fighter may take weeks.
- A fighter with the same XP has the same THAC0, 6d8 HP, and worse saves. His THAC0 improves significantly at 64,000 XP. Which is nothing compared to raising the dead.
- If both reach maximum level, their THAC0 is nearly identical and fighter saves are slightly worse. The cleric can now raise dead four times a day, AND requires fewer XP to get there.

- A magic user is frail when compared to a fighter, but has lots of firepower. A 1st level MU can put multiple fighters to sleep; a 10th-level MU can memorize three 10d6 fireballs, or even better-  summon a 16 HD fire elemental, which will give ANY single fighter a difficult fight. If he wins initiative, he cannot be stopped.
- Even with an awesome magic sword, the fighter cannot deal this kind of damage.

- Dwarves stop their careers at level 12... ONE SINGLE HP short of a 14th-level fighter, and with better saves, fewer XP needed, plus infravision. They have no restrictions on using swords. Dwarves are simply better fighters that require 10% more XP.

The thief is debatable - he often hits as hard as a fighter of similar XP, but is much frailer. I do not think the thief is a strong class in B/X (and it is even worse in BECMI). Halfings are not great either because their level cap is a significant limit.

Elves are more powerful than fighters because of magic, but slightly frailer in HP and saves. Overall, I find them much stronger than fighters.

How to fix that?

Well, that's the subject of the next post (I use feats, mostly; you might as well use the AD&D perks described above).

This one is about illustrating how the B/X fighter is weak compared to other fighters (0e, BECMI, 1e, etc.) and other B/X classes. 

Thursday, June 06, 2024

Author x Referee GMs

The GM is BOTH an author AND a referee.

The author creates a setting, adventures, NPCs, etc. - even rules.

The referee enforces the setting and rules.

As you can see, GMs do both - but these roles are somewhat incompatible.

How can you be expected to be the impartial judge with rules you create yourself?

One method many GMs use to deal with this is separating the functions in two distinct moments:

- I'm an author during "prep", when I create worlds, dungeons, NPCs, or even random tables to sue during play.

- I'm a referee during play. I stick to the rules, respect the rolls, and do not changing the HP of an important NPC during a combat.

But, since you are expected to play both roles, some GMs prefer do to BOTH AT ONCE.

This is what "improv" is about. You cannot improv as a referee - when you do that, you are in author stance.

To each their own - but I find that confusing.

For example, an author can definitely "fudge" the dice. If he can create a dragon (or nation) out of nowhere, deciding a single hit missed is well within his powers.

This solves a lot of problems.

For example, the classic "quantum ogre". There are two doors. The PCs choose the door on the right. Can the GM decide there is an ogre behind it AFTER the PCs choose?

Well, author-GM can, but referee-GM cannot.

Likewise, is changing the result of a random encounter roll adequate? It is for the author-GM, but not for the referee-GM.

And what about "balancing" encounters to suit the level of the PCs? Same thing.

Come to think of it, this separation is the reason why I avoid "improv". 

I want to use the author stance as little as possible during the game.

Maybe that is also why I feel the need to write (and sometimes publish) my own material, despite my constant urge to tinker and house-rule.

[Basic Wilderness Encounters was created precisely because of this reason. I didn't want to "author' encounters during the game. I wrote it is an author, and I use it as a referee].

And why I prefer to run other peoples' modules - and even campaigns - rather than creating my own.

To be clear, I LIKE to be an author-GM, just not during the game.

Anyway, I cannot be the first to have this thought, but I think the distinction should be more popular, and probably more discussed in game master's books.

There are probably more things to tackle around this subject. 

For example, I am thinking players have a specific stance during the game (they are almost NEVER authors in my games), but it is fine to give players author role between sessions, while writing a backstory, or when scheduling a game (e.g., agreeing in advance WHICH dungeon the PCs will tackle next).

But I think this is enough for today.