Overloading the Random Encounter Table

The perfect system for random encounters came to me in a dream. Often you’ll hear that as an insinuation that an idea was divinely inspired, a fated discovery from the muses, but perhaps it does happen. Perhaps the muses wanted me to know a dynamic but efficient random encounter method. Perhaps it was just a bit of undigested beef talking to me. Whatever the cause, I awoke in a cold sweat at 3am in order to write down this divine vision.

In Moldvay/Cook Basic & Expert (B/X) edition of the D&D rules, where so much clarity on running D&D stems to this day, a random encounter is a somewhat involved process. You need not only roll 1d6 to determine if an encounter is encountered, but you must also (1) roll to determine the type and number appearing of the monster encountered, (2) roll to determine the distance between said monster(s) and player characters, (3) roll to determine whether the party is surprised, and (4) make a reaction roll to see how the monsters react to the player characters’ presence. That is, frankly, too many discrete steps. Frequent readers of this blog know well that I desire nothing more than to cram as many results into a toss of dice as possible. And when my midnight muse visited me in my slumber, she told me exactly how.

Determining when and whether a random encounter occurs is a solved problem thanks to Necropraxis, my blogging idol. Necropraxis’ overloaded encounter die (and resulting hazard system) is a staple and heavily influential on my aforementioned desire to cram things into single rolls. In short, the referee still rolls 1d6 just like Mssrs. Moldvay and Cook prescribe, but now there are results beyond monster-or-not-monster, such as torches burning out or noticing monster tracks. In exploding the encounter die, I expand and contract this system for my needs: I cut the two results I was most lukewarm about (Locality and Exhaustion, this excellent post on the Doom Clock from my colleague at Illusory Sensorium detailed some of my issues with those results, along with issues with the entire institution of the overloaded encounter die), adding a potential rewarding result of quantum treasure, but also adding a push-your-luck aspect where the deeper you get into the dungeon, the deadlier the encounters will become but the sweeter the rewards. You get squeezed, but you get juice too. 

My exploded overloaded encounter die has been serving me well for a few years, but there is a glaring issue in that 2021 blogpost that I have now rectified. That is, it required 4 random encounter tables, each deadlier than the first. Even if each table is a d6 table, that is still too much prep, with 24 tables without even the benefit of bell curves. Bell curves are quite important to a good random encounter table, as any reader of Barkeep on the Borderlands, a random encounter table module if there ever were one, is well aware. The only use of a non-curved table that I thought seemed charming was the table for exploring civilized areas proposed by my colleague at Pointless Monument, and importantly it is paired with two other increasingly curvy tables. So in addition to consolidating the many steps for rolling up a random encounter into a single roll, I need a way to condense my multiple encounter tables into a single table that is capable of getting deadlier results as the player characters delve deeper. The demands upon my muse only grow. Will it live up to the hype I’m dousing it with?

Perfecting the Random Encounter Table

When the encounter die (overloaded, exploding, or otherwise) indicates that the players face a random encounter, roll 3d6. You will need an encounter table with results from 3 to 18, with results generally ascending in potential for peril. (Sidenote: this blogpost an excellent examination at the probabilities and potentials of a 3d6 roll.)

There is no need to roll separately for distance. If the roll includes doubles (two matching dice) or triples (all dice match), the monster is encountered near at hand. If the roll only includes singles (no dice match), the monster is encountered at a distance. I don’t specify the exact distance because what is reasonably will depend on where in the dungeon the creature is encountered, but as a rule of thumb a distant encounter will require at least a turn of movement before melee can be engaged, whereas a nearby encounter can be engaged in melee combat with less than a full turn of running full speed at the assailant. 

There is no need to roll separately for surprise. If the roll includes triples, then the players are surprised by the encounter. As stated above, triples are nearby encounters, so whether the monster drops down from the ceiling or sneaks up behind a player, the encounter is an in-your-face jumpscare. 

There is no need to roll separately for the monster’s reaction. Instead, evaluate the median die (the die result that is neither the highest or lowest result. If any dice match, the matching die will also be the median). That die result determines the encounter’s disposition toward the party. Below is the generic reaction table for this usage, but you could also develop monster-specific reaction tables, like Troika does with its “mein” tables, but in doing so, you will need to order the results with 1 being the least hostile and 6 being the most hostile.

1 — polite

2 — curious

3 — cautious 

4 — aggressive 

5 — hostile

6 — immediate attack

There is no need to roll separately for the number of monsters appearing. The result on the table should indicate how many are encountered. To the extent the number appearing should also be a variable amount, the dice already rolled could be use, e.g., for a small amount you could say the number appearing is equal to the lowest die rolled, for a medium amount you could say the number appearing is equal to the highest die rolled, for a larger amount it could be twice the highest die rolled, and for an amount beggaring belief, it could be triple the highest die rolled. The possibilities here are quite nearly endless. 

So a single roll tells you what monster, how many monsters, where they are, how they feel about the players, and whether the players are surprised by this all. But that is only half of the reason why I crouched over my phone harriedly writing down these notes in the wee hours of the morning. Not only can this roll cram in so much information, it can be dynamic for use with my aforementioned Exploding Encounter Die system or even just a less systematized approach where you want certain parts of the dungeon to be more dangerous than others.

When you want encounters to get more dangerous, roll an extra d6 and only reference the highest 3 results. Anyone who has rolled their stats using “4d6 drop the lowest” method for stat generation knows well how different the results are from your standard toss of the 3d6. Rolling 5d6 and dropping the two lowest amps this up even more. But the beauty, the elegance of this when applied to this overloaded random encounter table is not only that it makes higher results (and therefore more dangerous monsters) more likely, it also makes all results more likely to be hostile and increases the likely number appearing for encounters that aren’t solitary. For example, the odds of rolling a 12 increase when rolling 4d6-drop-lowest over 3d6, but also the odds that the 12 is a hostile 12 and not just cautious increases as well. This means you could use a single encounter table for multiple levels of a dungeon but cause the levels to vary in terms of danger simply by rolling differently on the deeper level. This method minimizes rolling and prep without much of a sacrifice. 

How to Stock Your Overloaded Random Encounter Table

Here is the catch: cramming so many results into a single toss of dice causes once-independent variables to become dependent. With the B/X method of rolling 5 times per random encounter, every encounter had a ⅙ chance of being a surprise. Here, only the following results will ever be a surprise: 3, 6, 9, 12, 15 and 18, with 3 and 18 always being surprises. Similarly, every monster was just as likely to be hostile as any other monster. Under this system, monsters with lower results are more likely to be less hostile and monsters with higher results are more likely to be hostile (although I don’t think of this as a bug). 

What this means is the referee using this system has to be less willy nilly with structuring their encounter table. Put monsters that are more likely to be friendly or at least curious toward adventurers toward the top and monsters that are more likely to eat adventurers toward the bottom. Put stealthy monsters on results divisible by 3. And although you want to strive to put more dangerous monsters lower on the table, if you don’t want that monster to sometimes be willing to negotiate with players, don’t make it a 17 or 18, as those are always fighting encounters.

The 3 and 18 results are the most overdetermined because both are ALWAYS a surprise (the only results for which that is true), 3 is ALWAYS polite and 18 is ALWAYS an immediate attack. This is because there is only one roll that can generate each of these (1,1,1 and 6,6,6 respectively). My solution is a simple one, and is inspired by my colleague Nick Whelan of the Papers Pencils blog’s advice (regarding 2d6 tables) that “2 is always a dragon, 12 is always a wizard”. Under this system, 3 is always a lost person from the surface world and 18 is always a demon. The reason for 18 (i.e., 6,6,6) being a demon is obvious from the dice being rolled, but happily fits with the prompt of “what is a dangerous, hostile monster that you are more likely to encounter the deeper you go?” The result for 3 is that it needed to be something polite and surprising. Coming across a lost villager or merchant in a dark and dreary dungeon is a surprise but a welcome one. They are likely in need of rescuing and will therefore always start out being polite to their potential saviors. (But note, in my more science fantasy setting of Prismatic Wasteland, the 1,1,1 result is a robot instead. Similar to the demon appearing on 666, the inspiration for the friendly robot on this result is due to this result being the only “binary” result)

When stocking your table, keep in mind the following meta-information about each result. The probabilities given below are based on a 3d6 roll, but note that it skews toward tougher and more hostile encounters when you move to 4d6-drop-lowest and 5d6-drop-lowest-2. Also, although I spell out all 16 possible permutations, I think a totally valid and lower prep version of this would lower the number of results to 8 by doubling up the numbers.

3d6	Reaction	Distance	Surprise
3 (0.46%)	100% Polite	100% Near	100% (+ Polite)
4 (1.39%	100% Polite	100% Near	0%
5 (2.78%)	50% Polite, 50% Curious	100% Near	0%
6 (4.63%)	33% Polite, 67% Curious	67% Near, 33% Far	33% (+ Curious)
7 (6.94%)	25% Polite, 50% Curious, 25% Cautious	75% Near, 25% Far	0%
8 (9.72%)	20% Polite, 40% Curious, 40% Cautious	60% Near, 40% Far	0%
9 (11.57%)	33% Curious, 50% Cautious, 17% Aggressive	50% Near, 50% Far	17% (+ Cautious)
10 (12.5%)	17% Curious, 50% Cautious, 33% Aggressive	50% Near, 50% Far	0%
11 (12.5%)	33% Cautious, 50% Aggressive, 17% Hostile	50% Near, 50% Far	0%
12 (11.57%)	17% Cautious, 50% Aggressive, 33% Hostile	50% Near, 50% Far	17% (+ Aggressive)
13 (9.72%)	40% Aggressive, 40% Hostile, 20% Attacks	60% Near, 40% Far	0%
14 (6.94%)	25% Aggressive, 50% Hostile, 25% Attacks	75% Near, 25% Far	0%
15 (4.63%)	67% Hostile, 33% Attacks	67% Near, 33% Far	33% (+ Hostile)
16 (2.78%)	50% Hostile, 50% Attacks	100% Near	0%
17 (1.39%)	100% Attacks	100% Near	0%
18 (0.46%)	100% Attacks	100% Near	100% (+ Attacks)

Example Overloaded Random Encounter Table

Here is a sample overloaded random encounter table for use with the method described in this post, taking into consideration the breakdown above:

3 — An escaped human prisoner from the nearby village

4 — An emaciated dwarf scrubbing orcish graffiti from the walls

5 — Forbidden orc and dwarf lovers fleeing from their families

6 — An animated dining room set restlessly rearranging itself (# appearing is highest die)

7 — A wise troglodyte admiring their treasure, a dwarf-crafted weapon

8 — Troglodytes gleefully dragging a fresh orc corpse (# appearing is highest die)

9 — An ooze that disguises itself as a puddle of putrid liquid

10 — An orc shaman looking for (consult lowest die): (1) someone who can teach her a new spell, (2) fresh, wriggling ingredients for a stew, (3) a worthy puppet to overthrow her former ogre-puppet who has become unruly

11 — Orc warriors that are (consult lowest die): (1) grumbling about their ogre of a boss, (2) carrying a wounded ally, (3) taking turns boasting over martial accomplishments (# appearing is double the median die)

12 — A sentient stalagmite-beast with tentacles, needle teeth and a single eye, that appears as a simple stalagmite

13 — An ogre warlord playing fetch with his pet dire wolves 

14 — A giant worm-beast whose face peels open like a banana to reveal sharp pincers

15 — A pile of bones knitting together to form the animated skeletons of orc warriors (# appearing is lowest die)

16 — A young black dragon, searching for materials to add to its burgeoning hoard

17 — The wraith of a dwarf mage wracked with crippling guilt and consumed with anger

18 — A demon that wants to possess the body of an innocent looking outsider so it can escape into the surface world and wreak havoc in a populated haven

Notes on the above table: I started with the 3 and 18 results since those were the most over-determined, following the advice I gave above. Next I filled in 6, 8, 12 and 15, the results that have chances of surprise. For each of these, there is a built-in ambush option. Because I was mostly taking encounters from a pre-existing published adventure (Forge of Fury), some of the results surprised me. For instance, the animated furniture is an auto-combat (of course, it is a 3e D&D adventure), but here the furniture will be either polite or curious when guests enter, which honestly makes more sense as a reaction for magical furniture (cf. Beauty and the Beast). After that, I just filled out the rest of the table. I generally just following the principle of putting deadlier and rarer encounters higher, more peaceful but rarer encounter lower, with middling difficulty but more common encounters in the middle. 10 and 11 are also examples of how to use “subtable” results without having to roll again, although you can really only get 3 sub-results this way (but 3 sub-results is plenty–I was greedy when I used 6 sub-results at times with Barkeep).

At 3:30am, I fell back asleep, and dreamt of demons (6,6,6) and damsels (1,1,1) in dungeons and dice, dice, dice.