Make your 2d6 Random Encounter Tables WAY BETTER!

This is amazing! I have a nice complication table in the game Stravagante! that is a d6 table with a nested d6 table for each number rolled. I also can use GMer' or Mythic's Chaos Rank to help with the difficulty raising, so when the tension raises in the adventure tougher encounters happen more often. I could probably use the Chaos Rank to influence the Number Appearing too! The possibilities!

These are great ideas and enough to cause fun sessions on their own. Thank you

One point I would make on the 2d systems. The ones with very dissimilar dice (d20+d8 for instance) do give a very flat region in the middle and only taper at the ends. For instance, the d20+d8 would have equal probabilities for the numbers 9 through 21 ... so 13 options in the middle that are all equal, and then some rare things at each end.

I'm wondering if using a deck of cards might not be a good idea ... each number could represent a different encounter and the suit might influence or determine something about the state or disposition of the encounter (♥ could be friendly, ♦might have extra loot, ♣could be an encampment or lair, ♠could be outright hostile). One of the advantages of this would be that once you have an encounter, it's in the discard pile and reduces the chances of encountering it again. Even if you do, the state would be different.

The mathematician in me has to point out that there is a difference between "truly random" and "random with equal probability"

Are you wearing a full suit or are you rocking the short? jorts even?

Why not use a simple matrix. Using a d6 for the header row with types of encounters and then a d6, 8, 10 for the encounters under any heading? The header row can be two, three, and on to six types. This allows weighting the types of encounters easily, or eliminating a type. Let's say you have a type as random encounters the forward the overall adventure with rumors or future hooks, and it could be eliminated if it will no longer be useful. Just more options with a matrix. You can have two players role dice to find the encounter and they will burn valuable brain processing time decode the table (should this be an added bit of gameplay with your group).

So what happens when, with so few encounter options, you roll the highest difficulty two times in a row right from the start? It seems like the bell curve problem could just be solved with more values total rather than ditching the bell curve altogether. If organized in difficulty, a best result of two rolls or worst result of two rolls could be used instead of adding values.

Why is “truly random” more desirable? A bell curve is used specifically to be less random so as to make the overall game more cohesive. If I’m in a goblin infested cave, I would have more chance of running into a goblinoid, put them at 7,8,9. I’d be less likely to run into another adventuring party, and maybe even less likely to run into a ghost, but I still want it to be a possibility…
Also, as a DM you can update the table, re-roll doubled results, or move the ghost to a more likely position if the party starts seeing more ghost related activity based on other table rolls….
This video seems to point out possible problems with one quick way of doing something and claiming that way is bad, but you could easily have a response to this way pointing out the problems with this quick way to do something… the trick is to not pretend that one quick way is the best way, but to learn to improvise and adjust on the fly.

Excellent as always. Many thanks.

I think the majority of this is really good. I just fundamentally disagree with the idea that weighted probabilities != random. I think that using mixed dice creates a better distribution that isn't as heavily dominated by the center. I would suggest 1d4 + 1d8 to still have rare encounters but start to flatten out the curve a bit. Or even better start treating 2's as 3's and 12's as 11's. This beefs up the chance of those rare encounters without just removing them.

Personally having rare unlikely encouters makes them more special. The 3rd time the party comes across a wyvern it is just as boring as the goblins. The fact that it is rare is what makes it special.

cool ideas

I do appreciate having a Disposition angle for random encounters. Hoblins rooting around for food (but otherwise neutral to the party) are potential allies if you offer food.

I did wonder about your random encounter table having things like Grifting for Gold as an option that non-sapient beings would have, but you made a point about innovating that with the Broken Wing Wyvern example. Just handwave a Sudden Enemy Necromancer NEARBY which alters the stakes considerably. (in my example, I guess the wolves can speak now. "Hey. We've been cursed. Only if a hero stout of heart pays us a gold apiece are we freed!")

Excellent video! A good and thoughtful demonstration of how to run fun random encounters.

One thing you can do with the 2d6 is turn it from random encounter to the random event table. So 7 would represent nothing happening with 6 and 8 would be the next more common events such as a moderate difficulty encounter and discovering a location. The idea is yes rolling a 2 might be rare but that is where you place a deadly or extreme difficulty encounter. That way it is possible for the party to discover some monsters that might flat out kill them, like the wyvern they are too low to realistic fight. It can then be used as a foreshadow of what the can encounter, or show how truly deadly and out of their depth they are.

Putting your method to the test I made a quick table with 12 slots. each with an entity, a behavior, and a complication.
Some test rolls got me:
A blind commoner lost in the grasslands trying to find his way home.
An Ankheg trying to find its lair but someone stole it's brain.
A mind controlled howler fighting a young blue dragon.


Some combinations made so little sense I HAD to reroll them, but its good enough for an on the fly event.

Found you because of your superior geopolitics videos. <3 This one is overall solid advice, especially regarding commonality of encounters; but I do disagree that encounters should in any way scale with the party. There's a reason the peasants are scared to go out in the forest. There's a reason they stay at home working even though they make almost no money and suffer terribly sometimes. They're tough laboring men, after all, and you aren't much more at first or even second level. If you rolled that white dragon at level one, guess you're finding some way to hide... or else you're clearly not the heroes of legend after all. This goes double for dungeons and enemy territory where you clearly aren't welcome. The next party will find your corpses. The one survivor will tell your stories at the local pub. The random encounter table is not only a harrying tool meant to tax resources... it's a race against time and fate.

I hate scaling creatures, its a big NONO in my campaigns; hate in in video game RPGs as well. Players should feel how strong and how weak they are against some creatures. You want to both provide challenge for them, maybe even so much so they will need to start making harsh choices, and sometimes let them indulge in some power fantasy - but never just one always all the time

This also keeps the world dynamic and lived in

Baron vs Bob: the random encounter fight of the century

What's a wervan?

Yo, this is great.

As far as raw dice numbers I do prefer a belt curve personally l. But having adjusted modifiers based on party level/civilization/danger distances as well as 3 other random tables that gives context to each other's a a genius idea and I'll 100% be using those

The bell curve distribution "problem" only occurs, if you use multiple dice. Using a single dice/die yields the same probability for each number. So just use 1d12 instead of 2d6. If you aim for a linear distribution of probability and therefore only use a single die/dice, the downside is, you can only have a result of 1 with that single dice. But the points mentioned are absolutely valid and spot on.

The more dice you use, the more evenly the probability of possible values will be distributed. So just use 12d2 instead of 2d6 for a value range of 12-24 therefore also having 12 possible slots to assign encounters to. Apply the same method to increase dice count and spread out probability.

Very helpful!

This is funny, because anyone who plays Monopoly or Craps knows about the bell curve.

If I have a cool encounter in mind, I roll for the time of the encounter, but not for what the encounter is. Like rolling a D100 and when I role something like a 15 or below the encounter happens. Alternatively I'll role a D6, or example and the encounter or complication happens in 1D6 turns.

Weird idea
Using a coin flip to determine which method you roll with
Still gives you a bell curve but somewhat flattens the bell curves

simple and helpful video

combinatorics dictate several tables multiplicatively produce a much larger number of possible random encounters than one list.

As a first time player myself who has to dm for a first time dnd group of 5... I thank you

These issues are why I don’t actually roll on random tables most of the time. I just use them for inspiration and ideas to choose from. I don’t personally find much of a benefit in subjecting my own choices and narratives as a DM to randomness. Usually feels better to design the encounters specifically for the situation, the mood, the party, etc.

How is adding 2 encounters to a 6 encounter d6 chart and adding +2 to whatever you roll and having that decide. Any way shape or form easier and less complicated then just erasing the first 2 and replacing them with the 2 new encounters. Because that's effectively what you are doing anyway just with WAY more steps.

Always use a D12 for random encounters. Because you have to use a D12 sometimes and it has many great equal possibilities.

Woow, you really are the Master. Behaviour and complications make that the encounters feels like real. Very nice tip.

Great idea about d6 on a bigger table. About the encounters..I recently discovered a method watching a livestream of Chris McDowall. You came up with 4 nouns for a location like a forest. (1. Tree 2. Leaves 3. Stream 4. Bandits) Than you make all combinations (1. Tree Leaves, 2. Tree Stream, 3. Tree bandits ... ) and come up with ideas like 1. A fallen tree on the road 2. A Treant crossing the stream 3. Bandits hiding in a tree and so on. It's a really great method.

To eliminate the bell curve I recommend rolling a percentile die, d100, and just having assigned ranges. This creates a curve with the shape you want... flat, bell, inverted bell, triangular, whatever you want. And if you add some monsters it is easy to adjust to the new group.

But, random encounters are really only needed if you're reusing an area with groups that might have heard about the adventure before. Most of us write one-offs so you don't need random encounters. Just make encounters that fit the story, the time and place, the terrain and the party level. As far as the players know, everything is random.

While this makes sense, I prefer not to include the concept of level scaling in my games. I don't run D&D anymore, admittedly, but my primary game is still fantasy themed. My method basically involves a different tables based on position relative to settlements and locations of interest, and terrain types. If you're only a few hexes away from a major settlement or traveling along well patrolled, frequently traveled roads, your chance of encountering certain creatures or individuals is increased or decreased. Additionally, this allows me to incorporate regional bleed. Essentially, while you have a good chance of encountering a large band of wood goblins within certain forests, you can also encounter such monsters in the adjacent hexes, though less frequently and likely in smaller numbers (raiding parties and such). Your Dark Boar, for instance could be encountered more commonly as the party moves towards the interior of the woods, with a small chance of being encountered in the hexes immediately adjacent to its forest.

It's more laborious, but I like the fact that the players understand that they are far more likely to encounter something horrible the further they journey from the safety of civilization.

Here's a different idea: If you like the bell curve system, which I do in my case, because I use a lot of automated imbedded 2D6 tables, you can still "level up" those tables by making rare more rare and more beneficial encounter lower on the table.

Let's say your characters travel on land and could meet a merchant in the area near a city. If that encounter is more rare compared to meeting guards or simple travelers, increasing the result for areas further of the city will effectively eliminate them from the table. Meeting guards will become even more rare and travelers too, because of how less densly populated the area is. An encounter with, let's say goblins, that would be less likely near a city could then become much more common, since you are going deeper into the wilderness. Like the other system in the video, certain "unique" encounters will also start to appear.

These tables are also useful for generating structures depending on location.

Bro…. WHY-vern

Possibly whi-vern

Def not worvin

Awesome video man! I have been DM'ing for some time now and have come to a phase where I've become so comfortable with the role, that I now start to challenge some of my newbie habbits to try and weed out some of the things I do that might be a little boring. And I love the way this challenges the simple random encounter tables that I find in the modules that I have relied so heavily on. So thanks for rocking the foundation for my understanding of DM resources, and for expanding my ability to do a "generated" narrative plug ^^

I don't think the bell curve is a problem, it's the purposeful point. Some things being more likely than others, it's a somewhat more rudimentary way of adjusting the probability weights. I use random generator tools to adjust it even more precisely, so a rare thing can be more rare.

"You meet a... dragon! For the 5th time in a row!" It also makes it so it's not so samey, there's ups and downs, not just TV static. Hard to explain briefly.

I first noticed in minecraft actually, there was a mod that added a biome that was just hideous, blocks just completely randomly placed mean no valleys, no hills, no rivers, no caves, just a vast land of entropy style, non-chaotic chaos, all different and yet feeling the same. Field after field of TV static like randomness. Procedural generation is not "fixed" or made better by removing the weights and making everything equal probability, but then, you're not going for procedural generation. (But I kind of am, lol.)

Randomization with weights and sub-categories and sub-sub-cagetories, with different things generated as sub-parts of other things generated, seems to the next step or two up. Then you can explore and find things that actually feel different, rather than more of the samey static that's just everpresent, yet technically not identical. If you're doing improv as a GM and just using it as inspiration, it's not as big of a deal though, and the curve might not suit your tastes or what you're trying to do. Still, I wouldn't call it "fixing" it, any more than using a different palette or medium would be "fixing" art.

Are you trying to tell a story, or simulate an environment? These are both legitimate goals. They result in very different tables. A 'story' table is not suitable for long term play after the campaign ends; a simulation table likely works indefinitely.

It seems to me that you are describing encounters that have zones. Like the zones Professor Dungeon Master uses for his UDT you could have zones extending from civilization. Adding a +2 to a d6 for say, questionable, dangerous and unknown zones would be a great way to scale encounter tables.

Great idea! The old solo module Ghost of Lion Castle uses a wandering monster table that is similar but without the added narrative.

god, I really like your content, But damnit man, the way you pronounce why-vurn as whir-vin drives me up a wall. Keep up with the great advice, I'll soldier on through the thick bits.

Also on the random encounter table, once something has been rolled, eliminate it from the list. If rolled again, you can re-roll until you get a different result. Or just select the random encounter you want.

Dumb solution. The reason those weak encounters occur 45% of the time is they are common while the others are rare. Using a single die removes the only real purpose of random encounter tables which is probability simulation. If you're looking to have zero probability where uncommon encounters occur as often as common just choose your encounters.

This is an oversight of the simulating aspect of sandbox gaming, not some clever insight into fixing randomness.

A better solution would be to have your dire boar/wyvern whatever occupy the higher values on your table then add the highest character level to the roll. As the group increase in level add slots of increasing difficulty but retain one or two common encounters at the median. That way your dire boar/wyvern become more common plus goblins or rogue bandits now and then. But again this breaks the world simulating fulcrum. For a true simulation of probability encounter tables should reflect regions. High level encounters being common in difficult areas so players can choose to go there knowing what they're in for.

In a different video, you talk about how a slot should be dedicated to Factions & Double Trouble. How would you incorporate that onto something like the devil table? I love that but it feels like double trouble is hard to come up with reasonable ideas for the behavior/complication and the faction one would change somewhat frequently depending on travel speed or factions in an area.

Why not use d100 and adjust the bell curve bent you don't like manually with percentage points? The linear approach you add does coincide nicely with a linear progression of difficulty so that bigger more common brackets are lowest and the closest encounters to 100 or 0, 0 on the die, would be rarer and more difficult.

A single d10 added for a hundreds digit presents huge world building by random encounter options too.

Your 666 method is neat, and smashes together moral 2d6+mod rolls and diverse Random encounter charts.

Really neat and underlines the need for tools to enable dungeonmastering as fun and gamers.

Thank for the video!

I really dig the proximity encounter tables. That's a pretty great idea to add different encounters without making entirely new lists.

That proximity idea is brilliant! I am happy I found this channel, you have many great ideas that are new to me.

Edit: The added behavior and complication idea is also great! I am not sure I totally understand the major threat part. Would that be a 4th column?