The Simple SECRET to Better RANDOM Encounter Tables 

💥 Czepeku: www.patreon.com/czepeku ✅ LIKE & SHARE: ru-vid.comvideos ✅ PATREON: www.patreon.com/bobworldbuilder

I love how its technically possible to have to fight every single encounter at once

An alternative way to do it is to roll one die, but assign a weight to each result, such as a result of 1 to 6 on the die always resulting in kobolds, but only resulting in dragon on a 20. The strength of this variation is that you can build the pacing into it, such as this: You don't want the dragon encounter to happen early into the dungeon? Have a d20 table but assign the rarer creatures to values above 20. Then, as each encounter passes, add an increasing bonus to your die roll, so these rare encounters become possible, or even probable as the game goes on.

5:09 “ I’m sure you’re familiar with the idea of food chain: Plant grows > deer eats plant > human eats deer > dragon eats human. Simple. “ 😂

Hahaha, “that’s not a secret it’s page one of the original dm guide” was my first thought. Props to you Bob!!❤

Great vid! A favorite dice mechanic of mine was for a homebrew where rather than using a d20, we rolled checks using d8+d12. The bell curve reigned in extreme rolls and felt less swingy. As an extra feature, the d12 was used to generate which body part you hit with an attack, and the d8 was used for damage. That way, a high attack result was almost guaranteed to also be a headshot for big damage. Very nifty go get 3 different results with 1 roll!

The game Cyberpunk Red uses 2d6 tables for this reason, and it works really well. A very dramatic example is the injury table (if critical damage is rolled), where it takes a roll of 2 or 12 to lose an arm or leg (versus a foreign object or torn muscle is in the middle). Naturally one of the first injuries my character took was a lost arm, and since it happened so early it became a big part of her character. But it only happened once to a PC in the whole campaign, so it always felt like a big deal that it had happened.

Depending on the game I sometimes cross out an option on my roll table and roll again if I hit that (could replace it with roll two so things escalate as players kill more) this means that your rarer encounters become more likely. I.e. your dragon is more likely to come out the more of its minions you kill.

I like using hexflowers for random encounters! Or random weather, terrain, treasure, dungeon rooms... They're cool because the last result affects the next one. Also makes it feel like a bit of a board game, *and* makes it harder for the players to learn what's actually on the table. You'll start in one hex, then roll dice to determine which direction and how many hexes you move. Tada, now you have the new result! On regular tables, rolling the same exact result will always get you the same thing, but on a hex flower it can end up being different.

On a random note, I really enjoyed your use of artwork from the 1st edition Monster Manual. I always say that D&D lost a little of its personality when the art started being so slick and realistic. That mimic with the fist coming out of its side is iconic!

I actually have used a special encounter table in the past that i call "weighted random" where you make a 1d100 table and put the results you want across various ranges. For example, kobolds would be a 45 number range where as dragon would be a 5 number range and so on. This led to a set of 100 where more likely results took up more real estate, thus making them more likely to appear

You can also bake the bell curve into a classic d100 table, and you still get the joy of rolling 2 dice. You can fine-tune the odds along a bell curve.

That 2d4+1d10 table would be perfect for your example dungeon your about equally likely of encountering kobolds, goblins or a trap but extremely unlikely to encounter the dragon in a random room

if you make the bad encounters at the higher numbers and not that bad at the lower numbers, you can do what d&d does during character creation. roll an extra advantage die or two. If monsters are on high alert, roll 3d6 and keep the 2 highest results per random encounter, or keep the lowest if the players are being particularly sneaky and good at traversing or something. just a fun thought.

The other really cool idea for random tables is weight it with less dangerous stuff at low numbers and more dangerous stuff at higher numbers, and then roll 1d6 + X where X is the “danger value” of the encounter. I think Dungeon Masterpiece did a video on that. Also, making sure you have good thematic tables (including non-combat encounters) and an “activity in progress when encountered” table keeps things interesting!

Back when I was a kid, basic and 1e days, it encounters didn’t matter as much and I used random tables all the time. When I returned to the hobby I found that I used them less and less favoring to preparing certain events/effects that may occur in a given session. Thanks for the great video!

The 2d4+1d10 roll looks like it could be useful for weather. If it's the middle of summer, the chances of a blizzard should be fairly low, while variations on "warm and sunny" are basically a given.

I've been using fewer random tables over the years and more planned encounters. Partially because I was disappointed that certain encounters didn't show up during the game because they weren't rolled, but mostly because random encounters just started slowing the game down. It started to feel bad when the end of the night seemed to always line up with a part of the adventure where the party was traveling and we were just rolling for random encounters for like the last hour or two of the game session. When I was using random encounters, I used to use the 2d6 method, typically with one result of "no encounter" so I'd have the random chance of having an encounter baked in with the result of which encounter we have, and if we were doing overland travel, I split my table into a day and night column. However, I started experimenting with a 1d12 method. It worked like this: I rolled Xd12, where X is the number of possible encounters I wanted to happen during the given time period, based on how dangerous the area was. Each player rolled 1d12. If any player roll matched one of my rolls, that encounter happened. That way, a bigger party naturally led to more chances for an encounter. If multiple dice matched, I would either run the encounter twice or would double the encounter, like there were twice as many monsters. Yeah, it resulted in some wild encounters sometimes, but it didn't happen often, and these encounters were usually one per day, so the party was typically at full health and spells, so each encounter could be more dangerous than normal without a TPK. Also, I sprinkled a number of beneficial encounters into the list as well, so the party had a chance of encountering friendly NPCs, or the remains of a previous battle, and could get some loot.

I love bell curves! PBTA style games are built on this maths, and I think they work really well. However in my experience, most D20 fans tend to dislike bell curves, feeling that they're 'cheap', or somehow deny them agency. Discuss? :)

The problem with 2d6 is that it exacerbates the problem of getting the same thing repeatedly, which risks becoming boring pretty quickly. This is why I prefer drawing from a deck of cards for random encounters; that way you can ensure you don't get the same thing twice.

The 1D10+2D4 would be great for a dungeon themed around a fight between 2 appossing forces.

My favorite random tables have always been 1d12+1d8. But those ones are best for larger tables I feel. Gives you five equally common spots to set a theme. And I like building the tables in a way that a roll 2 is actually just rolling 1d8+2 and 1d12+2 separately on the table for the mixed encounter.

When I saw the notification I immediately started thinking of my multi dice tables, then I saw the thumbnail and burst out laughing. Great video

if I was going for a specific theme, that would just be what's in the table. Still using a good bell curve has value

Anytime you mix dice types (d6 + d10 for example), you'll get flat spots in the middle. It's a great way to combine equally likely results with less likely edge cases.

That 2d4+1d10 table could be good for treasure. If you set the middle values for various amounts of gold, then start working towards common, uncommon, and even rare magic items as it gets to the edge. That would allow you to have some fun items you would want to give away, but still has a random chance at being in a treasure hoard or on some main monster in a random encounter. Or the party just gets some gold that they can then use to buy other items that they want/need.

As you told, this is a great way to push the story you want. And it reminds me of old school RPG on video games when you had to look for/avoid random encounters as you never knew when the big bad creature was about to ambush you !

I was actually working a random table for the last few days and found myself unhappy with the level potentially of all the events happening when I preferred some over others. This solves my issue perfectly and perfect timed! This subs for you...👍

5:29 dragon eats man… 😂 This video can also be called “How Does Gambling Help Build a Better Random Encounter Table”. Great job introducing statistics to the world building community! This is why Vegas makes 7 the Lucky Winning number on the Craps table… you know the movies where they throw the 2d6 but there’s NO miniatures on the table;) They are applying this bell curve to gambling. This is kind of what happens when you “Roll w/ Advantage”; even though you don’t get to add them up for a better average it creates the possibility of hitting the MEDIAN roll of 10! Love this video!

Brilliant combination of explanation and visual aids! Yes, I'd love to see another video and PDF about building dungeons around these tables. Thanks, Bob. You're a welcoming and insightful part of the community.

One of the most useful videos I have ever come across . Thanks Bob . I use this week sponsor and they make by far the best maps and the randomness has helped when I had no idea what do some weeks as the DM

I liked the pokemon reference and the heavy metal. Keep it up Bob, your content is great! I would also suggest turning every random encounter table into 1d100 and assign percentage values to all your options. You won't be trapped into using every face of the die for an option, where if you only had three things you wanted your players to encounter, 1-15 is your first encounter, 16-50 is your ssecond encounter, and 51-100 is your third. It takes one extra step in math to set up, but this way you aren't dealing with bell curves and all sorts of extra encounters you really weren't interested in running.

Great video, Bob! The relation to Trophic Levels was really cool. I hope we can see more of that background sneak into your scripts, I love the opportunity to learn a new concept connected to a hobby I love. Cheers!

Great video Bob! A different problem with random encounters unrelated to the table is the arbitrariness of when they occur. Sometimes you go 10+ dungeon turns without an encounter, then get several back-to-back. Last month 'Goblin Punch' published a great post on a countdown clock system. Check it out, it builds suspense, in a thematic way, but is a more reliable than straight rolls.

Perfectly timed video! I was just looking at making some random encounter tables, and this has given me quite a few new ideas to incorporate.

You bring up a strong point here that's actually the crux of the homebrew system I use. A d20 is too random and you're equally as likely to roll a 10 as a 20. In true-to-life performance, people aren't that equally likely to perform very well as they are averagely well. On average, they do ... average. This is why I use 2d6 for everything in my system and base mid-difficulty on being able to roll a 7, which is the most likely number to be rolled. As things get harder, it's actually MUCH harder to reach those target numbers reliably.

I like to use d4 + d6 and d4 + d8 and for larger areas d6 + d8. Using different dice values really allows you to make custom tables that suit what you want in the adventuring area. Great video as always!

This "tropic" would be true is there were no ability to teleport, gate, plane shift, spelljam etc..

Congrats on the CzePeku sponsorship! I love their maps so much

I'm thinking that if the dragon is the "boss", it wouldn't be on the _random_ table at all. By definition it would almost have to be _the_ last encounter. But although random encounters are usually for "between" the planned encounters, I can totally get behind the idea of a "themed" encounter table being used for the whole dungeon, with regular random rolls being uses in passageways between rooms and the rooms themselves being where one doubles or triples the random encounter rolls.

It's May 3rd and Bob still has his Yule tree up! Kudos to you my man! Great video! I love bell curve results. The old AD&D 1d12 and 1d8 is also a fun system. Where when you roll on the table it feels exactly like rolling a random encounter because of the two different dice denominations. Again, great video!!!

Another fun addition for a random encounter table would be a rival party exploring the same location 😊

If you like a straight 2d6 tables you'll probably like tables with drop high/low dice. They take the normal distribution and skew it higher or lower without changing the number of outputs. You can also drop high and low to skew more towards the middle.

Great video and idea. It makes so much sense that it is somewhat baffling, that this is not the standard. I personally like to include some encounters that can be solved by RP or are not necessarily threatening.(One time two PC on alert saw some small green humanoid - vegepygmy - doing weird stuff and communicating with bloops and drums. To this day, the other players do not believe this really happened) It was hilarious and dramatic at the same time.

If you do want to have a 0.1% encounter on your random encounter table I recommend rolling 1d4 d4's, It allows for 16 results with a distribution skewed heavily to the lower numbers, The distribution is a bit hard to predict, But as a rule of thumb numbers above 7 get increasingly less likely. Meaning 13 to 16 are the stuff of legend.

Great meat & potatoes content! One advantage of using a single D100 and a table is it's very easy to customize the percentage chance for each result by how many numbers it comes up on. Not limited to bell curve frequencies (though you could easily set that up too...), but still intuitive because we're so accustomed to thinking in percentages.

The Anydice site is a great resource for planning probability curves. As soon as I saw your first graph, I thought of it even though the graph was from a different source.

Shoutouts to AnyDice, the website you briefly featured towards the end! It's absolutely my favorite tool for RPG systems design. I use it extensively for my projects

Can’t tell you enough how helpful this explainer was. I’m now converting the Castle Ravenloft chapter from a bloated 30 page maze into a concise set of tables for each floor that’s less backstory focused and more party driven

Nothing to to do with the vid but I've been meaning to say how much I appreciate the North Watch print in the background. Parkinson was my favorite of the 'old school' TSR artists. I've still got the Dragon mag with that on the cover. Cheers.

Yeah Czepeku! Love their maps.

This is super helpful! Another way to adjust the randomness is to use 1d100, but make the results based on a range. For example, on a 1-5 you get the dragon; on a 6-20, you get the Ooze; on a 21-75, you get a kobold; etc. This method allows you to get very precise with the odds of certain things being rolled. If you want something to have exactly a 7% chance of being rolled, you give it a range of 7 (i.e. 1-7 or 34-40). It's not often that you need to be that precise, but sometimes its useful!

Bob I'd like to say I love the videos. I've started dming recently after playing and watching for a year now. And My friends are loving the game, so thank you

Another solid video with game changing advice. You rock!

One could also just make sure that when making the random table all of the options fit the theme they are going for. I do like the idea of kobolds (or any enemy group) having a pet mimic. They feed it and use it as a trap for those who enter their lair.

Automatically, as soon as you said "Bell..." I realized exactly what you meant. So simple! I can now do dungeon random encounters where the minions are more prevalent. It made no sense a dragon had the same amount of chance appearing in a small 30x30 foot room than a group of skeleton minions, not only that but appearing again in the next encounter.

Hey Bob, always love your take on things. I use a VTT [Fantasy Grounds Unity] with a card option so whenever I roll for encounters it never chooses the same encounter again. If I were using tabletop, and rolled the same encounter I add 1 to the roll if that would go over my 1d10 table I would just go back to 1 until all encounters were used. I have used both the add 2 encounters and the 2d6 table as well these are very good at what you say. I tend to use encounter tables for specific areas like woods, deserts etc. I also have a reason for them to be there, to create or keep some kind of realism in a fantasy setting. I did like your take, which I have not done in the past, in randomly rolling what the encounter is doing there. I like this idea. Keep up the good work!

I first saw this idea in the hex crawl video by Dungeon Masterpiece. Works great to 'show not tell' the overall feel for a region!

14:20 This is very similar to dungeon masterpiece's concept. In his random encounter video he mentioned having a behaviour and problem list that applied to everything on the table, further minimising prep, as instead of prepping 3 possible behaviours for each possible encounter, you only need to prep a single behaviour and problem for each, allowing for more variance for less work.

Imagine you have a forest or a dungeon ruled by four factions. That 2d4 + 1d10 table could make running into one of the four factions very likely, each of the four is equally likely to show up, and still allows other, rarer creatures and events to pop up. Great video, btw! Immediately useful advice!

A great addition to random encounters is action! That is, not just "what" but "why". What are the goblins up to? Why is the ghost there? Hunting, searching, guarding, fleeing, arguing, building, and so on.

you can also put the more dangerous results on the top, and better ones on the bottom. Maybe even have results of 1 or 13 on there. and then have modifiers based on the players actions. For example; being very loud gives a +1 on the random encounter table, or sneaking around in the dark a -1.

Czepeku is definitely worth it btw, I used their patron a few months ago after feeling ashamed for using the low-res versions of the maps that they post on reddit. They're great maps!

There's tons to love with rolling two dice! Also: Picking one die to reroll and picking higher or lower can help skew things to one side or the other of the table depending on the actions of the players. A modifier can also serve this purpose. I'd recommend going to anydice and perhaps using a spreadsheet to help visualize the probabilities! It is also possible to use this method to do other things as well, such as creating different ranges for outcomes such as negative 2-5, middling 6-8, positive 9-12. Bottom line is that messing around with the numbers can be fun!

this video is a great companion to your relevant encounters video from last month, great work, thanks!

A bell curve mechanic I really like is to roll three (or five) dice and choose the middle result. This feels faster than adding, allows smaller tables, and gets a more bell-like curve.

13:28 the great and crazy narrative-driven OSR RPG called "Grok?!" uses 1d4-1d6-1d10-1d12 rolls as attributes bonuses to d20 checks. Doing so keeps the 5% of a d20 roll at the flat part, while slowly lowing the probabilities at the borders of the bell curve. The more sides a attribute die gets, the slightly less likely is get results out of 5% flat top. You should check this RPG out, Bob.

I have a house-rule written for D20 that changes the base resolution to 2D10 instead. Makes crits and crit fails properly special. Pretty simple rule, only a few changes. But the bell curve is why GURPS uses 3d6. Note: there's a Dragon article from many many years ago that covered weird dice rolls, the core one for the article being 1d10 x 2d4 and similar calculations. That gives a skewed bell with an average of 27.5 but a huge long tail--80--at the top end for super-rare results. Could even go simpler with 1d6 x 1d6 x 1d6: minimum 3, average about 43, max 216. Even 1d6 x 1d6 is min 2, average about 12, max 36. If you want to stay with d20s, 1d20 x 1d20 is 2 / 110 / 400. Give you some fun tables. Or, as the article suggests, also damage results.

I like that you took dungeon ecology into account for making the table. I definitely like the concept of having the bell curve built around a theme. Overall a great concept.

Someone didn't watch Dungeon Masterpiece's excellent video on why bell curve encounter tables are a terrible idea! Basically you end up with sessions full of battles against wolves, and none against dragons :D

Oh I love that authors random tables. Did he make a hard cover compilation that I missed? Or maybe I'm thinking of someone else...

This is also a bit of throwback to how DnD started. Modern DnD is based on d20 rolls. The game started with multiple d6 . . . because walking to your local store and buying a d20 was not a thing for most people. The game was later updated for those new fangled non-cubes.

Again, i HAVE to comment how much i like your hair pulled back. Looking good my man! Keep up the good work

I just build a table giving 4 shares for common, 2 shares for uncommon, and 1 share for rare creatures. I just heap them in alphabetically but grouped thematically or regionally. With a computer, you can roll 1d, so the final number of total shares doesn't matter at all. This is better than a bell curve because it ensures proper representation regardless of how many entities are rare or common.

Really helpful and a very clear explanation! Thank you

Thanks Bob! this was very useful!! Yes a PDF! I like where this is leading!

Really like that Keith Parkinson poster in the background. I remember owning that Dragon Magazine back in the day. Memories!

This was essentially the method used in the AD&D Monster Manual 2. That used 1d12+1d8 to create tables ranging from 2-20 but with a (very slightly) different distribution of results than 2d10 would create (because 2d10 has 100 possible rolls, d12+1d8 has 96 possible rolls)

Worth noting that a lot of the flat encounter tables are written in a way to put the most interesting options at the extremes anyhow. So you can often roll 2 of what it says to roll and half it and get a decent bell-curve version of the table

I quibble with your reasoning... the goal should be to create a fun experience, which may or may be "naturalistic". And the encounter table has to viewed in the context of the planned encounters of course, so on it's own the table doesn't' necessarily need to make a lot sense. But for sure, using 2 or more dice is an easy way to squeeze the distribution of outcomes, should you judge that is desirable (and you don't feeling like making a d100 table with ranges). I had never thought about mixing two different dice and how that would affect the distribution, so that was cool. Great video.

I have at least three encounter tables for different zones in my campaigns, and depending, they are either 1d6 or 2d6. The 1d6 areas are usually open and sparsely populated. But for the 2d6 tables, I always use doubles as the "roll two" option. That has several effects: 1) all the even numbers have possible double rolls, so I make those easily pairable, and not too strong on their own, and 2) both 2 and 12 are exclusively doubles, so never come up on their own, which means they need to be exclusively parable, like something that could supercharge any other rolled encounter. Then the rarest encounters I put at 3 and 11.

Hey it’s one of the few things the 5e DMG got right! I love 1d12+1d8. Room for variety with 4 equally likely results, most likely results only about 8% while the most extreme are a cool 1% ish. If only the book’s guidance and writing were as good!

Just as tech bros that reinvented the bus, you just successfully reinvented the 2-20 (1d8+1d12) random table. Congrats!

I was not expecting an ecology lesson in here! That was hands down the most interesting part of this video to me, and I would love to see a video going into more detail about how to apply trophic levels to tabletop RPGs!

I’ve always been highly cognizant of dice probably when making random tables. I switched to a 2dX system like your examples long ago. But I've always customized my tables to fit the environment. So a swamp table would have swamp-like stuff in it; lizardmen, bullywugs, crocodiles, will-o-wisps, maybe a black dragon; a witch or hag (which are great RP encounters; especially when you have them hand out quests). I highly recommend it.

variations of encounters work for wandering monsters because it provides some narrative element to the randomness by providing the type of encounter but also the circumstances upon which these encounters happen. for example: …6 = 2 Kobolds patrolling 7 = a group of kobolds hazing another kobold 8 = a stealthy kobold tracking the party 9 = 8 kobolds returning to the dungeon with stolen treasure and goods ect.

The 2D4+1D10 roll could be used to fill a merchant caravan's transported goods, 1D4+1 to determine the variety in the shipment. A merchant will usually be transporting a few specific goods, and only a small number of goods outside of their usual trade route. The least likely outcomes could be nothing, or rare things they are bringing to a buyer, or are paid to drop/hand off at a place on their route. If you rolled a nothing, and/or are missing a common trade good, that could be a story element.

Another way to control the result balance, or to further control it past the bellcurve: On each entry add in a box with the “stockpile #” in it, each time you roll that result lower the stock by 1, if it hits 0 reroll the result. This lets you fill a dungeon with a random distribution, but still have balanced numbers. You can also make it so certain encounters are built up slightly by having them add a fraction to another event’s stock, so say for every 10 goblin rooms it generates a hobgoblin or goblin shaman room chance (still needs to be rolled). Additionally, as you stated, some alternate dice pools, or mix-matching multiple rolls, can really change how things happen. For example, the 4D6dL that is used for character stats leads to a skewed, sideways banister cap looking graph, or 10D10 rr duplicates once can be a really odd castled box looking thing

Thank you for the useful tips. I wish there was more content about the "science" behind the dice... Nice Gnawa music in the background! 😊

Good work Bob. This will help with my next campaign.

This was one of the best and most helpful videos I have seen from you. I thought you would have lost your power after cutting your samson-like hair, but I guess not. Keep it up.

I thank Catan and it's colored sized tokens for teaching me this some years ago... But also, great use of this bell for encounters! Neat idea to reinforce theme and have a little more control over results.

Great concepts. I've been having a lot of fun using 1d6+1d8 for my encounter tables. The distribution curve is a bit wider and flatter, which makes the edges (dragons) a bit more rare and the middle (kobolds) a bit less repetitive. This way I can have multiple, equally likely common encounters.

When I started running Curse of Strahd for my group I wondered why it made me roll 1d8 + 1d12 for encounters. After I researched the matter the propabilities came up which I hadn't even considered. These days I use it for a lot of stuff in addition to encounters, for example when I need a random NPC I roll on a table where some species are more rare to encounter than others. Also here's a tip for those who want to get everything out of this trick: You can have a larger table with some more dangerous results at the top, but you use smaller dice except in some more dangerous areas, so those extreme scenarios are limited to those times/places 😄

This is the video we didnt know we needed.

Yes please, a video on trophic levels would be amazing. The only thing I know about trophic levels is trophic Cascade from when they release wolves back into Yellowstone; which was amazing. It would be really cool to hear more about that kind of stuff for the purpose of world-building.

Here's one 2d6 table for a Primal Forest environment I use when solo-rpging if anyone wants to use it: Primal Forest 1. Gravitational ore 2. d6 poachers 3. Directionless Mushrooms (spores disorient travelers if inhaled, roll perception to spot on time) 4. creaking Sequoias 5. 'saurian ground nest w/d4 eggs 6. Footprint ferns (fragrant ferns that grow to fill giant footprints) 7. furtive shrews 8. creeping moss 9. gigantic skeleton 10. distant roar 11. Wand Wood (items made from this tree have their own pool of mana mages can draw on) 12. Monster (megafauna usually)

Whoa, where'd you amazing hair go? You look great Bob, just startled me

I love it. I forget why I started making tables of all the ways I could achieve different probabilities. I think it was because I wanted the option of making crits much more rare (both failures and successes). I know it can be a tall order to basically rewrite the whole handbook's suggested dice, but if I was making an electronic game I definitely wouldn't be loyal to D&D's systems.

So great. I love this video so much!