I’ve decided that January is “Make Your Own Island” month. This started out as me puttering about with some encounter tables, but things escalated quickly, and I decided to turn it into a fully-organic, quick-and-dirty summary of how you can create a little campaign setting. In last week’s post, I ended up creating an island map and broached the concept of nested encounter tables. More on that below, but first…
Some Housekeeping
To provide a sense of scale, I added the Regional Hex template as an overlay to the MapGen2 map of Minocra. Last week’s version used the CC3 version of the template, but before this exercise is over, I’ll want to place some symbols on it, add some labels, and embed notes. All these things will be easier in Hexographer, so in keeping with the quick-and-dirty approach I’m after, the map at right is rendered with the Hexographer version of the Regional Hex template. The map itself is semi-transparent, so the colours don’t “pop” as much, but it’s a trade-off between making the map or the hex overlay visible. You can’t make out the hex numbers, but you still get an idea of what’s what.
Added bonus: this paves the way for creating a true hex-only map, which may or may not materialise—all depends on how jiggy I get with it. Now, back to the business at hand…
Nested Encounter Tables
As suggested in last week’s post, the best encounter table is one that you customise for your setting and your group. In other words, you’re defining the setting by the encounters in it, but with a particular eye for what challenges your players and keeps them interested. Put that way, an encounter table is more than just a roster of what the PCs might meet 1-in-6 times every few hours—it’s really more of a blueprint for who lives in the setting and what they do there.
I like nested tables, which requires a multi-roll, drill-down approach to determining encounters. One of my favourite models is from the Rules Cyclopedia, which provides a chart for each terrain type, and the results lead to sub-tables of different monster types—not, as might be expected, actual monsters. For example, if you wanted an encounter for grasslands, you’d roll on the “Plains” table, which might give you a result of “Insect” or “Human.” Then you go to the appropriate sub-table, where you roll again to see what Insect or Human actually shows up.
This is where the format becomes useful. Instead of interpreting “Insect” as an actual monster, think of it as an encounter type. Now the Insect sub-table can include any manner of bug-related stuff.
With that in mind, the last tweak is weighting the table to reflect the odds of one encounter occurring over another. The original RC encounter tables were equal-distribution—there were 8 results, and you rolled 1d8 to determine which one you got, meaning that each entry had a 12.5% chance of coming up. This is fine if you want equal odds, but I suggest you weigh your top-level encounter tables because it makes it easier to devise encounter types.
Chimera provides some guidelines for monster frequency, suggesting Common (40%), Uncommon (30%), Rare (20%), and Very Rare (10%). You can apply these frequencies in general terms and create tables using 1d10 or 1d20, like this:
| Encounter Frequency | Die Range (1d10) | Die Range (1d20) | Frequency |
|---|---|---|---|
| Common | 1-4 | 1-8 | 40% |
| Uncommon | 5-7 | 9-14 | 30% |
| Rare | 8-9 | 15-18 | 20% |
| Very Rare | 10 | 19-20 | 10% |
Note that your table need not be limited to 4 results: don’t confuse frequency (as a percentage) with roll result (as a single number). What’s important to remember is that Common encounters occur 40% of the time represent 40% of the table’s composition. That means, on a 1d10 table, 4 out of 10 results must be Common. This could be a single Common encounter in the 1-4 range, giving you a 4-in10 chance of a Common result when you check the table.
If you need more “slots” on the table, you might split that 4-in-10 chance into roll results 1-2 and 3-4, or even 1-3 and 4. It really doesn’t matter how you arrange the table entries, so long as Common entries don’t occupy more than 40% of the possible roll results. However, doing this changes the probability of those results coming up (e.g., 1-2 equals 20%, 1-3 equals 30%, 4 equals 10%). So while your table’s composition is 40% Common, the probability of getting a Common result might be less than that (thanks to Roger for pointing this out).
Alternatively, you can accomplish similar wizardry by rolling multiple dice and adding the results. This forms a bell curve, familiar to many gamers the world over, where results in the middle of the range occur more frequently than results on the extreme upper and lower ends. It works, and involves more math (albeit not very difficult math), but it can be time-consuming. You can get clever and use 2d6, 1d4+1d6, 3d10, whatever combination you want, but each combination offers up a different set of probabilities, and very few are going to be as tidy as the model shown above. My advice is to keep things quick and easy to create and maintain by sticking with 1d10 or 1d20 (or, if you’re into big tables with lots of results, 1d100).
Thus, our revised sub-tropical jungle table from the last instalment might end up like this:
| Die Roll (1d10) | Encounter | Sub-table | Frequency |
|---|---|---|---|
| 1-2 | Animal | Animal | Common |
| 3 | Insect | Insect | Common |
| 4 | Plant | Plant | Common |
| 5-6 | Human | Human | Uncommon |
| 7 | Sub-men | Sub-men | Uncommon |
| 8 | Event | Event | Rare |
| 9 | Monster | Monster | Rare |
| 10 | Unusual | Unusual | Very Rare |
From this, you would generate sub-tables for each Encounter Type. When you make sub-tables, the frequency variation is up to you. My rule of thumb is this: if the table has a sub-table, I use the distribution above. But if the table is final (i.e., there are no sub-tables), I use equal-odds distribution. Continuing with our example, if I get a 3 on the table above, I roll on the Insect sub-table:
| Die Roll (1d10) | Insect Type | Sub-table |
|---|---|---|
| 1-4 | Crawler | Insect (crawler) |
| 5-7 | Flyer | Insect (flyer) |
| 8-9 | Burrower | Insect (burrower) |
| 10 | Parasite | Insect (parasite) |
Again, I’m creating encounter types, though they’re limited to categories of insects. Because there are more sub-tables, I use a weighted distribution on the Insect table. But the remaining sub-tables Crawler, Flyer, Burrower, and Parasite) are the end of the line, so I’ll use an equal-odds distribution for each. Here’s the Crawlers sub-table, which uses 1d6:
| Die Roll (1d6) | Insect (Crawler) |
|---|---|
| 1 | Ant, giant |
| 2 | Beetle, luminary |
| 3 | Beetle, tiger |
| 4 | Rock Mantis |
| 5 | Spider, hunting |
| 6 | Swarm |
Final Words
We’re just touching the tip of the encounter table iceberg, and there are a few more details I’d like to add to the mix in the next instalment.
As a side note, my apologies for missing last Wednesday’s post. These are busy times, and I’ve been heads-down ramping up my new Chimera RPG offering, which will happen by the end of the month.
This is shaping up to be a fantastic series/resource, Erin. I’m really looking forward to following along. Enjoyed your other world building articles. Great stuff indeed.
@Reese : Thanks for the encouragement. This series is probably going to be bigger than I had anticipated–there are many nooks and crannies to explore–but I’m really glad you’re along for the ride!
This is great and I’m along, too! You make all this things look fun and awesome.
Good stuff Erin. It would be great if this process could be tied some how into your “Monster Turf” article.
Well timed! I planned on creating my encounter tables this week for four geographic regions near the starting place in my next campaign (due to start in 2 weeks.) I was considering something along the same lines; but, your post formalized my thoughts and made it easier to just get writing.
The main difference in my approach and yours is that I’ll not go with even distribution on the sub-charts (though fairly close to even.)
Thanks!
@Benjamin : Thanks – I hope I can deliver to your expectations. Next up are some ways to get more meaning out of encounter table results.
@Josh : Good catch – you have a hell of a memory 😉 I’ll work on adding that into the next bit.
@Eric Wilde : Cool – glad this is helpful. I hear you on the even distribution–I can’t say I always use it for the final sub-table, but I tend to. My thinking is that the weighted distribution on the parent table(s) steers you toward the odds you want by the time you get to the last sub-table. But customisation is key, so whatever approach gives you the results you want is the right one to use. Care to send a link when your tables are done?
I’ve had this thread in the back of my mind for a few days now Erin. I saw the awesome maps put out by that island generator, and wished there was something similar for planetary continental maps. Then yesterday I was half-listening when my kids were watching TV and heard the word “Caldera”.
Something clicked and now I have this idea to create a world where the entire known world is smallish island continents contained within the caldera of an enormous volcano. Thank you, sir!
Sure, I’ll send a link when I’m done. Probably on Monday.
@Reese : That’s a cool idea! You could also generate a bunch of islands with mapgen2 and place them geomorph-like on an Atlas template as “tiles.” And they don’t have to be equal-sided tiles, either–you could resize some islands to 1×2 Atlas squares or 2×3 or whatever.
As a side note, I’ve found that the Atlas template is a lot of territory–plenty of space for an entire campaign. Each of those templates, at 25 miles per hex, is 390,625 sq. miles, or the equivalent of about 4.13 United Kingdoms.
That’s exactly what I had in mind Erin. Multiple mapgen islands spread about a round “flat earth” world bordered by an impassable ring of mountains. Are they truly impassable? What could lie beyond?
I’ll be making good use of your templates and Hexographer to map the islands in hex format. I hadn’t thought to resize some of them, so thanks for that tip too!
Wonderful articles, thank you for posting them.
I have one math related nit-pick, but it does end up being important if people are really wanting to get their distributions correct.
You stated that “That means, on a 1d10 table, 4 out of 10 results must be Common. This could be a single Common encounter in the 1-4 range, or two Common encounters split up 1-2 and 3-4, or even 1-3 and 4.”
Which isn’t quite right. What actually happens is that you’ve taken a common occurrence (1-4/10) and turned it into two rare (1-2/10) and (3-4/10) or one very rare (1/10) and one uncommon (2-4/10). Dice rolls are independent. Two separate entries spanning four positions on the die are not the same as one entry spanning four entries on the die. The first has probability 20% each, but only once, the second has probability 40% once. You’ve made one common occurrence into 2 rare occurrences. You’ve counterintuitively made rare occurrences common.
You’ve gone from 1 rare event to many rare events because it’s the distribution that matters not the designation. If you fill up your chart with 8 (1/10 chance: very rare) encounters and 1 (2/10 chance: rare) then you have exactly that, 8 encounters that happen 1/10 times and one encounter that happens 2/10 times.
I got burned by exactly this when I decided to make random spell scroll tables, I had so many spells in my “uncommon” group that you had less chance of getting a given uncommon spell than you did a rare spell.
@Reese : That sounds like a great project – I hope this series helps (and I hope even more that you share whatever you come up with).
@Roger3 Ack! You are correct, sir, and I apologise for the misinformation (because I
sometimesnearly always suck at math). I’ll post an update anon. Thank you for pointing this out!I puzzled over your explanation a few times, because I couldn’t quite get my head round it at first. But I finally got it. To explain why you’re right, let’s assume I have a 1d10 table of spells, like this:
1: hold portal (1st-level)
2: light (1st-level)
3: magic missile (1st-level)
4: read magic (1st-level)
5: knock (2nd-level)
6-7: web (2nd-level)
8: dispel magic (3rd-level)
9: fireball (3rd-level)
0: wizard eye (4th-level)
Strictly speaking–looking at the table only–40% of the results give a 1st-level spell, 30% of the results give a 2nd-level spell, 20% give a 3rd-level spell, and 10% give a 4th-level spell.
However, the actual probability of rolling a 1st-level spell is equal to rolling, say, a 4th-level spell (i.e., 1-in-10). IOW, the chance of rolling “hold portal” is the same on the 1d10 as rolling “wizard eye”.
I’ll be dipped, Roger. Thanks for pointing out this grievous error. I’ll update the above and address it in the next post. More than anything, this puts forth the need for the bell-curve tables I dismissed above.
Thanks for keeping me honest (because, again, I suck at math)!
Here’s the tables. I ended up going for more even distribution then I expected in the end.
http://ewildebeast.blogspot.com/2012/01/swords-and-dark-sorcery-encounter.html