Monday, June 28, 2021

Trophic Encounter Tables

We had a fabulous chat yesterday on the Wandering DMs Sunday talk show about our favorite Content Generators, starting with the monster tables in Original D&D provided as a way to populate dungeons and wilderness with a minimum of DM effort.

On that note, here's another guest article by our friend Angela Black: can we use real-world Trophic levels as a way to structure our random-table ecology simulators?

I've recently become bothered by encounter tables that propose that a certain wilderness area might be infested with all manner of huge and lethal monsters. Perhaps I'm too literal, but when I see that this particular forest contains dragons and owlbears and displacer beasts and wargs and and and... I always ask myself, "what are the EATING?" I suppose the answer might be 'each other,' but that's oftentimes too glib for me. I frequently find myself wishing for a more 'realistic' model of what monsters live in a given area.

Fortunately, scientists are great at modeling all aspects of the real world, and this sort of thing is not exception. Ecologists use a concept called "trophic levels" to describe the flow of energy through a given environment, which incidentally gives us a nice model of the relationships between predators and prey. Without wading too deeply into the very complicated math of this fascinating concept, we can use it to gain some traction on a more realistic model of encounter tables for fantasy roleplaying.

To sum up the ideas at play very briefly, we can imagine a pyramid-shape divided into zones that are called Trophic Levels.

  • At the base, the largest level is Trophic Level 1, where we find plants that form the base of the ecological system. Plants are rarely - but not never! - threats to PCs, so we don't normally list all the plants in a given region on the encounter tables.
  • Just above this we have Trophic Level 2, where we find the herivores that feed on the lower level These tend to be small creatures and likewise not *usually* a threat to PCs, so we don't tend to put all the squirrels and rabbits and whatnot on the encounter tables, either.
  • Trophic Level 3 is for smaller predators that feed on the herbivores in Trophic Level 2, like foxes and weasels and whatnot. These are usually only threats to humans in the real world if they feel directly threatened or they are protecting their young.
  • The next level if Trophic Level 4, where we find predators who feed at least partially on other predators, like hawks that might eat foxes. The important observation here is that creatures that occupy Trophic Level 4 are accustomed to attacking other creatures that are in themselves dangerous, as long as those creatures seem vulnerable. That will be important for gaming!
  • Finally, at the very top, we have so-called "apex predators," who can and will prey on anything in the area. Not every region will have an apex predator, but if it does, the apex predator will present a clear threat to PCs!

Now, we can already make some observations that will inform our encounter tables. We can (for now) ignore anything that would live at Trophic Level 1 or 2 in our region. On the other hand, we must account for the creatures that live at levels 3 and 4, as well as the apex predator(s), if any.

But the populations at these trophic levels live in very predictable relationships - for instance, there can't be more creatures at level 4 than at level 3, unless the creatures at level 4 are vastly smaller (for some reason) than the ones at level 3. If creatures at level 3 were rare and creatures at level 4 were abundant, the creatures at level 4 would quickly run out of food!

Generally speaking, the ratio of creatures from Trophic Level N to Trophic Level N + 1 is 10:1 or thereabouts. This is a very rough gloss of the actual math, but it works for our purposes. This also assumes similar body mass, but we'll ignore that for now. This figure of 10:1 gives us a great starting point for roughing out an encounter table.

Let's start with creatures at Trophic Level 3, since anything at levels 1 or 2 will only present a threat to PCs under very rare circumstances. Let's look at the Temperate Wilderness Forest tables in the MMII to get some ideas.

We should start by picking out a handful of animals for Trophic Level 3. These are animals that feed largely on herbivores (though they may themselves be omnivores that also eat plants). Some good candidates from the table we're look at might be badgers, boars, poisonous toads, snakes, weasels, and wolves. That's a great start.

2Then we move on to Trophic Level 4. What are some creatures that might prey on (at least some of) those creatures? We should choose fewer of them, maybe about three or four. We'll pick bears, hawks, and - just to add a fantasy element - owlbears.

Finally, we should choose an apex predator (and it is recommended there not be more than one, but this rule can be broken with careful consideration). For fun, I'll say there are fire lizards in this forest. Found under "Lizard, Giant," in the Monster Manual, the fire lizard is essentially a drake, a very dragon-like creature that is not intelligent but more of an animal. It has 10 HD and can breathe fire so it can certainly fill the role of apex predator!

Now to create the table. Let's start by observing the ratios. Collectively, the creatures at level 4 should be approximately ten times more likely to be encountered than the apex predator, and likewise the creatures at level 3 should be approximately ten times more likely to be encountered than the creatures at level 4. Now, we can't quite accommodate that on a standard percentile table, but we can approximate with the following standard layout:

  • 1%: Apex Predator
  • 2% to 11%: all the creatures at level 4
  • 12% to 99%: all the creatures at level 3
  • 100%: everything else

This layout preserves a realistic relative frequency of encounters - it wouldn't make sense for fire lizards to be massively more common than, say, bears, or even wolves! Likewise, the creatures at a higher level should be more dangerous overall than creatures at a lower level.

If the DM wants to adjust some, a 1:5 ratio could be used without stretching credibility too much, leading to an alternate table layout as such:

  • 1% to 3%: Apex Predator
  • 3% to 18%: Level 4
  • 19% to 93%: Level 3
  • 94% to 100%: everything else

DMs may find this layout is more congenial to gameplay while still preserving the same general ratios.


  • The strength of the apex predator can be a guide as to whether the land is ripe for settlement or destined to remain wilderness for the foreseeable future. It should be assumed that humans in a fantasy setting will, as they were in the real middle ages, attempting to clear land into which they hope to expand. In fantasy settings, however, there are things which can easily repel even determined groups of humans, and if the apex predator in a given environment is sufficiently difficult to kill, certain areas may remain uncleared and unused regardless of the wishes of the local civilization! For instance, a region where the apex predator is the lion may present danger, to be sure, but can eventually be cleared for use by determined humans. Compare this to a forest where the apex predators are phase spiders!"everything else" is a useful catch-all for the rare, miscellaneous encounters from levels 1 or 2 that might actually prove dangerous to PCs. For instance, AD&D has numerous dangerous plants, like whipweed or yellow musk creepers. These plants cannot be exceedingly common, however, or there'd be no creatures from level 1 left to eat the plants! Putting such encounters in "everything else" is a good way to model this.

  • Placement of intelligent humanoids requires some judgment. If the area is truly wilderness, the humanoids must either be non-cultivators (like ogres) or exceedingly rare. For instance, if the DM wants to include human bandits in the woods, it's probably best to list them as an option under "apex predator," since if humans were in the woods in any greater abundance, they'd have cleared out the other apex predators and started to tame the land. Goblins, however, who avoid the larger creatures and mostly live by trapping herbivores and small predators actually qualify for level 4!

  • The level of the encounter is a fine guide to the behavior of the creature encountered.
    • Creatures in level 3 are non-confrontational unless desperate or frightened - they will fight if their homes are invaded, their young threatened, or they are backed into a corner, but that's it. It may be that characters experienced in woodscraft can "defuse" an encounter with a creature at level 3.
    • Creatures at level 4 will only attack with an advantage - if the target is smaller and weaker, if the target is injured, or if the target is outnumbered. Again, goblins make a fine level 4 creature, don't they?
    • Apex predators can and will attack anything they choose, and are frequently good at stalking prey and assessing the danger they present. Apex predators may hunt singly or in groups, depending on type and, if semi intelligent or intelligent, may present a massive threat to PCs.

Delta back here again -- Personally, I think that's pretty nifty, and the solid real-world math being used as a foundation is the kind of thing that gives me really good results in the past. Would you consider using that as a basis for your tables?


  1. Hmm fun.
    I can see for a given wilderness zone, roll to generate your apex predator (with in some range of Hit Die/EHD) and that is your defining monster, then that allows you to target a number of tophic level with HD ranges for the rest of the monsters/creatures. Possibly tying apex HD to number of hexes they "control" (almost like castle/lord....) So now you have a general idea of the ecosystem and sphere of influence for a given wilderness..

    I wonder what how this would muck up the traditional dungeon encounter levels. I guess it speaks to the "fairy tale" underworld concept, in order for all these monsters to subsist in dungeon.

  2. I like the idea, and think it's very useful in thinking about "stocking" a wilderness. However, I wonder if the biomass ratios are the same as the chance for a person to encounter a specific animal.

    1. They def aren't, but they're the best starting place we have.

    2. Good detail to consider!

  3. For my 'Lost Valley of the Hutaaken' campaign I wanted some big predators for the party to avoid/deal with. I used the average hunting range for bears, wolves, lions, guide me on how dense these would be. I used the terrain differences to squeeze them in a bit. Owlbear stayed mostly in the heavy forest while the pride of lions prowled the grasslands and everyone avoided the lake claimed by the giant crocodile. I used the 1:10 ratio loosely to fill in the rest for smaller, less dangerous stuff.

    So far so good. The party has avoided certain areas (the lake especially) just to avoid certain creatures and changed their security habits (no posting guards on top of buildings around the owlbear, who flies).

    The area isn't so tightly packed it feels like a zoo but each zone has a distinct feel and danger to it.

    1. That's very nice, I like that!

      also slightly jealous you stocked the Lost Valley successfully when I failed as a teenager


  4. Very nice approach to stocking an environment, though. I like it.

    That 10:1 ratio of lower to higher trophic levels, mind, refers strictly to the metabolic demands of creatures at each level. There’s a lot of loss or waste, so to speak, as the higher-level predator can’t absorb every bit of energy available from the lower-level prey.

    But: not all predators have the same metabolic demands. You could have a large number of big snakes and crocodiles, for instance, preying on almost the same number of mammalian prey, because the reptiles have much lower metabolic demands. The same number of antelope that might support 10 lions could in principle support 100 crocodiles.

    Similarly, predators and prey don’t necessarily have the same body mass. You could have 100 predators supported by 100 prey, provided the predators were much smaller animals, their total body mass only 1/10 that of their prey.

    The example that springs to mind, in a D&D context, is the dragon that sleeps for months or years at a time, wakes briefly to ravage the countryside, then goes back to sleep. If its flight and breath weapon are magical (not supported by biological functions) then you could plausibly have dozens of dragons in an area that has only a few times their collective mass in prey animals (and humans). If they were all actively predating all the time, their prey would indeed run out quickly. But if they have super-low average metabolic needs (because they sleep for years at a time) there could actually be quite a few of them around, seemingly disproportionate to their base of support.

    I’m not sure this changes the ratios for a random-encounter table, because less metabolically active predators are also likely to be less energetically travelling about the environment, and hence less likely to be encountered. It would certainly affect estimates of how many dragons are lurking underground, and similarly numbers of drakes, giant serpents, terrasques, dragon turtles, and the like, and for that matter predators that conceal themselves with immobility, like ropers, gargoyles, etc. There are lots of traditional D&D monsters that seem to spend time in suspended animation until the party blunder into them; they could be very numerous indeed, despite limited prey.

    Just my $0.02. I quite like this approach to the design of encounter tables. It might be worth extending it to the number of bandits who can plausibly support themselves on a few passing PC groups, or asking how pirates continue to exist when “None dares sail the Karlang Archipelago, these many long years”, or indeed how quite small towns can support magic shops only well-heeled PCs can possible afford. A fertile set of ideas.

    1. The BX clone I play has an economic system based on tiers. Each settlement is given a number between 1 (a metropolis e.g. Lanhkmar) and 6 (tiny village). The amount of goods you can get are limited based on cost. Simple system handled with a tiny chart so I know how many flasks of oil are available for sale in the town. It gives PCs a reason to go to the big city to sell and buy stuff. Adds realism w/o complexity.

    2. That's a great point, I like that very much!

  5. Large herbivores are both a potential threat to pcs AND a food source to apex predators. I think Gygax assumed a whole Lotta herd animals in the wilderness of fantasy worlds. In the monster manual it lists possible herds of over 200 of them roaming around. Imagine the buffalo herds of the early americas, but all over. That could feed a hydra or two.

    As for dungeons, I always worry about water sources for creatures, but only worry about food sources for sentient non-magical things (e.g. not worried about the mimic or any of the vermin).

  6. This greatly resembles the suggested method of creating encounter tables from the 2nd Edition Dungeon Master's Guide. That method, however, is integrated with the creature frequency ratings given in the Monstrous Manual and Compendia, with suggested ratios of 70% common, 20% uncommon, 7% rare, and 3% very rare creatures. There's even an example with a set of creatures picked out for a desert environment, and how to fit them into both a 2-20 table and a percentile table while approximating this "ideal" ratio. The sample percentile table looks like this:

    01-11 Camel
    12-22 Giant centipede
    23-33 Herd animal
    34-44 Ogre
    45-55 Orc
    56-66 Huge spider

    67-70 Basilisk
    71-74 Caravan
    75-78 Hobgoblins
    79-82 Nomads
    83-86 Giant scorpion

    87-88 Chimera
    89-90 Pilgrims
    91-92 Harpy
    93-95 Dervishes
    96-97 Salamander

    Very Rare
    98-99 Lamia
    100 Djinni

    Of course, these frequency ratings are a subjective mixture of both prevalence and likelihood to encounter a particular creature. Humanoids in particular tend to be assigned a higher frequency regardless of terrain or level of civilization, presumably with the intention of more interesting encounters.

    1. Hey, thanks for that reminder -- I tend not to be too up on 2E materials. Interestingly similar %'s are given in the notes of the 1E MM (65/20/11/4), and it says these are "considered" in construction of the DMG tables, but it's never explicated exactly how (and also weren't printed until 2 years later).

  7. I'm tempted to handwave that the 10:1 increase in numbers as you move down the ladder is exactly balanced by the decrease in likelihood that they'll be a dangerous encounter. We can discard the 99% of rolls that have the PCs encounter a harmless herb or whatever.

    Thus, you could could have encounter tables with 6 rows (the five Trophic levels plus Other), roll a d6 to select one, then roll on that row to select the actual encounter.

    If a truly random encounter is needed, not selected for danger, you could use exploding dice and count the number of dice to determine the trophic level. That removes the issue of percentile dice not having probabilities small enough to capture the odds of a given creature being a dragon.

    1. That's actual a pretty strong argument. Of course the OD&D tables have really no "non-dangerous" encounters at all (unless one follows an option in a particular footnote).

  8. I wonder if a 2 die bell curve (2d8, or even 2d10 or 2d12) might give a better probability distribution? I tend to prefer those for tables instead of percentile rolls.

    1. Of course you cannot use smaller than 1% increments with d100, but you can definitely set up a d100 table to approximate the probabilities of 2d8/2d10 fairly well.

    2. Definitely an option. Gygax even specified a d8 + d12 system as a new default in the 1E MM II.

  9. No, I would not consider these as a basis. I believe in a kind of 'heightened reality' where interesting shit happens to the PCs at an unrealistically high rate (but still randomly!). So I'll go with a high monster population density and a high encounter rate. That said, the ideas in the article *are* useful as a point of comparison.

  10. AD&D MMII (138) gives a method for designing random encounter tables. It uses 1d12 + 1d8 to generate a number from 2 to 20. Very rare creatures are encountered at the two extremes, while common creatures appear in “a large flat spot of equal probability in the 9-13 range.”

    In the system, there are not five levels but only four: very rare, rare, uncommon, and common.

    Looking at the probability for each result (using, it cuts close to the 1:5 ration Angela gives.

    Percentage to dice results (1d12 + 1d8)
    1% to 3% (3%): Apex Predator — 2 or 20 (2.08%)
    4% to 18% (15%): Level 4 — 3 to 6 (14.59%)
    19% to 93% (75%): Level 3 — 7 to 16 (73.94%)
    94% to 100% (7%): everything else — 17-19 (9.3%)

    I also notice that, in the book’s two examples (same page), animals populate the large (common) flat spot.

    (AD&D, first edition, if there is ever any doubt.)

    1. Those charts were an inspiration to this! I just disagreed a lot with how they sorted creatures into "common/uncommon/rare/etc."

  11. I'm a bit late to the party, but two things are worth noting.
    1) You're using the count of creatures in the area, but a better metric is rate of biomass exchange. On land this rate usually corresponds roughly to extant biomass, which strongly suggests bringing in groups at lower tropic levels. This also works out nicely given the prevalence of groups at lower trophic levels; there's lots of herd animals around.
    2) You'll notice I specified "on land" earlier. There's a reason for that, and it's that marine biomass distributions are shaped by some extremely short lifespans with absolutely ridiculous reproduction rates in terms of biomass expansion. A given mass of phytoplankton replicates a lot faster than grasses, let alone trees, and this ends up almost inverting the pyramid and putting the bulk of oceanic biomass in top predators at any given time. This provides a nice justification for horrifying sea monsters having an outsized space on encounter tables and the ocean getting to be extra scary, which works out really well in game terms.

    1. Thanks for that! Angela & I have actually been cycling the last few days and (1) is very much on my mind, trying to work that into a system. (2) is news to me, fascinating!