Avoiding Unintended Responses to Incentives

Over the last few months, I have been devoting a ridiculous amount of effort into determining how much treasure different kinds of monsters should have. ('Ridiculous' especially because my foibles mean so much of this effort involves going down dead ends, recalculating the same thing different ways, etc.) As Alex eloquently said in his blog post about behavioral rewards, this data is important to the Adventurer Conqueror King System's goal of being "among the best-designed fantasy RPGs from the point of view of connecting the rules of the game (its laws) with the incentives of its players (its economics) design." A monster's average yield of gold pieces is even more of a reward for the players to defeat it than its yield of experience points; specifically, the design goal is for it to be four times as much of a reward.

But, precisely because players respond to incentives, I am worrying about whether the information about the reward should it be made obvious to the reader, or intentionally kept somewhat obscure.

In 3E, what Justin Alexander has called the "fetishization of balance" arises when players expect that all monsters their characters encounter will always be a challenge appropriate to their level. As Justin has pointed out, 3E's DM-level instructions about the challenge rating system specifically say that's not how it's meant to be used; challenge rating is a guideline not a rule, and the mechanics of the guideline ought to procuce many encounters both above and below what is exactly level-appropriate for the PCs. Other challenge-scaling systems are as old as the old school; see the 1979 JG "Temple of Ra Accursed by Set," or the Monstermark articles in the earliest issues of White Dwarf. Why did 3E, but not these other attempts, result in the fetishization of balance despite the designers' stated intent?

There are many answers to that question, but I think an important one is that the CR rating appears in the monster entry. As you read through that, most of the numbers are invariants that you might feel you need to know to design an effective character. A player of a mathy turn of mind who is thinking about choosing the Power Attack feat will want to look at the AC and hit points of a likely range of monsters to decide whether the feat's to-hit/damage tradeoff is worth it. Even some things that are more clearly guidelines, like the environment in which a monster appears, can be relevant to player decisions like "which type of favored enemy should my arctic ranger choose?"

So when a player sees a challenge rating in the monster entry, the tendency will be to treat it like a monster's armor class or environmental range: a fact about the world which they can use to make choices.

In paying a lot of attention to the ratio of treasure to experience yielded by defeating a monster - as in many other examples of ACKS design - I am taking inspiration from the "Red Box" Basic D&D set which is our most essential inspiration. In a post over at Delta's D&D Hotspot which emerged from our email conversations about monsters and treasure, Delta notes that original Red Box designer Tom Moldvay

shows an exquisite awareness of the average results produced by the treasure table system (as evidenced by his correct 3/4 ratio statement; and listing the correct average values for each treasure type, unique to his rules).

It's worth noting, however, that the place in which Moldvay lists these average values is in the treasure section. What he lists in the individual monster entries is a letter code, which requires some page-flipping to turn into the average payout in gold pieces. The analogy isn't exact because B/X is in one book and 3E in three, but I wonder: would the fetishization of balance have happened if the monster entry said "Challenging" instead of "CR 8", and you had to go to a separate table to figure out that "Challenging" meant that half of the monsters an eighth-level party encountered should be this tough?

Moldvay's treasure table average values only tell part of the story - basically the part before I started devoting insane amounts of work to figuring out the rest. In order to calculate how many GP a monster is worth, you have to divide the number appearing in an average encounter into the number in its lair, which is where those listed values appear.

What he chose not to do, but I am contemplating, is listing that computed value directly in the monster entry: when you encounter an average group of goblins, they have 1d6 gp each, and 300 gp back in their lair. It's possible that he didn't do this because he lacked the "benefit of enormously powerful spreadsheets, computer modeling, and random generators at our desktop" which Alex mentioned in his last post. Certainly, those have been invaluable in helping me work out each monster's payoff - but I wonder if Moldvay didn't list this directly to avoid an unintended consequence.

Here, then, are the questions I'm going through:

  • What would the consequence be? Players would be motivated to seek out the monsters that offer the richest reward for the least difficulty.
  • Is it likely? Perhaps not. It takes a fair amount of geekiness to rank monsters on profit-to-threat ratio, but the folks I play with at New York Red Box are at least this geeky. Although they have done the math necessary for pillaging by the numbers, I don't know that it has affected our behavior in the game all that much. In my White Sandbox campaign, player freedom of action is as wide-open as I can make it, but so far no one has turned down any of my enticing adventure hooks to search out and grind leet monsters like trogdolytes or giant rats.
  • Is it undesirable? It might have been for Moldvay, who was likely confined by the requirement to work with existing treasure types and their allocation to specific creatures. Arneson presumably gave merchants and nobles and dwarves hefty allocations of treasure because this was what made sense in the macro-economics of the Blackmoor. Sticking to these values meant Moldvay is in the curious position of having created both the world's best introduction to gaming for new players, and a powerful argument that the best way to advance in life is by murdering Lawful creatures for their stuff whenever they show up on the wandering monster table.

Adventurer Conqueror King values fidelity with previous editions of the world's most popular fantasy roleplaying game, both for the sake of backwards compatibility and because not fixing what ain't broke is a good way to avoid unintended consequences. However, we do have more freedom to change things as we see fit, and one of the changes I've made is to create a dungeon wandering monster table that only generates encounters that are enemies or rivals. And the treasure tables I'm creating awards treasure to monsters based as much on the player behavior we want to incentivize as on what makes sense for Gygaxian naturalism.

So I feel like it's OK to make it easy for players to see that the benefit/risk ratio of fighting a mummy is better than a hydra. In sandbox play, making these kind of important decisions is essential to the player's agency; and in a world where adventurers might retire to serve as advisors to up-and-coming barons, this is exactly the kind of knowledge that characters might seek out and codify.


I've been thinking about the problem of transparency in this economic model - balancing risk vs reward - the concern being it only works as expected in the mega dungeon setting. There, the players already have an over-arching goal (whatever they're doing in the dungeon in the first place) and economic transparency might effect planning micro-missions and delves when the group is choosing between otherwise similar options. The player-driven wilderness hex crawl would work this way as well.

Outside of that context, I tend to think story motivations and soft drivers play a greater role in influencing players than an economic analysis; either that, or the explicit monetary rewards offered by the patron or story hook.

But I am really intrigued about setting up a campaign that would lend itself to this type of analysis by the players, and whether it can be applied in other venues.

I agree that, in the White Sandbox, "soft drivers" seem to trump economic analysis. There is a lot of concern with pursuing profitable ventures, but it might be that the number of other uncertainties like "what monsters will actually be there" makes a risk-benefit analysis on monster types impossible or at least low in the mix.

That said, setting up a framework like this isn't meant to be player-facing; it is only the possibility that it could be taken that way, like CR in the 3E MM, that I'm concerned with here. Thinking about risk vs. reward is definitely something GMs do, and even if the effect of making it easy to run these calculations was slightly negative for players it might be worth it to make it lots more easy for the Judge,

That's true, the approach you guys are taking to presenting monster encounters and No. Encountered will make it easy (and interesting) to stock; adding average rewards right into the monster blocks would certainly streamline the DM's effort even if it's not typically actionable info for players.

Tavis, you might think about taking a clue from EPT. I dunno of you have looked at those treasure tables recently, but they are similar to the D&D ones, except the amounts vary a good bit from D&D. Anyway, in the monster description a column is given for Treasure, and another is given for Treasure in Lair. Each is followed by the familiar letter types - and a percentage.

The percentage is the chance that any treasure is there at all. Exampe T A: 40
So you could use the same method and add an average value of A, (800 gp frex) while retaining the chance that there is no treasure at all. To players this would mean that there are no gurantees and no real point to trying to maximize the monster treasure hunting.

As of earlier today, I have a set of tables I am happy with. However, I also have a bunch of custom-generated spreadsheets made for analyzing and generating treasure tables. At some point I will enjoy using these to break down the EPT tables, and will use that insight when I do my White Sandbox supplement.

Just as Alex's Auran Empire campaign has been a great asset for ACKS in helping us decide which assumptions to start with, one of my goals with the White Sandbox will be to show how the ACKS tools can be re-genericized and then customized to fit another campaign; and having Prof. Barker's awesome & long-running campaign to start with will be excellent practice! (I already use his city of Jakalla as the Nameless City of Limbo in my setting, so called because at first I didn't know where the map I had came from or what it was supposed to represent.)

White SandBox supplement? That's Awsome!!!

One of the odd things I noticed about Basic D&D a few years ago was that fighting kobolds was a losing proposition. They have the worst treasure rating by far. You're betting off fighting almost anything else.

Aldarron: There are lots of cool things on the Autarch release schedule, and the White Sandbox supplement is unlikely to be among the first of them, especially since many are spearheaded by Alex who writes publication-ready materials for his home game, whereas I operate on the principle that half-baking ideas mean I can have twice as many. Eventually though I would like this to be like what I think the Blackmoor supplement should have been: one part setting, one part house rules variants, one part new items and spells and monsters and stuff, all unified by notes talking about how it hung together in campaign play and the process of fitting it into ACKS.

mhensley, there are still nuisance monsters - one advantage of implementing a 4 gp to 1 xp ratio at the design level, rather than via the adjucation of individual encounters, is that you can bake in systemic variation instead of implying that rats and dragons and orcs and wolves all drop coins when killed at the same rate. But the variation is reduced overall, with a smaller gap between similar monsters like kobolds and goblins - and customized to monster types, so that kobolds more reliably have coins whereas rats are more swingy, sometimes socking away a shiny gem but also often having nothing.

The mention of Dwarves as one of the most profitable enemies to pillage got me wondering: will it be possible and/or desirable to stat up a Dwarf Lord, his (dwarf) fortress and followers instead of rolling up a # appearing? Can ACKS populate the wilderness with strongholds that behave and look just like the ones the PCs are going to make?

cr0m, at one point I had the idea of building the monster treasures from other data - like you would arrive at how much treasure an orc had by looking at how much it cost to hire one as a mercenary, then apply some calculation about how many month's wages the orc would save in order to reach the 4:1 treasure value. This proved to be unworkable because creatures you fight in the wilderness or dungeon need to have many more GP than the civilized economy provides for. This makes sense - think about how many random interactions you'd need to have to get a gram of opium if you were a stranger in NYC, as compared to being in a similar situation in a bandit town in Afghanistan or Cambodia - but it means you need different systems to handle each. To generate dwarf strongholds in the wilderness, use the domain rules. This reinforced my decision not to include Lawful creatures on the dungeon encounter tables; they're kind of apples and oranges.

In the case of NPC parties, I decided it was justified because they will often be returning from dungeon-economy scores. The average opium yield of beating up an American teenager depends strongly on whether you encounter them in America or in Afghanistan!

Tavis, you are an obsessive DM after my own heart. That project to build the treasure tables from derived data sounds like something I would have tried back when I was running 2e back in the 90s.

Thanks for the answers. I'm looking forward to generating some strongholds.

casque monster beats pro et aussi Inters experience Geffen community relationsPhysician
La conceiving d'insertion verticale,

Add new comment