Home > Game Mechanics > Risk vs. Reward

Risk vs. Reward

Calculating the payoff for derring-do

Last week, I presented my rationale for awarding experience in Chimera; my belief is that it addresses the needs of a majority of campaigns. But Anthony over at of Pedantry raises an oft-forgotten aspect of role-to-advance models: how does the system account for risk vs. reward?

In other words, adjusting the Advancement Roll in proportion to assumed risk. Also, it needs to apply to non-combat activities. Plus, the system must be objective, elegant, and intuitive. One that looks nice, and not too expensive…

I’ve been rolling 3 potential solutions around; this post contains the synthesis.

Defining Risk

Risk-reward curve

Risk/Reward Curve

My working definition of risk is “a chance of loss.” Literally. I’m not concerned with what you might lose—just the odds of losing it. Ergo, a 75% chance of failing a Diplomacy roll represents a greater risk than a 55% chance of falling into a pit of liquid hot magma. For purposes of experience, the act of exposing yourself to risk is more significant than the consequences of failure.

If you’re looking for a quantifiable measure of risk, you might start with the relationship between an action’s Target Number (TN) and a character’s chance of achieving it, after applying his Action Roll (AR) modifiers.

I’ve decided that a chance of failure greater than 50% constitutes risk. Based on a d20, this means an adjusted Target Number 12 or more. Let me plant this stake in the ground:

  • An action is risky if (base TN – AR adjustment) ≥ 12

Experiencing Risk

Let’s start with the last bit first: Attempting a risky action qualifies you for a base Advancement Roll bonus equal to +1 per point above TN 11 (i.e., +1 at TN 12, +2 at TN 13, +3 at TN 14, etc.).

Now, let’s put some parameters around this, lest we end up with the same bean-counting issues that prompted me to ditch XP in the first place:

  • Only Clutch Situation actions are eligible for this bonus (i.e., a Clutch Situation damage roll, for example, does not qualify).
  • You need a Normal Success or a Critical Failure to get the bonus (a Normal Failure isn’t sufficiently instructional, and a Critical Success could be a fluke).
  • The actual outcome of the attempt (i.e., what happens in the game) is unrelated to the bonus—what’s important is that you either succeeded or failed miserably.
  • Multiple bonuses are additive, though the total bonus cannot be split across more than one Advancement Roll.

That’s the nickel tour. Let me blather about my rationale for a bit.

I figure that Clutch Situations are important enough to leave a lasting impression on the character—certainly enough to affect his shot at improvement.  In other words, a Clutch Situation is more than just a “mulligan”—it can be a developmental milestone.[1]

This approach also lets the player chose the situations—and level of risk—he wants to “apply” toward his Advancement Roll (he could even use multiple Clutch Situations to get a bigger bonus). This optionally decouples experience awards from combat—now the character can get a bonus for any action that involves sufficient risk. And, appropriately enough, the bonus is proportional to the risk assumed.

Finally, by limiting “risky” rolls to Clutch Situations, I’ve taken care of a few pesky details: (1) it provides a framework for what rolls constitute meaningful risk, (2) it ensures that the player applies some discipline to his acquisition of Advancement Roll bonuses, (3) it offsets the “current level penalty” applied to the Advancement Roll (thus overcoming level limits), and (4) it doesn’t require a lot of bookkeeping.

Qualitative Tweaks

You’ll note that I said “base Advancement Roll bonus” above. You know I’m a sucker for tweaks, right? Let’s talk about the qualitative value of risk.

risk v. reward

I can haz pelican?

This is harder to quantify because different things have different values to different characters at different times. That said, we can make the sensible generalisation that temporary or tangible rewards (money, consumable goods, tactical gains) have less overall impact on experience than permanent or intangible rewards (survival, influential connections, strategic advantage).

If this distinction makes sense in your game, halve the Advancement Roll bonus for temporary rewards (round fractions down, minimum +1), but leave the bonus as-is for permanent rewards. My rational is that short-term gains provide a quick, one-time learning lesson, but long-term gains, having more lasting consequences, provide a longer “learning tail.”

Tying It All Together

Klar the barbarian (Vet 2; AB Athletics +3, Fight +2, Shoot +1, Sneak +2; AdCost 10) has 2 Clutch Situtations (until he reaches 3rd-level). During an adventure, Klar is the “point man” for the party, and he needs to sneak past a pair of orc sentries. The attempt is TN 16, adjusted by Klar’s Sneak +2, for an adjusted TN 14. Klar uses one of his Clutch Situations and rolls 2d20 for results of 8 and 17. He takes the best result (17) for a Normal Success.

Because the action was risky (TN 12 or more) and because Klar used a Clutch Situation,[2] the Advancement Roll bonus is +3 [adjusted TN 14 - 11]. Because this was a tactical gain, the GM rules that the +3 bonus is halved. His Advancement Roll TN equals his AdCost of 10; adjusting for his level (-2) and his bonus (+1), he must roll 11 or better [10 + 2 - 1 = 11].

He makes the Advancement Roll and gets 2 Improvement Points. He opts not to level-up, since he still has 1 unused Clutch Situation (let’s say he improves his WL and Fight by +1 each instead). In the next adventure, Klar has an opportunity to make an ally of an influential priest. This requires a Diplomacy roll against TN 12. It’s definitely risky, because Klar does not have the Diplomacy Ability and therefore suffers a penalty of -4 to the attempt [TN 12 - (-4) = 16]; Klar uses his remaining Clutch Situation.

He rolls 2d20 and comes up with 10 and 4—both short of the mark—but he takes the 4 since it’s a Critical Failure and, while he’s probably insulted the priest hopelessly, he can at least learn something from his mistake. The Advancement Roll bonus is +5 [adjusted TN 16 - 11], which the GM leaves as-is because the attempt was for long-term gain. As before, his Advancement Roll TN equals his AdCost of 10; adjusting for his level (-2) and his bonus (+5), he must roll 7 or better [10 + 2 - 5 = 7].

Final Words

Well, that was crunchy, yeah?

I believe this delivers on the requirements of objective and elegant. Intuitive remains to be seen, as I’m seeing this through the author’s lenses. My only additional comment: this system can be used in conjunction with (or as a substitute for) story-based Advancement Roll bonuses, so you can make it optional without breaking anything.

As always, I’m interested in your thoughts and feedback. If it makes sense, it’ll be added to the 5th Printing.


  1. Just to keep you up to date, this is how Clutch Situations are defined in the 5th Printing: Heroes tend to come through in a tight spot. You get 1 Clutch Situation per level; each Clutch Situation you use grants 1 additional die on any roll you make, and you may select the best result. You must declare how many Clutch Situations you’re using before a roll is made, and they’re lost once spent, regardless of the roll’s outcome. Clutch Situations renew when you level up; unused Clutch Situations do not carry-over from level to level.
  2. Klar’s player may or may not have known the actual TN (depending on how the GM feels about that sort of thing) but remember that, to be eligible for an Advancement Roll bonus, the action must be TN 12 or more and be made as a Clutch Situation.

  1. March 10th, 2011 at 08:53 | #1

    Once again, forgive me of my ignorance of the fine details of your system and if I get them screwed up.

    From my own perspective, I want to look at risk and reward in a broader sense. You give the example of specific rolls that have a significant failure rate, but I don’t think that correlates to risk. There are always going to be actions that carry a greater than 50% failure chance (D&D example: any to-hit where you need an 11 or higher), but that doesn’t necessarily signify a risky situation.

    Ok, let’s use another D&D example. I’m a 3rd level fighter and I am fighting a goblin, AC 7. I need a 10 or higher to hit him, so it’s not a risky action. Put the goblin in scale mail and now I need a 12 or higher to hit him, so now it is a risky action. Regardless, a lone goblin is not match for a 3rd level fighter.

    Forgive me if I am misinterpreting your post. I also understand you are tracking advancement per action rather than per encounter or session. Plus the limited amount of Clutch situations puts a frame around the number of times someone may advance. What you have here looks almost like a skill based system wherein you advance skills by using them. Not a problem at all, just a totally different implementation. In this case, your system does make sense, practice makes perfect after all.

    Thanks either way!

  2. March 10th, 2011 at 11:02 | #2

    @Anthony : I think your interpretation is essentially spot-on, and your D&D example is apt–while the “to-hit” roll may be 12 or higher, a 3rd-level fighter vs. a lone goblin doesn’t seem at all risky.

    However, I wouldn’t think the player would expend a Clutch Situation on the attack. Even if they did, the bonus would be only +1, which would seem a wasted opportunity. That’s why the bonus requires a TN 12+ and a Clutch Situation. The TN simulates an appropriate level of difficulty (i.e., defined numerically instead of via GM fiat), and the Clutch Situation ascribes in-game significance to the action (i.e., elevates it to “bonus-worthy” status).

    I also understand you are tracking advancement per action rather than per encounter or session. Plus the limited amount of Clutch situations puts a frame around the number of times someone may advance. What you have here looks almost like a skill based system wherein you advance skills by using them.

    This is a little off the mark, but maybe this will help put the above in proper context:

    Chimera doesn’t track advancement on a “per action” basis (maybe you meant “per adventure”?). Instead, experience is more about completing an adventure than what you did on it. IOW, you’re assumed to leverage all your capabilities to survive the mission–otherwise, advancement is a moot point.

    Clutch Situations have no bearing on the number of times you can advance, though the number of Clutch Situations does dictate how many advancement bonuses you could earn between levels. IOW, you don’t need Clutch Situations to advance, you just don’t get the advancement bonus above without them.

    And you’re right, Chimera is generally a skill-based system, but with an important difference: while using skills can improve your chances of advancement (as above), you don’t have to improve those skills when you advance (e.g., Klar used Sneak to get an Advancement Roll bonus, but then improved his Fight and WL).

    Still, I don’t think I’ve accounted for risk in the broader sense, as you put it. I think I need some more examples of what you’re after.

  3. March 10th, 2011 at 17:21 | #3

    @Anthony : UPDATE Unless you’re basically saying, “I want to decouple Clutch Situations from advancement bonus.” As in, Adventure A (broadly speaking) is more dangerous than Adventure B. So how does the GM reward PCs who undertake the risker Adventure B? Is that closer to what you’re after?

    (Forgive me–I’m slow on the draw these days…)

  4. March 11th, 2011 at 11:13 | #4

    I use D&D examples since it is the system I use, but don’t think I consider it the be all and end all.

    The classic D&D convention is that the deeper you go into the dungeon, the harder it gets. This is reflected on the XP side, the PCs have a measure of control on their own advancement and acceptance of risk. The XP and treasure system work together to put emphasis on these tougher monsters by giving more XP for the kill and/or their treasure. This all works in a general sense, balancing dungeon levels, risks, and rewards.

    My main goal is to somehow incorporate something with a similar feeling to other game play areas. Obviously, the above example is only useful in the dungeon setting or the 3e string-of-encounters setting.

    Let me try a different approach. You mention “experience is more about completing an adventure than what you did on it. IOW, you’re assumed to leverage all your capabilities to survive the mission.” What about a game without missions? What if you put PCs in to a town that is populated with NPCs, organizations, conflicts, alliances, monsters, magic, religious persecution, rebellion, etc and just let the players run rampant? In this hypothetical, the DM has NO missions planned.

    Ok, the PCs decide that they have an old grudge against someone who is now a petty thief in the town. They spend the evening tracking him down, trussing him up, and disposing of him. The next day, they decide that they want to get involved in the looming civil strife that is threatening to boil over into war. They poke around town, talk to some movers and shakers, position themselves into the overall scheme, maybe conduct some espionage or magical scrying, stockpile weapons and food while they are cheap, etc etc. Then, some time down the road, a conflict breaks out between two of these groups and a small skirmish erupts in which the players get involved.

    So in these 3 hypothetical sessions of gaming, how can we systematize risk, reward, and advancement? My original thought was to just give 1 exp each night, in effect, awarding experience for everything (or nothing), but that created all the problems we’ve been talking about. Clearly, each session has its own set of challenges, risks, rewards, pitfalls, tough situations, easy situations, role playing, combat, skill usage, etc.

    I guess this problem does come closer to your second comment (risky adventure A vs B). But it is also born of a mismatch in gaming style and system. D&D is based around combat and loot, so how do we incorporate its experience system into a setting that is more robust than combat and loot? Tangentially, how would you assign risks and rewards to game actions that don’t require explicit skill checks?

    I can get behind the general idea of “any risk at all is good enough” or “the act of exposing yourself to risk is more significant than the consequences of failure.” Meaning, you wouldn’t get experience for buying low and selling high, but it would count if you bought low in a safe city and sold high in the war-torn, monster-filled border fort.

    Just to be contrary, there is one more issue here: Throw away all the hypotheticals and design questions. What do players do on a nightly basis? They get themselves into conflicts and fights! So is this all just pedantry at this point, do we just trust in the fact that any given night of role-playing will eventually cross over to the realm of combat and conflict and leave the experience system alone?

    Because that is the point I am at. I won’t use fiat and I can’t come up with a better system. My current system is to put faith in the players that they will get into trouble on their own and the ‘combat and loot’ style of XP will balance out in the end. Or, ultimately, despite all the scheming the PCs do, the XP rewards don’t come along until they actually go out there and bash some heads, figuratively and literally.

    Let me know what thought pop up on your end :)

  5. March 12th, 2011 at 10:45 | #5

    @Anthony : I’m pretty sure we’re on the same wavelength, but it seems I’m making some broad assumptions that may not be accurate.

    First, while everyone gets an advancement roll for surviving an adventure, the advancement bonus described above does depend on specific actions.

    Second, eschewing fiat requires quantification, but I don’t want to do a lot (or even a little) bookkeeping; by extension, this means avoiding “adventure-scale” calculations the GM might have to do in order to assign some risk value to a given mission or undertaking he runs for the PCs.

    Clearly, each session has its own set of challenges, risks, rewards, pitfalls, tough situations, easy situations, role playing, combat, skill usage, etc.

    I think the system above flexes well against this “mission-less” style, assuming the GM provides challenging opportunities for PCs to use Clutch Situtations. In situations where the GM has nothing planned, this approach lets the players decide what activities will impact their advancement.

    In that sense, it really doesn’t matter what the adventure is about, or how inherently dangerous it is–what matters is how much each PC participates. Think of it this way: I’m on a mission to retrieve the Golden Fleece. If I’m an Argonaut, I’m pitching in, but I’m not taking a lot of risks. If I’m Jason, I got more skin in the game (no pun intended).

    Now, both Jason and each Argonaut get Advancement Rolls after the adventure–because they did it, because they survived, because they saw and experienced amazing things. But Jason’s advancement bonus potential is higher because he extended himself in risky things, unlike the typical Arognaut.

    If we’re to abandon fiat, I don’t think you can decouple advancement bonus from skill rolls (or, on a higher level, some pre-calculated adventure challenge rating). There are plenty of skill rolls to be made in each of the hypotheticals you provided, and I think it’s fair that the player controls his advancement bonus by taking (or avoiding) risks along the way.

    Hunting down a petty thief could be easy or hard or extremely challenging, based on the rolls required to achieve the desired goals. If the PCs decide that they want to pursue this course, they might first check around to assess risk, but ultimately, how much they engage in the challenge determines their advancement bonus (if any). Maybe the PCs are good at tracking people down, or maybe they’re really tough–finding and trussing up the thief won’t be very challenging, and it’s possible that no opportunity for advancement bonus comes up.

    But maybe one of the PCs has to do some diplomatic side-stepping when an underworld boss asks uncomfortable questions, or maybe a PC mage uses a spell to cover the party’s escape from the thief’s hideout. These are individual opportunities for the PC to earn an advancement bonus. They’re not part of the adventure per se but they do represent a character’s choice to exert over-and-above effort during the course of his adventuring career.

    At the risk of boring you even more, I’ll go on to say that this approach is completely independent of head-bashing, which I don’t think is really required to advance. A GM might wish to run his game that way, but it doesn’t have to be so. I ran an urban sandbox, and one night the characters decided to find and hire an assassin. In six hours of playing, there was 1 fight, but definitely a lot of other stuff going on–a lot of rolls for diplomacy, sneaking, observing, academics, athletics, and mettle.

    (BTW, the best way to break the ice when asking around for an assassin? Stick a dagger in the bar and ask the innkeeper, “Do you know anyone who knows how to use this?”)

  6. deimos3428
    March 13th, 2011 at 17:11 | #6

    Not a fan of this approach, but it’s an admirable attempt at rewarding player engagement without thinking about it. (Of course it’s not completely fiat-free, as the GM must assign the TN and any situational modifiers).

    The automatic nature of XP awards has me bristling somewhat, but the main reason I don’t like the system above is the lack of a teamwork component in its cold quantitative calculation.

    Did the action in question benefit the party? Assist in the mission? Succeed but end up killing half the party? Is it at all meaningful to the adventure, or just hard? “Hey, I figured out how to trisect an angle!”. I’d imagine a fiat-less system would potentially benefit uncooperative players far too much without human analysis of the tasks accomplished. (I’m also reminded of DragonAge, where I seem to level up no matter how badly I try to play.)

  7. March 13th, 2011 at 20:13 | #7

    @deimos3428 : I may have to concede defeat on this one, at least for now. I agree that it doesn’t address teamwork as an integral component, but since the Advancement Roll is an individual attempt, the exclusion didn’t bother me. But this is based on the assumption that teamwork would be necessary to complete the mission to begin with (if you run a party of opportunists, perhaps double the award if the action benefits the group–i.e., full reward for tactical and twice the reward for strategic).

    Outside of that, the modifications I might suggest to address the objections raised require either a more pronounced level of GM fiat, or some method in pre-calculating a value that tells you just how “hard” an adventure is. Personally, I’m not upset with fiat–I don’t think it can be entirely purged from experience rewards–but I definitely don’t want to go down the “challenge rating” road (which assumes a level of predictability that I just don’t agree exists during play). Plus, some activities (like Anthony’s sandbox stuff) could easily defy such systems of categorisation.

    For now, perhaps this is best applied as an optional rule–see if it works for your group and tweak (or discard) as necessary. That said, I’m open to suggestions.

    BTW, Deimos–your comment above is #3000. How does it feel?

  8. deimos3428
    March 14th, 2011 at 11:16 | #8

    Sorry, I didn’t mean to dump on the idea. It’s just the polar opposite of how I handle things. GM fiat is a key component to RPGs in my opinion; calculation or even die rolls being secondary.

    That said if one were to soldier on with a fiat-light system, I’d say you’re on the right track. It’s an incremental improvement over previous attempts at the same. There may be a few more components you need to consider, but I wouldn’t give up on the idea outright. At most, shelve it for now and come back to it with a fresh mind later.

    “BTW, Deimos–your comment above is #3000. How does it feel?”

    Kind…old, actually. :D

  9. March 14th, 2011 at 11:59 | #9

    @deimos3428 : No apology necessary – I think you and Anthony have come up with valid concerns. This is definitely not a “one-size-fits-all” solution, though I do believe it’s more flexible than the “automatic XP –> level” approach.

    I believe I’ll include this as an optional rule, something called “Heroics,” as an adjunct to Clutch Situations. I think it has merit for some playing styles–perhaps the ultimate in GM fiat is whether or not this system fits in with yours?

    As you suggest, I’m going to continue playtesting it for the Core (I won’t include it in 5th Printing Basic) and see where it can be improved or made more universally accessible.

    As always, I appreciate the comments and suggestions – if there’s more on the topic, let’s move it over to the discussion forum (http://groups.google.com/group/chimerarpg)

  10. March 14th, 2011 at 12:28 | #10

    @Erin D. Smale
    I agree, we are on pretty much the same wavelength. I really like your system since people can opt to improve in discrete steps for certain skills rather than periodic dings that raise a level and all a character’s abilities. I feel that a standard roll to advance system for overall character level plus the Clutch mechanic for individual advancement in skills sounds like a great system. If I can find a way to graft that onto D&D, then I most certainly would.

    Personally, I am just going to stick with gold earned and HD of monsters slain. I am going to expand both a little bit and use your more broad sense of risk. Like the example I used before, taking a long, dangerous journey through war torn lands to make a profit selling goods would qualify the profit made for XP-for-GP. In the end, I feel it is possible to equate ‘risky’ behavior with some type of tangible end result that ties to either money or foes defeated. With this kind of framework, I think I can work with it without requiring a major refit to the D&D level standard.


  11. March 14th, 2011 at 16:40 | #11

    @Anthony : I appreciate the encouragement. I do believe this approach has merit, but I still don’t get the feeling it addresses all concerns. It does a good job with individuals, but not the entire party, nor does it capture the “whole” of an adventure’s challenges.

    These don’t worry me so much, but I realise they’re pretty big abstractions, and may be difficult for newcomers to get used to. I’ll continue to tinker–maybe I need more time for these bits to “sink in” before I can expand it to a truly workable-in-all-cases system.

  12. March 16th, 2011 at 20:52 | #12

    Yes, “the ‘whole’ of an adventure’s challenges” is pretty tough to capture objectively.

    Quick question, you stated that you get 1 Clutch situation per level. Do they ever replenish or is that 1 Clutch all you get until you level up and get another? Controlling the amount of Clutch situations that a given player has at any moment is a good way to control XP and throttle advancement to approximate the whole of an adventure’s challenge. Tougher adventures should have more opportunities to encounter tough situations with qualifying target numbers and therefore more opportunities to use Clutch points and advance. Easier adventures might not present any or few opportunities to use Clutch points on a qualifying challenge, thus approximating an adventure’s whole challenge. Hence, players can weigh risk and reward and have that calculation reflected in earned XP/chances to level.

  13. March 17th, 2011 at 09:59 | #13

    @Anthony You’re correct on Clutch Situtations–you get one per level, and your pool doesn’t replenish until you level up again. Put another way, at 2nd-level (for example), you get 2 Clutch Situations until you achieve 3rd-level.

    But I like where your head’s at–very interesting suggestion to use Clutch Situations as an indirect measure of challenge. It would still require some interpretation, but I could see a “2-Clutch” adventure being more difficult than a “No-Clutch” adventure. That approach could also handle your “missionless” sessions.

    It also suggests the possibility of awarding PCs with additional Clutch Situations. Gotta think on that one, but assuming PCs do cool things during a session, an ascending scale of rewards might be:

    1: +1 Bonus to Advancement Roll (to be applied after the adventure)
    2: +1 Clutch Situation (to be added to the character’s CS pool)
    3: +1 Improvement Point (to be spent immediately, regardless of Advancement Roll result)

    UPDATE – I think maybe #1 and #2 should be switched?

  14. deimos3428
    March 17th, 2011 at 21:58 | #14

    Also, I forgot to post this last week. Risk/reward is best summed up via the chart in the penultimate panel:


  15. March 17th, 2011 at 22:56 | #15

    @deimos3428 : Indeed. Not so difficult after all.

  1. No trackbacks yet.